Grade_Symbol_Identifer.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "provenance": []
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "source": [
        "Student name:\n",
        "Adam GREEN\n",
        "\n",
        "Tags used for the data generation: Mario(my inputs) and Data1,Data2,Data3\n",
        "\n",
        "Maximum accuracy achieved: NN = 96.25% (with train size 1740) & CNN = 93.92%\n",
        "\n",
        "Types of neural network used: NN and CNN"
      ],
      "metadata": {
        "id": "XAK2xvy_rQY2"
      }
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ZTtoXAi59o4d"
      },
      "source": [
        "# Instaling PyMongo, this is the interface to connect to MongoDB with Python\n",
        "! python -m pip install pymongo\n",
        "\n",
        "# Those libraries are only required for the drawing GUI we're going to use to manually give some input to our neural network.\n",
        "from IPython.display import HTML, Image\n",
        "from google.colab.output import eval_js\n",
        "from base64 import b64decode\n",
        "from datetime import datetime\n",
        "\n",
        "# Library required to connect to the database where we will store your dataset\n",
        "from pymongo import MongoClient\n",
        "from urllib.parse import quote_plus\n",
        "from ssl import SSLContext, CERT_NONE\n",
        "\n",
        "# library used for randomness\n",
        "import random\n",
        "\n",
        "# Import for multi-dimensional array manipulation\n",
        "import numpy as np\n",
        "\n",
        "# OpenCV2 library used to manipulate images\n",
        "import cv2\n",
        "\n",
        "# This import will allow us to display and plot our data\n",
        "import matplotlib.pyplot as plt\n",
        "\n",
        "# This import is needed to be able to save the model and download it\n",
        "from google.colab import files\n",
        "\n",
        "# Imports about Tensorflow. You may need to complete this section\n",
        "from tensorflow import keras\n",
        "from tensorflow.keras.models import Sequential\n",
        "from tensorflow.keras.layers import Dense, Flatten\n",
        "from keras.preprocessing.image import ImageDataGenerator"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "K6TKuM-90hhR"
      },
      "source": [
        "All the data that will be generated will be stored in a dedicated database."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "1thieBFd_Qv_"
      },
      "source": [
        "uri = f'mongodb+srv://tbs:jiTUWJJNzJRvyunL@tbs.2qto0jb.mongodb.net/?retryWrites=true&w=majority&appName=tbs'\n",
        "\n",
        "client = MongoClient(uri)\n",
        "db = client.tbs\n",
        "grades = db['grades']"
      ],
      "execution_count": 2,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "gyKCkEKc8pi2"
      },
      "source": [
        "canvas_html = \"\"\"\n",
        "<canvas width=%d height=%d style=\"border:1px solid #000000;\"></canvas><br/>\n",
        "<button>Finish</button>\n",
        "<script>\n",
        "var canvas = document.querySelector('canvas')\n",
        "var ctx = canvas.getContext('2d')\n",
        "ctx.lineWidth = %d\n",
        "ctx.lineCap = 'round';\n",
        "var button = document.querySelector('button')\n",
        "var mouse = {x: 0, y: 0}\n",
        "canvas.addEventListener('mousemove', function(e) {\n",
        "  mouse.x = e.pageX - this.offsetLeft\n",
        "  mouse.y = e.pageY - this.offsetTop\n",
        "})\n",
        "canvas.onmousedown = ()=>{\n",
        "  ctx.beginPath()\n",
        "  ctx.moveTo(mouse.x, mouse.y)\n",
        "  canvas.addEventListener('mousemove', onPaint)\n",
        "}\n",
        "canvas.onmouseup = ()=>{\n",
        "  canvas.removeEventListener('mousemove', onPaint)\n",
        "}\n",
        "var onPaint = ()=>{\n",
        "  ctx.lineTo(mouse.x, mouse.y)\n",
        "  ctx.stroke()\n",
        "}\n",
        "var data = new Promise(resolve=>{\n",
        "  button.onclick = ()=>{\n",
        "    resolve(canvas.toDataURL('image/png'))\n",
        "  }\n",
        "})\n",
        "</script>\n",
        "\"\"\"\n",
        "labels = ['A+', 'A', 'A-', 'B+', 'B', 'B-', 'C+', 'C', 'C-', 'D+', 'D', 'D-', 'E+', 'E', 'E-']\n",
        "\n",
        "# With this method you can draw manually a grade and save it to the database\n",
        "# The tags have to be comma-separated. Example: test,test2,test3\n",
        "def drawAndSaveInDatabase(tags,  filename='drawing.png', w=280, h=280, line_width=20):\n",
        "  label = random.choice(labels)\n",
        "  print(f\"Draw a {label} and then click on finish to store it.\")\n",
        "  display(HTML(canvas_html % (w, h, line_width)))\n",
        "  data = eval_js(\"data\")\n",
        "  b64 = data.split(',')[1]\n",
        "  document = {\n",
        "      'tags': tags.split(','),\n",
        "      'label': label,\n",
        "      'data': b64,\n",
        "      'date': datetime.now()\n",
        "  }\n",
        "  grades.insert_one(document)\n",
        "  print(f\"Succesfully saved grade with tags [{tags}] into the database.\")\n",
        "\n",
        "# Retrieve the images with at least one of the given tags.\n",
        "# The tags have to be comma-separated. Example: test,test2,test3\n",
        "def getDatasetFromDatabase(tags, size = 28):\n",
        "  n = grades.count_documents({'tags':{\n",
        "            '$in':tags.split(',')\n",
        "        }})\n",
        "  X = np.zeros((n,size,size))\n",
        "  Y = np.zeros(n)\n",
        "  i = 0\n",
        "  for document in grades.find({'tags':{\n",
        "            '$in':tags.split(',')\n",
        "        }}):\n",
        "    nparr = np.frombuffer(b64decode(document['data']), np.uint8)\n",
        "    img_np = (cv2.resize(cv2.imdecode(nparr, cv2.IMREAD_UNCHANGED), (size, size))[:,:,3]) / 255\n",
        "    X[i,:,:] = img_np\n",
        "    Y[i] = labels.index(document['label'])\n",
        "    i = i + 1\n",
        "  return X, Y\n",
        "\n",
        "\n",
        "def getCompleteDatasetFromDatabase(size = 28):\n",
        "  n = grades.count()\n",
        "  X = np.zeros((n,size,size))\n",
        "  Y = np.zeros(n)\n",
        "  i = 0\n",
        "  for document in grades.find():\n",
        "    nparr = np.frombuffer(b64decode(document['data']), np.uint8)\n",
        "    img_np = (cv2.resize(cv2.imdecode(nparr, cv2.IMREAD_UNCHANGED), (size, size))[:,:,3]) / 255\n",
        "    X[i,:,:] = img_np\n",
        "    Y[i] = labels.index(document['label'])\n",
        "    i = i + 1\n",
        "  return X, Y"
      ],
      "execution_count": 3,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "soyCtkWR9V-p",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 360
        },
        "outputId": "aa775782-ffde-40ab-fe3a-0f330ba81490"
      },
      "source": [
        "drawAndSaveInDatabase('test')"
      ],
      "execution_count": 8,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Draw a A- and then click on finish to store it.\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              "\n",
              "<canvas width=280 height=280 style=\"border:1px solid #000000;\"></canvas><br/>\n",
              "<button>Finish</button>\n",
              "<script>\n",
              "var canvas = document.querySelector('canvas')\n",
              "var ctx = canvas.getContext('2d')\n",
              "ctx.lineWidth = 20\n",
              "ctx.lineCap = 'round';\n",
              "var button = document.querySelector('button')\n",
              "var mouse = {x: 0, y: 0}\n",
              "canvas.addEventListener('mousemove', function(e) {\n",
              "  mouse.x = e.pageX - this.offsetLeft\n",
              "  mouse.y = e.pageY - this.offsetTop\n",
              "})\n",
              "canvas.onmousedown = ()=>{\n",
              "  ctx.beginPath()\n",
              "  ctx.moveTo(mouse.x, mouse.y)\n",
              "  canvas.addEventListener('mousemove', onPaint)\n",
              "}\n",
              "canvas.onmouseup = ()=>{\n",
              "  canvas.removeEventListener('mousemove', onPaint)\n",
              "}\n",
              "var onPaint = ()=>{\n",
              "  ctx.lineTo(mouse.x, mouse.y)\n",
              "  ctx.stroke()\n",
              "}\n",
              "var data = new Promise(resolve=>{\n",
              "  button.onclick = ()=>{\n",
              "    resolve(canvas.toDataURL('image/png'))\n",
              "  }\n",
              "})\n",
              "</script>\n"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Succesfully saved grade with tags [test] into the database.\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "This is how you can retrieve the data set from the database."
      ],
      "metadata": {
        "id": "zxrLxE2395jw"
      }
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "q51YPU8hjFY9"
      },
      "source": [
        "# Retrieve the raw dataset, with the images with the associated tags\n",
        "X_raw, Y_raw = getDatasetFromDatabase(tags='mario,data1,data2,data3',size=28)\n",
        "Y = keras.utils.to_categorical(Y_raw, 15)\n",
        "print(Y.shape)\n",
        "\n",
        "# Convert into float numbers\n",
        "X = X_raw.astype('float32')\n",
        "print(X.shape)\n",
        "\n",
        "\n",
        "# This is one way to split the dataset, taking the first 50 elements in the training set, and the remaining in the validation set\n",
        "X_train = X[0:1500,:,:]\n",
        "X_test = X[1500:,:,:]\n",
        "Y_train = Y[0:1500,:]\n",
        "Y_test = Y[1500:,:]"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "### START CODE HERE\n",
        "y = len(X_raw)\n",
        "\n",
        "print(y)\n",
        "\n",
        "#we see its a 3D array\n",
        "print(X_raw)\n",
        "print(X_raw.shape)\n",
        "print(X_raw[0,:,:])\n",
        "\n",
        "print(\"Y vars\")\n",
        "print(Y_raw)\n",
        "\n",
        "\n",
        "\n",
        "print(Y_raw.shape)\n",
        "\n",
        "### END CODE HERE"
      ],
      "metadata": {
        "id": "PKvtjpVXI8z4"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "print(Y_raw.shape)\n"
      ],
      "metadata": {
        "id": "QZeKPpAEI8wA"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "#import numpy as np\n",
        "\n",
        "#Check the unique values in Y_raw and their maximum\n",
        "unique_vals = np.unique(Y_raw)\n",
        "max_val = np.max(unique_vals)\n",
        "\n",
        "print(\"Unique values in Y_raw:\", unique_vals)\n",
        "print(\"Maximum value in Y_raw:\", max_val)\n"
      ],
      "metadata": {
        "id": "Cv6jYVdII8Pa"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "Y_train = keras.utils.to_categorical(Y_raw, 15)\n",
        "Y_test = keras.utils.to_categorical(Y_raw, 15)\n",
        "print(Y_train.shape)\n",
        "print(Y_test.shape)\n",
        "print(Y_train[0:10, :])\n",
        "\n",
        "X_train_raw = X_raw.astype('float32')\n",
        "X_test_raw = X_raw.astype('float32')\n",
        "\n",
        "X_train = X_train_raw / 255\n",
        "X_test = X_test_raw / 255"
      ],
      "metadata": {
        "id": "51myIP1iI8l3"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "#USE THIS TO VIEW DATA\n",
        "\n",
        "# This method should display the picture at index i in the dataset X, and also print the corresponding label in Y\n",
        "def visualize(X, Y, i):\n",
        "  ### START CODE HERE\n",
        "\n",
        "    plt.imshow(X[i,:,:],cmap=\"gray\",vmin=0,vmax=1)\n",
        "\n",
        "    plt.show()\n",
        "    print(\"This is a\", np.argmax(Y[i]))\n",
        "\n",
        "\n",
        "  ### END CODE HERE\n",
        "\n",
        "visualize(X_raw, Y_raw, 6)\n",
        "\n"
      ],
      "metadata": {
        "id": "XwdgwMZqMMU5"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "def visualize_training(history):\n",
        "  plt.plot(history.history['accuracy'])\n",
        "  plt.plot(history.history['val_accuracy'])\n",
        "  plt.title('model accuracy')\n",
        "  plt.ylabel('accuracy')\n",
        "  plt.xlabel('epoch')\n",
        "  plt.legend(['training', 'validation'], loc='lower right')\n",
        "  plt.show()\n",
        "\n",
        "  plt.plot(history.history['loss'])\n",
        "  plt.plot(history.history['val_loss'])\n",
        "  plt.title('model loss')\n",
        "  plt.ylabel('loss')\n",
        "  plt.xlabel('epoch')\n",
        "  plt.legend(['training', 'validation'], loc='upper left')\n",
        "  plt.show()\n",
        "\n"
      ],
      "metadata": {
        "id": "a7a-03AGN3bB"
      },
      "execution_count": 41,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "def drawAndGuess(model, filename='drawing.png', w=280, h=280, line_width=20):\n",
        "  print(\"Draw a score and then click on finish to try to recognize it.\")\n",
        "  display(HTML(canvas_html % (w, h, line_width)))\n",
        "  data = eval_js(\"data\")\n",
        "  binary = b64decode(data.split(',')[1])\n",
        "  image = np.zeros((1,28,28));\n",
        "\n",
        "  with open(filename, 'wb') as f:\n",
        "    f.write(binary)\n",
        "  image[0,:,:] = (cv2.resize(cv2.imread(filename, cv2.IMREAD_UNCHANGED), (28, 28))[:,:,3]) / 255\n",
        "  result = model.predict(image)\n",
        "  index = np.argmax(result[0])\n",
        "  confidence = np.round(result[0][index] * 100)\n",
        "  print(f\"This is a {labels[index]} with confidence {confidence}%.\")\n",
        "\n"
      ],
      "metadata": {
        "id": "wPEJPuHvMMBF"
      },
      "execution_count": 84,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "def indexes_bad_prediction(X, Y):\n",
        "    result = []\n",
        "    # Perform the prediction on the full dataset\n",
        "    Y_pred = model.predict(X)\n",
        "    # Loop over the full dataset and find the misclassified predictions\n",
        "    for i in range(len(X)):\n",
        "        if np.argmax(Y[i]) != np.argmax(Y_pred[i]):\n",
        "            result.append(i)\n",
        "    return result"
      ],
      "metadata": {
        "id": "TE_eEx5JxC9M"
      },
      "execution_count": 137,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "# **Base Model: Augmentation 1**"
      ],
      "metadata": {
        "id": "AO3HVeKycSDV"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Take only the first 500 samples of the training dataset\n",
        "X_train_small = X_train[0:500, :, :].reshape((-1, 28, 28, 1))\n",
        "Y_train_small = Y_train[0:500, :]\n",
        "\n",
        "# The ImageDataGenerator will randomly apply some transformation to the input images before passing them to the neural network\n",
        "generator = ImageDataGenerator(rotation_range=20, width_shift_range=0.2, height_shift_range=0.2, zoom_range = 0.1, shear_range = 10)\n",
        "\n",
        "# Define the model\n",
        "model = Sequential()\n",
        "model.add(Flatten(input_shape=(28, 28)))\n",
        "model.add(Dense(32, activation='relu'))\n",
        "model.add(Dense(16, activation='relu'))\n",
        "model.add(Dense(15, activation='softmax'))\n",
        "\n",
        "model.compile(loss=keras.losses.categorical_crossentropy,\n",
        "              optimizer=keras.optimizers.Adam(),\n",
        "              metrics=['accuracy'])\n",
        "\n",
        "# Train the model\n",
        "# Use Model.fit, which supports generators.\n",
        "history = model.fit(\n",
        "    generator.flow(X_train_small, Y_train_small, batch_size=32),\n",
        "    steps_per_epoch=len(X_train_small) / 32,  # Adjusted to match the size of X_train_small\n",
        "    epochs=2000,\n",
        "    validation_data=(X_test, Y_test)  # Ensure X_test is properly reshaped if needed\n",
        ")\n",
        "\n",
        "\n",
        "# Visualize the training evolution\n",
        "visualize_training(history)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "GVdfQRbZVXq4",
        "outputId": "53f81805-ab32-4388-9446-3a989df60274"
      },
      "execution_count": 82,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/2000\n",
            "3/3 [==============================] - 2s 273ms/step - loss: 2.7588 - accuracy: 0.0500 - val_loss: 2.6941 - val_accuracy: 0.0426\n",
            "Epoch 2/2000\n",
            "3/3 [==============================] - 0s 54ms/step - loss: 2.7157 - accuracy: 0.0200 - val_loss: 2.6945 - val_accuracy: 0.0638\n",
            "Epoch 3/2000\n",
            "3/3 [==============================] - 0s 29ms/step - loss: 2.6834 - accuracy: 0.0800 - val_loss: 2.7022 - val_accuracy: 0.0426\n",
            "Epoch 4/2000\n",
            "3/3 [==============================] - 0s 39ms/step - loss: 2.6948 - accuracy: 0.0900 - val_loss: 2.7065 - val_accuracy: 0.0426\n",
            "Epoch 5/2000\n",
            "3/3 [==============================] - 0s 34ms/step - loss: 2.6743 - accuracy: 0.0500 - val_loss: 2.7014 - val_accuracy: 0.0213\n",
            "Epoch 6/2000\n",
            "3/3 [==============================] - 0s 33ms/step - loss: 2.6520 - accuracy: 0.0900 - val_loss: 2.6994 - val_accuracy: 0.0213\n",
            "Epoch 7/2000\n",
            "3/3 [==============================] - 0s 29ms/step - loss: 2.6393 - accuracy: 0.0600 - val_loss: 2.7049 - val_accuracy: 0.0213\n",
            "Epoch 8/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 2.6361 - accuracy: 0.1200 - val_loss: 2.7129 - val_accuracy: 0.0213\n",
            "Epoch 9/2000\n",
            "3/3 [==============================] - 0s 40ms/step - loss: 2.6343 - accuracy: 0.1000 - val_loss: 2.7190 - val_accuracy: 0.0213\n",
            "Epoch 10/2000\n",
            "3/3 [==============================] - 0s 33ms/step - loss: 2.6210 - accuracy: 0.1200 - val_loss: 2.7260 - val_accuracy: 0.0213\n",
            "Epoch 11/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 2.6011 - accuracy: 0.0800 - val_loss: 2.7461 - val_accuracy: 0.0426\n",
            "Epoch 12/2000\n",
            "3/3 [==============================] - 0s 38ms/step - loss: 2.5999 - accuracy: 0.1400 - val_loss: 2.7595 - val_accuracy: 0.0213\n",
            "Epoch 13/2000\n",
            "3/3 [==============================] - 0s 45ms/step - loss: 2.6103 - accuracy: 0.1000 - val_loss: 2.7637 - val_accuracy: 0.0213\n",
            "Epoch 14/2000\n",
            "3/3 [==============================] - 0s 38ms/step - loss: 2.5829 - accuracy: 0.1300 - val_loss: 2.7601 - val_accuracy: 0.0213\n",
            "Epoch 15/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 2.5693 - accuracy: 0.1300 - val_loss: 2.7698 - val_accuracy: 0.0000e+00\n",
            "Epoch 16/2000\n",
            "3/3 [==============================] - 0s 32ms/step - loss: 2.6085 - accuracy: 0.1100 - val_loss: 2.7724 - val_accuracy: 0.0000e+00\n",
            "Epoch 17/2000\n",
            "3/3 [==============================] - 0s 30ms/step - loss: 2.5763 - accuracy: 0.1400 - val_loss: 2.7595 - val_accuracy: 0.0000e+00\n",
            "Epoch 18/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 2.5665 - accuracy: 0.1300 - val_loss: 2.7521 - val_accuracy: 0.0000e+00\n",
            "Epoch 19/2000\n",
            "3/3 [==============================] - 0s 29ms/step - loss: 2.5628 - accuracy: 0.1100 - val_loss: 2.7531 - val_accuracy: 0.0000e+00\n",
            "Epoch 20/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 2.5579 - accuracy: 0.1300 - val_loss: 2.7487 - val_accuracy: 0.0000e+00\n",
            "Epoch 21/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 2.5525 - accuracy: 0.1100 - val_loss: 2.7386 - val_accuracy: 0.0000e+00\n",
            "Epoch 22/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 2.4827 - accuracy: 0.2100 - val_loss: 2.7201 - val_accuracy: 0.0000e+00\n",
            "Epoch 23/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 2.5356 - accuracy: 0.1300 - val_loss: 2.6848 - val_accuracy: 0.0000e+00\n",
            "Epoch 24/2000\n",
            "3/3 [==============================] - 0s 38ms/step - loss: 2.5064 - accuracy: 0.1500 - val_loss: 2.6635 - val_accuracy: 0.0000e+00\n",
            "Epoch 25/2000\n",
            "3/3 [==============================] - 0s 38ms/step - loss: 2.5286 - accuracy: 0.2000 - val_loss: 2.6540 - val_accuracy: 0.0213\n",
            "Epoch 26/2000\n",
            "3/3 [==============================] - 0s 33ms/step - loss: 2.5073 - accuracy: 0.1600 - val_loss: 2.6535 - val_accuracy: 0.0213\n",
            "Epoch 27/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 2.5356 - accuracy: 0.1600 - val_loss: 2.6572 - val_accuracy: 0.0213\n",
            "Epoch 28/2000\n",
            "3/3 [==============================] - 0s 33ms/step - loss: 2.4671 - accuracy: 0.2200 - val_loss: 2.6549 - val_accuracy: 0.0213\n",
            "Epoch 29/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 2.5119 - accuracy: 0.1600 - val_loss: 2.6590 - val_accuracy: 0.0213\n",
            "Epoch 30/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 2.4439 - accuracy: 0.1500 - val_loss: 2.6621 - val_accuracy: 0.0213\n",
            "Epoch 31/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 2.5151 - accuracy: 0.1200 - val_loss: 2.6605 - val_accuracy: 0.0213\n",
            "Epoch 32/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 2.4332 - accuracy: 0.1900 - val_loss: 2.6551 - val_accuracy: 0.0638\n",
            "Epoch 33/2000\n",
            "3/3 [==============================] - 0s 39ms/step - loss: 2.4869 - accuracy: 0.1900 - val_loss: 2.6423 - val_accuracy: 0.0638\n",
            "Epoch 34/2000\n",
            "3/3 [==============================] - 0s 39ms/step - loss: 2.4768 - accuracy: 0.1300 - val_loss: 2.6432 - val_accuracy: 0.0638\n",
            "Epoch 35/2000\n",
            "3/3 [==============================] - 0s 31ms/step - loss: 2.4784 - accuracy: 0.1600 - val_loss: 2.6364 - val_accuracy: 0.0851\n",
            "Epoch 36/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 2.4536 - accuracy: 0.1900 - val_loss: 2.6195 - val_accuracy: 0.1064\n",
            "Epoch 37/2000\n",
            "3/3 [==============================] - 0s 28ms/step - loss: 2.4401 - accuracy: 0.1900 - val_loss: 2.6048 - val_accuracy: 0.1277\n",
            "Epoch 38/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 2.4395 - accuracy: 0.2000 - val_loss: 2.5938 - val_accuracy: 0.1489\n",
            "Epoch 39/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 2.4862 - accuracy: 0.1500 - val_loss: 2.5729 - val_accuracy: 0.1277\n",
            "Epoch 40/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 2.4600 - accuracy: 0.1900 - val_loss: 2.5578 - val_accuracy: 0.1489\n",
            "Epoch 41/2000\n",
            "3/3 [==============================] - 0s 31ms/step - loss: 2.4301 - accuracy: 0.2600 - val_loss: 2.5552 - val_accuracy: 0.1489\n",
            "Epoch 42/2000\n",
            "3/3 [==============================] - 0s 34ms/step - loss: 2.4132 - accuracy: 0.2000 - val_loss: 2.5731 - val_accuracy: 0.1489\n",
            "Epoch 43/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 2.3636 - accuracy: 0.2000 - val_loss: 2.5870 - val_accuracy: 0.1489\n",
            "Epoch 44/2000\n",
            "3/3 [==============================] - 0s 32ms/step - loss: 2.4097 - accuracy: 0.2200 - val_loss: 2.5821 - val_accuracy: 0.1702\n",
            "Epoch 45/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 2.3501 - accuracy: 0.1800 - val_loss: 2.5765 - val_accuracy: 0.1702\n",
            "Epoch 46/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 2.3236 - accuracy: 0.2500 - val_loss: 2.5589 - val_accuracy: 0.1702\n",
            "Epoch 47/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 2.3810 - accuracy: 0.2200 - val_loss: 2.5411 - val_accuracy: 0.1915\n",
            "Epoch 48/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 2.3814 - accuracy: 0.2900 - val_loss: 2.5130 - val_accuracy: 0.1915\n",
            "Epoch 49/2000\n",
            "3/3 [==============================] - 0s 29ms/step - loss: 2.3747 - accuracy: 0.2100 - val_loss: 2.4889 - val_accuracy: 0.1915\n",
            "Epoch 50/2000\n",
            "3/3 [==============================] - 0s 29ms/step - loss: 2.3808 - accuracy: 0.1800 - val_loss: 2.4768 - val_accuracy: 0.1915\n",
            "Epoch 51/2000\n",
            "3/3 [==============================] - 0s 33ms/step - loss: 2.2918 - accuracy: 0.3300 - val_loss: 2.4837 - val_accuracy: 0.1702\n",
            "Epoch 52/2000\n",
            "3/3 [==============================] - 0s 39ms/step - loss: 2.4139 - accuracy: 0.2100 - val_loss: 2.4971 - val_accuracy: 0.1702\n",
            "Epoch 53/2000\n",
            "3/3 [==============================] - 0s 45ms/step - loss: 2.4075 - accuracy: 0.2200 - val_loss: 2.4877 - val_accuracy: 0.1702\n",
            "Epoch 54/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 2.2901 - accuracy: 0.2700 - val_loss: 2.4572 - val_accuracy: 0.2128\n",
            "Epoch 55/2000\n",
            "3/3 [==============================] - 0s 48ms/step - loss: 2.3609 - accuracy: 0.2200 - val_loss: 2.4239 - val_accuracy: 0.2340\n",
            "Epoch 56/2000\n",
            "3/3 [==============================] - 0s 47ms/step - loss: 2.2919 - accuracy: 0.2400 - val_loss: 2.3974 - val_accuracy: 0.2340\n",
            "Epoch 57/2000\n",
            "3/3 [==============================] - 0s 49ms/step - loss: 2.3025 - accuracy: 0.2200 - val_loss: 2.3739 - val_accuracy: 0.2340\n",
            "Epoch 58/2000\n",
            "3/3 [==============================] - 0s 54ms/step - loss: 2.3725 - accuracy: 0.1900 - val_loss: 2.3655 - val_accuracy: 0.2128\n",
            "Epoch 59/2000\n",
            "3/3 [==============================] - 0s 50ms/step - loss: 2.3601 - accuracy: 0.2500 - val_loss: 2.3642 - val_accuracy: 0.2128\n",
            "Epoch 60/2000\n",
            "3/3 [==============================] - 0s 64ms/step - loss: 2.2677 - accuracy: 0.2400 - val_loss: 2.3709 - val_accuracy: 0.1702\n",
            "Epoch 61/2000\n",
            "3/3 [==============================] - 0s 49ms/step - loss: 2.2964 - accuracy: 0.2400 - val_loss: 2.3651 - val_accuracy: 0.1702\n",
            "Epoch 62/2000\n",
            "3/3 [==============================] - 0s 59ms/step - loss: 2.2974 - accuracy: 0.1700 - val_loss: 2.3624 - val_accuracy: 0.1489\n",
            "Epoch 63/2000\n",
            "3/3 [==============================] - 0s 57ms/step - loss: 2.3410 - accuracy: 0.2200 - val_loss: 2.3671 - val_accuracy: 0.1277\n",
            "Epoch 64/2000\n",
            "3/3 [==============================] - 0s 38ms/step - loss: 2.3082 - accuracy: 0.2900 - val_loss: 2.3854 - val_accuracy: 0.1277\n",
            "Epoch 65/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 2.2733 - accuracy: 0.2500 - val_loss: 2.3956 - val_accuracy: 0.1702\n",
            "Epoch 66/2000\n",
            "3/3 [==============================] - 0s 32ms/step - loss: 2.2176 - accuracy: 0.2600 - val_loss: 2.3776 - val_accuracy: 0.2128\n",
            "Epoch 67/2000\n",
            "3/3 [==============================] - 0s 38ms/step - loss: 2.3034 - accuracy: 0.2700 - val_loss: 2.3550 - val_accuracy: 0.2128\n",
            "Epoch 68/2000\n",
            "3/3 [==============================] - 0s 29ms/step - loss: 2.2774 - accuracy: 0.2100 - val_loss: 2.3321 - val_accuracy: 0.2128\n",
            "Epoch 69/2000\n",
            "3/3 [==============================] - 0s 33ms/step - loss: 2.2478 - accuracy: 0.3000 - val_loss: 2.3089 - val_accuracy: 0.2128\n",
            "Epoch 70/2000\n",
            "3/3 [==============================] - 0s 53ms/step - loss: 2.3067 - accuracy: 0.2400 - val_loss: 2.2892 - val_accuracy: 0.2340\n",
            "Epoch 71/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 2.2713 - accuracy: 0.2500 - val_loss: 2.2920 - val_accuracy: 0.2340\n",
            "Epoch 72/2000\n",
            "3/3 [==============================] - 0s 29ms/step - loss: 2.2301 - accuracy: 0.2800 - val_loss: 2.3025 - val_accuracy: 0.2340\n",
            "Epoch 73/2000\n",
            "3/3 [==============================] - 0s 38ms/step - loss: 2.2180 - accuracy: 0.2700 - val_loss: 2.3098 - val_accuracy: 0.2340\n",
            "Epoch 74/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 2.3040 - accuracy: 0.1900 - val_loss: 2.3357 - val_accuracy: 0.2128\n",
            "Epoch 75/2000\n",
            "3/3 [==============================] - 0s 32ms/step - loss: 2.2054 - accuracy: 0.2200 - val_loss: 2.3669 - val_accuracy: 0.1702\n",
            "Epoch 76/2000\n",
            "3/3 [==============================] - 0s 38ms/step - loss: 2.1426 - accuracy: 0.3100 - val_loss: 2.3606 - val_accuracy: 0.1915\n",
            "Epoch 77/2000\n",
            "3/3 [==============================] - 0s 32ms/step - loss: 2.2268 - accuracy: 0.2500 - val_loss: 2.3649 - val_accuracy: 0.1702\n",
            "Epoch 78/2000\n",
            "3/3 [==============================] - 0s 29ms/step - loss: 2.1190 - accuracy: 0.2600 - val_loss: 2.3037 - val_accuracy: 0.1702\n",
            "Epoch 79/2000\n",
            "3/3 [==============================] - 0s 29ms/step - loss: 2.1611 - accuracy: 0.2600 - val_loss: 2.2668 - val_accuracy: 0.1915\n",
            "Epoch 80/2000\n",
            "3/3 [==============================] - 0s 29ms/step - loss: 2.1746 - accuracy: 0.2600 - val_loss: 2.2783 - val_accuracy: 0.2340\n",
            "Epoch 81/2000\n",
            "3/3 [==============================] - 0s 30ms/step - loss: 2.2401 - accuracy: 0.2400 - val_loss: 2.3068 - val_accuracy: 0.2340\n",
            "Epoch 82/2000\n",
            "3/3 [==============================] - 0s 29ms/step - loss: 2.0979 - accuracy: 0.2500 - val_loss: 2.3038 - val_accuracy: 0.2340\n",
            "Epoch 83/2000\n",
            "3/3 [==============================] - 0s 40ms/step - loss: 2.1488 - accuracy: 0.2700 - val_loss: 2.2450 - val_accuracy: 0.1915\n",
            "Epoch 84/2000\n",
            "3/3 [==============================] - 0s 31ms/step - loss: 2.2478 - accuracy: 0.2200 - val_loss: 2.2229 - val_accuracy: 0.1915\n",
            "Epoch 85/2000\n",
            "3/3 [==============================] - 0s 40ms/step - loss: 2.0944 - accuracy: 0.2900 - val_loss: 2.2092 - val_accuracy: 0.2128\n",
            "Epoch 86/2000\n",
            "3/3 [==============================] - 0s 38ms/step - loss: 2.1006 - accuracy: 0.3400 - val_loss: 2.1915 - val_accuracy: 0.2553\n",
            "Epoch 87/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 2.1574 - accuracy: 0.2800 - val_loss: 2.1477 - val_accuracy: 0.2553\n",
            "Epoch 88/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 2.1687 - accuracy: 0.3100 - val_loss: 2.1244 - val_accuracy: 0.2766\n",
            "Epoch 89/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 2.1467 - accuracy: 0.2900 - val_loss: 2.1087 - val_accuracy: 0.2553\n",
            "Epoch 90/2000\n",
            "3/3 [==============================] - 0s 28ms/step - loss: 2.1276 - accuracy: 0.3200 - val_loss: 2.1112 - val_accuracy: 0.2553\n",
            "Epoch 91/2000\n",
            "3/3 [==============================] - 0s 33ms/step - loss: 2.1916 - accuracy: 0.2700 - val_loss: 2.1491 - val_accuracy: 0.2128\n",
            "Epoch 92/2000\n",
            "3/3 [==============================] - 0s 33ms/step - loss: 2.1911 - accuracy: 0.2600 - val_loss: 2.2240 - val_accuracy: 0.1702\n",
            "Epoch 93/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 2.0509 - accuracy: 0.2900 - val_loss: 2.2232 - val_accuracy: 0.1915\n",
            "Epoch 94/2000\n",
            "3/3 [==============================] - 0s 32ms/step - loss: 2.1042 - accuracy: 0.3200 - val_loss: 2.1517 - val_accuracy: 0.2340\n",
            "Epoch 95/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 2.0565 - accuracy: 0.2800 - val_loss: 2.0911 - val_accuracy: 0.2340\n",
            "Epoch 96/2000\n",
            "3/3 [==============================] - 0s 33ms/step - loss: 2.1384 - accuracy: 0.2100 - val_loss: 2.0472 - val_accuracy: 0.2340\n",
            "Epoch 97/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 2.1714 - accuracy: 0.3300 - val_loss: 2.0420 - val_accuracy: 0.2553\n",
            "Epoch 98/2000\n",
            "3/3 [==============================] - 0s 28ms/step - loss: 2.1422 - accuracy: 0.2400 - val_loss: 2.0339 - val_accuracy: 0.2553\n",
            "Epoch 99/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 2.0740 - accuracy: 0.2700 - val_loss: 2.0241 - val_accuracy: 0.2553\n",
            "Epoch 100/2000\n",
            "3/3 [==============================] - 0s 31ms/step - loss: 2.0875 - accuracy: 0.2900 - val_loss: 2.0203 - val_accuracy: 0.2553\n",
            "Epoch 101/2000\n",
            "3/3 [==============================] - 0s 40ms/step - loss: 2.1251 - accuracy: 0.2900 - val_loss: 2.0077 - val_accuracy: 0.2553\n",
            "Epoch 102/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 1.9913 - accuracy: 0.2900 - val_loss: 1.9939 - val_accuracy: 0.2553\n",
            "Epoch 103/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 2.0885 - accuracy: 0.3000 - val_loss: 1.9911 - val_accuracy: 0.2340\n",
            "Epoch 104/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 2.0041 - accuracy: 0.3300 - val_loss: 2.0118 - val_accuracy: 0.2340\n",
            "Epoch 105/2000\n",
            "3/3 [==============================] - 0s 39ms/step - loss: 2.0282 - accuracy: 0.3200 - val_loss: 2.0382 - val_accuracy: 0.2553\n",
            "Epoch 106/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 1.9790 - accuracy: 0.3600 - val_loss: 2.0255 - val_accuracy: 0.2340\n",
            "Epoch 107/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 2.0334 - accuracy: 0.3000 - val_loss: 1.9783 - val_accuracy: 0.2553\n",
            "Epoch 108/2000\n",
            "3/3 [==============================] - 0s 29ms/step - loss: 1.9656 - accuracy: 0.3500 - val_loss: 1.9558 - val_accuracy: 0.2553\n",
            "Epoch 109/2000\n",
            "3/3 [==============================] - 0s 30ms/step - loss: 2.0393 - accuracy: 0.3100 - val_loss: 1.9398 - val_accuracy: 0.2766\n",
            "Epoch 110/2000\n",
            "3/3 [==============================] - 0s 38ms/step - loss: 2.0396 - accuracy: 0.2600 - val_loss: 1.9205 - val_accuracy: 0.2979\n",
            "Epoch 111/2000\n",
            "3/3 [==============================] - 0s 40ms/step - loss: 2.0109 - accuracy: 0.3100 - val_loss: 1.9293 - val_accuracy: 0.3191\n",
            "Epoch 112/2000\n",
            "3/3 [==============================] - 0s 39ms/step - loss: 1.9877 - accuracy: 0.2800 - val_loss: 1.9474 - val_accuracy: 0.2979\n",
            "Epoch 113/2000\n",
            "3/3 [==============================] - 0s 30ms/step - loss: 1.9465 - accuracy: 0.3200 - val_loss: 1.9588 - val_accuracy: 0.2766\n",
            "Epoch 114/2000\n",
            "3/3 [==============================] - 0s 29ms/step - loss: 1.9771 - accuracy: 0.3500 - val_loss: 1.9901 - val_accuracy: 0.2766\n",
            "Epoch 115/2000\n",
            "3/3 [==============================] - 0s 40ms/step - loss: 1.9109 - accuracy: 0.3500 - val_loss: 1.9811 - val_accuracy: 0.2766\n",
            "Epoch 116/2000\n",
            "3/3 [==============================] - 0s 33ms/step - loss: 1.9341 - accuracy: 0.3200 - val_loss: 1.9447 - val_accuracy: 0.2553\n",
            "Epoch 117/2000\n",
            "3/3 [==============================] - 0s 45ms/step - loss: 2.0371 - accuracy: 0.2200 - val_loss: 1.9178 - val_accuracy: 0.2979\n",
            "Epoch 118/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 1.9750 - accuracy: 0.3300 - val_loss: 1.8871 - val_accuracy: 0.3191\n",
            "Epoch 119/2000\n",
            "3/3 [==============================] - 0s 39ms/step - loss: 1.9249 - accuracy: 0.3700 - val_loss: 1.8646 - val_accuracy: 0.3191\n",
            "Epoch 120/2000\n",
            "3/3 [==============================] - 0s 38ms/step - loss: 1.8711 - accuracy: 0.3600 - val_loss: 1.8420 - val_accuracy: 0.2766\n",
            "Epoch 121/2000\n",
            "3/3 [==============================] - 0s 39ms/step - loss: 1.9132 - accuracy: 0.3400 - val_loss: 1.8325 - val_accuracy: 0.3191\n",
            "Epoch 122/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 1.8848 - accuracy: 0.4200 - val_loss: 1.8525 - val_accuracy: 0.3404\n",
            "Epoch 123/2000\n",
            "3/3 [==============================] - 0s 31ms/step - loss: 2.0101 - accuracy: 0.3500 - val_loss: 1.8558 - val_accuracy: 0.3404\n",
            "Epoch 124/2000\n",
            "3/3 [==============================] - 0s 30ms/step - loss: 1.8655 - accuracy: 0.3700 - val_loss: 1.8387 - val_accuracy: 0.3191\n",
            "Epoch 125/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 1.9459 - accuracy: 0.3800 - val_loss: 1.8197 - val_accuracy: 0.3191\n",
            "Epoch 126/2000\n",
            "3/3 [==============================] - 0s 32ms/step - loss: 1.8962 - accuracy: 0.3700 - val_loss: 1.8089 - val_accuracy: 0.3404\n",
            "Epoch 127/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 1.8026 - accuracy: 0.4300 - val_loss: 1.8169 - val_accuracy: 0.3830\n",
            "Epoch 128/2000\n",
            "3/3 [==============================] - 0s 38ms/step - loss: 1.9110 - accuracy: 0.3500 - val_loss: 1.8343 - val_accuracy: 0.3617\n",
            "Epoch 129/2000\n",
            "3/3 [==============================] - 0s 38ms/step - loss: 1.8360 - accuracy: 0.3500 - val_loss: 1.8532 - val_accuracy: 0.3404\n",
            "Epoch 130/2000\n",
            "3/3 [==============================] - 0s 44ms/step - loss: 1.7567 - accuracy: 0.3800 - val_loss: 1.8613 - val_accuracy: 0.3191\n",
            "Epoch 131/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 1.8671 - accuracy: 0.3800 - val_loss: 1.8488 - val_accuracy: 0.3191\n",
            "Epoch 132/2000\n",
            "3/3 [==============================] - 0s 30ms/step - loss: 2.0093 - accuracy: 0.2800 - val_loss: 1.8303 - val_accuracy: 0.3617\n",
            "Epoch 133/2000\n",
            "3/3 [==============================] - 0s 31ms/step - loss: 1.9691 - accuracy: 0.3400 - val_loss: 1.8128 - val_accuracy: 0.4043\n",
            "Epoch 134/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 1.8210 - accuracy: 0.3500 - val_loss: 1.7988 - val_accuracy: 0.4043\n",
            "Epoch 135/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 1.8516 - accuracy: 0.4100 - val_loss: 1.7872 - val_accuracy: 0.4894\n",
            "Epoch 136/2000\n",
            "3/3 [==============================] - 0s 38ms/step - loss: 1.7950 - accuracy: 0.4200 - val_loss: 1.7742 - val_accuracy: 0.5106\n",
            "Epoch 137/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 1.8043 - accuracy: 0.3500 - val_loss: 1.7780 - val_accuracy: 0.4468\n",
            "Epoch 138/2000\n",
            "3/3 [==============================] - 0s 34ms/step - loss: 1.8227 - accuracy: 0.3500 - val_loss: 1.7749 - val_accuracy: 0.4681\n",
            "Epoch 139/2000\n",
            "3/3 [==============================] - 0s 55ms/step - loss: 1.9348 - accuracy: 0.3200 - val_loss: 1.7653 - val_accuracy: 0.4681\n",
            "Epoch 140/2000\n",
            "3/3 [==============================] - 0s 48ms/step - loss: 1.9293 - accuracy: 0.3300 - val_loss: 1.7645 - val_accuracy: 0.4255\n",
            "Epoch 141/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 1.9650 - accuracy: 0.3600 - val_loss: 1.7603 - val_accuracy: 0.4043\n",
            "Epoch 142/2000\n",
            "3/3 [==============================] - 0s 40ms/step - loss: 1.8225 - accuracy: 0.3700 - val_loss: 1.7435 - val_accuracy: 0.4468\n",
            "Epoch 143/2000\n",
            "3/3 [==============================] - 0s 65ms/step - loss: 1.8125 - accuracy: 0.3800 - val_loss: 1.7110 - val_accuracy: 0.4468\n",
            "Epoch 144/2000\n",
            "3/3 [==============================] - 0s 51ms/step - loss: 1.8988 - accuracy: 0.3300 - val_loss: 1.6993 - val_accuracy: 0.4468\n",
            "Epoch 145/2000\n",
            "3/3 [==============================] - 0s 50ms/step - loss: 1.6912 - accuracy: 0.4100 - val_loss: 1.7097 - val_accuracy: 0.4043\n",
            "Epoch 146/2000\n",
            "3/3 [==============================] - 0s 54ms/step - loss: 1.8580 - accuracy: 0.3800 - val_loss: 1.7019 - val_accuracy: 0.4681\n",
            "Epoch 147/2000\n",
            "3/3 [==============================] - 0s 30ms/step - loss: 1.7439 - accuracy: 0.4200 - val_loss: 1.7059 - val_accuracy: 0.4255\n",
            "Epoch 148/2000\n",
            "3/3 [==============================] - 0s 40ms/step - loss: 1.8011 - accuracy: 0.3800 - val_loss: 1.7141 - val_accuracy: 0.4255\n",
            "Epoch 149/2000\n",
            "3/3 [==============================] - 0s 30ms/step - loss: 1.7920 - accuracy: 0.3600 - val_loss: 1.6728 - val_accuracy: 0.4468\n",
            "Epoch 150/2000\n",
            "3/3 [==============================] - 0s 38ms/step - loss: 1.7135 - accuracy: 0.4400 - val_loss: 1.6411 - val_accuracy: 0.4043\n",
            "Epoch 151/2000\n",
            "3/3 [==============================] - 0s 31ms/step - loss: 1.7345 - accuracy: 0.3900 - val_loss: 1.6317 - val_accuracy: 0.4468\n",
            "Epoch 152/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 1.6761 - accuracy: 0.4200 - val_loss: 1.6264 - val_accuracy: 0.4681\n",
            "Epoch 153/2000\n",
            "3/3 [==============================] - 0s 31ms/step - loss: 1.7137 - accuracy: 0.4200 - val_loss: 1.6134 - val_accuracy: 0.4681\n",
            "Epoch 154/2000\n",
            "3/3 [==============================] - 0s 44ms/step - loss: 1.7665 - accuracy: 0.3700 - val_loss: 1.5911 - val_accuracy: 0.4681\n",
            "Epoch 155/2000\n",
            "3/3 [==============================] - 0s 29ms/step - loss: 1.7625 - accuracy: 0.4000 - val_loss: 1.5808 - val_accuracy: 0.4255\n",
            "Epoch 156/2000\n",
            "3/3 [==============================] - 0s 34ms/step - loss: 1.7193 - accuracy: 0.4600 - val_loss: 1.5741 - val_accuracy: 0.4894\n",
            "Epoch 157/2000\n",
            "3/3 [==============================] - 0s 38ms/step - loss: 1.7593 - accuracy: 0.3800 - val_loss: 1.5565 - val_accuracy: 0.5106\n",
            "Epoch 158/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 1.7174 - accuracy: 0.3600 - val_loss: 1.5499 - val_accuracy: 0.5745\n",
            "Epoch 159/2000\n",
            "3/3 [==============================] - 0s 29ms/step - loss: 1.6925 - accuracy: 0.4100 - val_loss: 1.5619 - val_accuracy: 0.5106\n",
            "Epoch 160/2000\n",
            "3/3 [==============================] - 0s 38ms/step - loss: 1.7872 - accuracy: 0.4300 - val_loss: 1.5737 - val_accuracy: 0.5106\n",
            "Epoch 161/2000\n",
            "3/3 [==============================] - 0s 31ms/step - loss: 1.7347 - accuracy: 0.4300 - val_loss: 1.5773 - val_accuracy: 0.4681\n",
            "Epoch 162/2000\n",
            "3/3 [==============================] - 0s 52ms/step - loss: 1.8473 - accuracy: 0.3500 - val_loss: 1.6034 - val_accuracy: 0.4043\n",
            "Epoch 163/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 1.7814 - accuracy: 0.3500 - val_loss: 1.6141 - val_accuracy: 0.4043\n",
            "Epoch 164/2000\n",
            "3/3 [==============================] - 0s 38ms/step - loss: 1.7183 - accuracy: 0.3700 - val_loss: 1.6111 - val_accuracy: 0.4043\n",
            "Epoch 165/2000\n",
            "3/3 [==============================] - 0s 40ms/step - loss: 1.8894 - accuracy: 0.3100 - val_loss: 1.5942 - val_accuracy: 0.4894\n",
            "Epoch 166/2000\n",
            "3/3 [==============================] - 0s 38ms/step - loss: 1.7367 - accuracy: 0.4600 - val_loss: 1.5930 - val_accuracy: 0.4468\n",
            "Epoch 167/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 1.7733 - accuracy: 0.3800 - val_loss: 1.6190 - val_accuracy: 0.4043\n",
            "Epoch 168/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 1.6865 - accuracy: 0.4800 - val_loss: 1.6151 - val_accuracy: 0.4043\n",
            "Epoch 169/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 1.6936 - accuracy: 0.3900 - val_loss: 1.5559 - val_accuracy: 0.4468\n",
            "Epoch 170/2000\n",
            "3/3 [==============================] - 0s 38ms/step - loss: 1.7246 - accuracy: 0.4200 - val_loss: 1.5285 - val_accuracy: 0.5106\n",
            "Epoch 171/2000\n",
            "3/3 [==============================] - 0s 30ms/step - loss: 1.7000 - accuracy: 0.4200 - val_loss: 1.5212 - val_accuracy: 0.5532\n",
            "Epoch 172/2000\n",
            "3/3 [==============================] - 0s 38ms/step - loss: 1.7524 - accuracy: 0.4400 - val_loss: 1.5197 - val_accuracy: 0.5319\n",
            "Epoch 173/2000\n",
            "3/3 [==============================] - 0s 31ms/step - loss: 1.7724 - accuracy: 0.3700 - val_loss: 1.5267 - val_accuracy: 0.5319\n",
            "Epoch 174/2000\n",
            "3/3 [==============================] - 0s 34ms/step - loss: 1.6796 - accuracy: 0.4400 - val_loss: 1.5374 - val_accuracy: 0.5106\n",
            "Epoch 175/2000\n",
            "3/3 [==============================] - 0s 48ms/step - loss: 1.7085 - accuracy: 0.4200 - val_loss: 1.5409 - val_accuracy: 0.5106\n",
            "Epoch 176/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 1.7351 - accuracy: 0.4100 - val_loss: 1.5518 - val_accuracy: 0.5319\n",
            "Epoch 177/2000\n",
            "3/3 [==============================] - 0s 39ms/step - loss: 1.6433 - accuracy: 0.4600 - val_loss: 1.5586 - val_accuracy: 0.5319\n",
            "Epoch 178/2000\n",
            "3/3 [==============================] - 0s 39ms/step - loss: 1.6511 - accuracy: 0.4500 - val_loss: 1.5554 - val_accuracy: 0.5319\n",
            "Epoch 179/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 1.6530 - accuracy: 0.4500 - val_loss: 1.5497 - val_accuracy: 0.5745\n",
            "Epoch 180/2000\n",
            "3/3 [==============================] - 0s 30ms/step - loss: 1.6530 - accuracy: 0.4800 - val_loss: 1.5640 - val_accuracy: 0.5319\n",
            "Epoch 181/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 1.7714 - accuracy: 0.3800 - val_loss: 1.5560 - val_accuracy: 0.5532\n",
            "Epoch 182/2000\n",
            "3/3 [==============================] - 0s 38ms/step - loss: 1.5962 - accuracy: 0.4800 - val_loss: 1.5466 - val_accuracy: 0.4894\n",
            "Epoch 183/2000\n",
            "3/3 [==============================] - 0s 39ms/step - loss: 1.6690 - accuracy: 0.4100 - val_loss: 1.5365 - val_accuracy: 0.5106\n",
            "Epoch 184/2000\n",
            "3/3 [==============================] - 0s 44ms/step - loss: 1.5977 - accuracy: 0.5100 - val_loss: 1.5326 - val_accuracy: 0.4681\n",
            "Epoch 185/2000\n",
            "3/3 [==============================] - 0s 40ms/step - loss: 1.6770 - accuracy: 0.4400 - val_loss: 1.5363 - val_accuracy: 0.4468\n",
            "Epoch 186/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 1.5489 - accuracy: 0.4400 - val_loss: 1.5199 - val_accuracy: 0.4894\n",
            "Epoch 187/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 1.6204 - accuracy: 0.4600 - val_loss: 1.5113 - val_accuracy: 0.5319\n",
            "Epoch 188/2000\n",
            "3/3 [==============================] - 0s 40ms/step - loss: 1.6306 - accuracy: 0.4200 - val_loss: 1.5058 - val_accuracy: 0.5106\n",
            "Epoch 189/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 1.5354 - accuracy: 0.5000 - val_loss: 1.4964 - val_accuracy: 0.5319\n",
            "Epoch 190/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 1.5211 - accuracy: 0.4700 - val_loss: 1.4838 - val_accuracy: 0.5319\n",
            "Epoch 191/2000\n",
            "3/3 [==============================] - 0s 38ms/step - loss: 1.6296 - accuracy: 0.4700 - val_loss: 1.4697 - val_accuracy: 0.5106\n",
            "Epoch 192/2000\n",
            "3/3 [==============================] - 0s 52ms/step - loss: 1.6833 - accuracy: 0.3500 - val_loss: 1.4727 - val_accuracy: 0.5319\n",
            "Epoch 193/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 1.5984 - accuracy: 0.5000 - val_loss: 1.4770 - val_accuracy: 0.5106\n",
            "Epoch 194/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 1.5940 - accuracy: 0.4600 - val_loss: 1.4725 - val_accuracy: 0.4894\n",
            "Epoch 195/2000\n",
            "3/3 [==============================] - 0s 31ms/step - loss: 1.6378 - accuracy: 0.4400 - val_loss: 1.4693 - val_accuracy: 0.4468\n",
            "Epoch 196/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 1.5158 - accuracy: 0.4900 - val_loss: 1.4644 - val_accuracy: 0.5106\n",
            "Epoch 197/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 1.5718 - accuracy: 0.4500 - val_loss: 1.4663 - val_accuracy: 0.5745\n",
            "Epoch 198/2000\n",
            "3/3 [==============================] - 0s 38ms/step - loss: 1.6547 - accuracy: 0.5300 - val_loss: 1.4586 - val_accuracy: 0.6170\n",
            "Epoch 199/2000\n",
            "3/3 [==============================] - 0s 38ms/step - loss: 1.4554 - accuracy: 0.5100 - val_loss: 1.4329 - val_accuracy: 0.6170\n",
            "Epoch 200/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 1.5353 - accuracy: 0.4900 - val_loss: 1.4313 - val_accuracy: 0.5745\n",
            "Epoch 201/2000\n",
            "3/3 [==============================] - 0s 38ms/step - loss: 1.5491 - accuracy: 0.4500 - val_loss: 1.4271 - val_accuracy: 0.5957\n",
            "Epoch 202/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 1.6086 - accuracy: 0.5000 - val_loss: 1.4412 - val_accuracy: 0.5745\n",
            "Epoch 203/2000\n",
            "3/3 [==============================] - 0s 31ms/step - loss: 1.5806 - accuracy: 0.4500 - val_loss: 1.5057 - val_accuracy: 0.4894\n",
            "Epoch 204/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 1.6269 - accuracy: 0.3900 - val_loss: 1.5557 - val_accuracy: 0.5319\n",
            "Epoch 205/2000\n",
            "3/3 [==============================] - 0s 45ms/step - loss: 1.5916 - accuracy: 0.4400 - val_loss: 1.4950 - val_accuracy: 0.5532\n",
            "Epoch 206/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 1.7369 - accuracy: 0.4400 - val_loss: 1.4268 - val_accuracy: 0.5319\n",
            "Epoch 207/2000\n",
            "3/3 [==============================] - 0s 34ms/step - loss: 1.4505 - accuracy: 0.4900 - val_loss: 1.3970 - val_accuracy: 0.5957\n",
            "Epoch 208/2000\n",
            "3/3 [==============================] - 0s 38ms/step - loss: 1.4550 - accuracy: 0.5000 - val_loss: 1.3901 - val_accuracy: 0.5957\n",
            "Epoch 209/2000\n",
            "3/3 [==============================] - 0s 38ms/step - loss: 1.5577 - accuracy: 0.4500 - val_loss: 1.3785 - val_accuracy: 0.5532\n",
            "Epoch 210/2000\n",
            "3/3 [==============================] - 0s 31ms/step - loss: 1.4993 - accuracy: 0.5000 - val_loss: 1.3671 - val_accuracy: 0.5957\n",
            "Epoch 211/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 1.4703 - accuracy: 0.5100 - val_loss: 1.3612 - val_accuracy: 0.6383\n",
            "Epoch 212/2000\n",
            "3/3 [==============================] - 0s 39ms/step - loss: 1.5946 - accuracy: 0.4500 - val_loss: 1.3655 - val_accuracy: 0.5957\n",
            "Epoch 213/2000\n",
            "3/3 [==============================] - 0s 38ms/step - loss: 1.5909 - accuracy: 0.4400 - val_loss: 1.3692 - val_accuracy: 0.5532\n",
            "Epoch 214/2000\n",
            "3/3 [==============================] - 0s 44ms/step - loss: 1.6579 - accuracy: 0.4700 - val_loss: 1.3867 - val_accuracy: 0.5532\n",
            "Epoch 215/2000\n",
            "3/3 [==============================] - 0s 32ms/step - loss: 1.6092 - accuracy: 0.4300 - val_loss: 1.3941 - val_accuracy: 0.5745\n",
            "Epoch 216/2000\n",
            "3/3 [==============================] - 0s 31ms/step - loss: 1.4893 - accuracy: 0.5100 - val_loss: 1.4039 - val_accuracy: 0.5532\n",
            "Epoch 217/2000\n",
            "3/3 [==============================] - 0s 45ms/step - loss: 1.5164 - accuracy: 0.4700 - val_loss: 1.4073 - val_accuracy: 0.5319\n",
            "Epoch 218/2000\n",
            "3/3 [==============================] - 0s 31ms/step - loss: 1.4433 - accuracy: 0.5100 - val_loss: 1.4126 - val_accuracy: 0.5745\n",
            "Epoch 219/2000\n",
            "3/3 [==============================] - 0s 44ms/step - loss: 1.5435 - accuracy: 0.4800 - val_loss: 1.4107 - val_accuracy: 0.5532\n",
            "Epoch 220/2000\n",
            "3/3 [==============================] - 0s 53ms/step - loss: 1.5858 - accuracy: 0.4500 - val_loss: 1.4035 - val_accuracy: 0.5532\n",
            "Epoch 221/2000\n",
            "3/3 [==============================] - 0s 49ms/step - loss: 1.4983 - accuracy: 0.5700 - val_loss: 1.3883 - val_accuracy: 0.5532\n",
            "Epoch 222/2000\n",
            "3/3 [==============================] - 0s 55ms/step - loss: 1.3221 - accuracy: 0.5700 - val_loss: 1.3642 - val_accuracy: 0.5532\n",
            "Epoch 223/2000\n",
            "3/3 [==============================] - 0s 51ms/step - loss: 1.4560 - accuracy: 0.5200 - val_loss: 1.3385 - val_accuracy: 0.5957\n",
            "Epoch 224/2000\n",
            "3/3 [==============================] - 0s 50ms/step - loss: 1.5589 - accuracy: 0.4900 - val_loss: 1.3093 - val_accuracy: 0.6809\n",
            "Epoch 225/2000\n",
            "3/3 [==============================] - 0s 52ms/step - loss: 1.3356 - accuracy: 0.5200 - val_loss: 1.3074 - val_accuracy: 0.6596\n",
            "Epoch 226/2000\n",
            "3/3 [==============================] - 0s 58ms/step - loss: 1.4850 - accuracy: 0.4400 - val_loss: 1.3106 - val_accuracy: 0.6596\n",
            "Epoch 227/2000\n",
            "3/3 [==============================] - 0s 63ms/step - loss: 1.4976 - accuracy: 0.4800 - val_loss: 1.2981 - val_accuracy: 0.6383\n",
            "Epoch 228/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 1.5017 - accuracy: 0.5200 - val_loss: 1.3030 - val_accuracy: 0.6170\n",
            "Epoch 229/2000\n",
            "3/3 [==============================] - 0s 39ms/step - loss: 1.5652 - accuracy: 0.4200 - val_loss: 1.3148 - val_accuracy: 0.6596\n",
            "Epoch 230/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 1.4847 - accuracy: 0.4900 - val_loss: 1.3325 - val_accuracy: 0.5957\n",
            "Epoch 231/2000\n",
            "3/3 [==============================] - 0s 39ms/step - loss: 1.4767 - accuracy: 0.4900 - val_loss: 1.3306 - val_accuracy: 0.5532\n",
            "Epoch 232/2000\n",
            "3/3 [==============================] - 0s 30ms/step - loss: 1.4477 - accuracy: 0.5300 - val_loss: 1.3265 - val_accuracy: 0.5745\n",
            "Epoch 233/2000\n",
            "3/3 [==============================] - 0s 33ms/step - loss: 1.5192 - accuracy: 0.5600 - val_loss: 1.3315 - val_accuracy: 0.6170\n",
            "Epoch 234/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 1.5136 - accuracy: 0.5000 - val_loss: 1.3561 - val_accuracy: 0.5957\n",
            "Epoch 235/2000\n",
            "3/3 [==============================] - 0s 29ms/step - loss: 1.5440 - accuracy: 0.5100 - val_loss: 1.3641 - val_accuracy: 0.6170\n",
            "Epoch 236/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 1.4297 - accuracy: 0.4600 - val_loss: 1.3716 - val_accuracy: 0.5745\n",
            "Epoch 237/2000\n",
            "3/3 [==============================] - 0s 38ms/step - loss: 1.3601 - accuracy: 0.6200 - val_loss: 1.3455 - val_accuracy: 0.6596\n",
            "Epoch 238/2000\n",
            "3/3 [==============================] - 0s 29ms/step - loss: 1.3328 - accuracy: 0.6300 - val_loss: 1.3137 - val_accuracy: 0.7234\n",
            "Epoch 239/2000\n",
            "3/3 [==============================] - 0s 32ms/step - loss: 1.4397 - accuracy: 0.5400 - val_loss: 1.3166 - val_accuracy: 0.6383\n",
            "Epoch 240/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 1.5067 - accuracy: 0.5700 - val_loss: 1.3267 - val_accuracy: 0.5957\n",
            "Epoch 241/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 1.5176 - accuracy: 0.5100 - val_loss: 1.3267 - val_accuracy: 0.5957\n",
            "Epoch 242/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 1.4434 - accuracy: 0.5200 - val_loss: 1.3115 - val_accuracy: 0.6596\n",
            "Epoch 243/2000\n",
            "3/3 [==============================] - 0s 30ms/step - loss: 1.4464 - accuracy: 0.5200 - val_loss: 1.3138 - val_accuracy: 0.6383\n",
            "Epoch 244/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 1.3765 - accuracy: 0.5600 - val_loss: 1.3095 - val_accuracy: 0.5957\n",
            "Epoch 245/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 1.3016 - accuracy: 0.6500 - val_loss: 1.3213 - val_accuracy: 0.6596\n",
            "Epoch 246/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 1.3672 - accuracy: 0.5900 - val_loss: 1.3146 - val_accuracy: 0.6596\n",
            "Epoch 247/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 1.4457 - accuracy: 0.5700 - val_loss: 1.3134 - val_accuracy: 0.6596\n",
            "Epoch 248/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 1.4187 - accuracy: 0.5600 - val_loss: 1.2936 - val_accuracy: 0.6809\n",
            "Epoch 249/2000\n",
            "3/3 [==============================] - 0s 38ms/step - loss: 1.3990 - accuracy: 0.6000 - val_loss: 1.2962 - val_accuracy: 0.6596\n",
            "Epoch 250/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 1.6072 - accuracy: 0.5100 - val_loss: 1.2828 - val_accuracy: 0.6596\n",
            "Epoch 251/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 1.4429 - accuracy: 0.5900 - val_loss: 1.2899 - val_accuracy: 0.6383\n",
            "Epoch 252/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 1.5600 - accuracy: 0.5100 - val_loss: 1.2960 - val_accuracy: 0.6170\n",
            "Epoch 253/2000\n",
            "3/3 [==============================] - 0s 46ms/step - loss: 1.3014 - accuracy: 0.6200 - val_loss: 1.2865 - val_accuracy: 0.6170\n",
            "Epoch 254/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 1.4473 - accuracy: 0.5500 - val_loss: 1.2918 - val_accuracy: 0.6383\n",
            "Epoch 255/2000\n",
            "3/3 [==============================] - 0s 38ms/step - loss: 1.3499 - accuracy: 0.5600 - val_loss: 1.2790 - val_accuracy: 0.6596\n",
            "Epoch 256/2000\n",
            "3/3 [==============================] - 0s 39ms/step - loss: 1.3329 - accuracy: 0.5800 - val_loss: 1.2639 - val_accuracy: 0.7021\n",
            "Epoch 257/2000\n",
            "3/3 [==============================] - 0s 31ms/step - loss: 1.3142 - accuracy: 0.5100 - val_loss: 1.2573 - val_accuracy: 0.6383\n",
            "Epoch 258/2000\n",
            "3/3 [==============================] - 0s 46ms/step - loss: 1.5099 - accuracy: 0.5100 - val_loss: 1.2498 - val_accuracy: 0.6596\n",
            "Epoch 259/2000\n",
            "3/3 [==============================] - 0s 55ms/step - loss: 1.4524 - accuracy: 0.5300 - val_loss: 1.2392 - val_accuracy: 0.6809\n",
            "Epoch 260/2000\n",
            "3/3 [==============================] - 0s 45ms/step - loss: 1.4562 - accuracy: 0.4600 - val_loss: 1.2292 - val_accuracy: 0.7447\n",
            "Epoch 261/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 1.3615 - accuracy: 0.5600 - val_loss: 1.2298 - val_accuracy: 0.6809\n",
            "Epoch 262/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 1.3155 - accuracy: 0.4900 - val_loss: 1.2467 - val_accuracy: 0.6170\n",
            "Epoch 263/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 1.4558 - accuracy: 0.5400 - val_loss: 1.2397 - val_accuracy: 0.6170\n",
            "Epoch 264/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 1.2767 - accuracy: 0.6100 - val_loss: 1.2046 - val_accuracy: 0.6383\n",
            "Epoch 265/2000\n",
            "3/3 [==============================] - 0s 39ms/step - loss: 1.4380 - accuracy: 0.5800 - val_loss: 1.2224 - val_accuracy: 0.6383\n",
            "Epoch 266/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 1.4716 - accuracy: 0.5000 - val_loss: 1.2383 - val_accuracy: 0.6170\n",
            "Epoch 267/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 1.2951 - accuracy: 0.5700 - val_loss: 1.2069 - val_accuracy: 0.6383\n",
            "Epoch 268/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 1.3309 - accuracy: 0.5800 - val_loss: 1.1932 - val_accuracy: 0.7021\n",
            "Epoch 269/2000\n",
            "3/3 [==============================] - 0s 34ms/step - loss: 1.4726 - accuracy: 0.5600 - val_loss: 1.2048 - val_accuracy: 0.7021\n",
            "Epoch 270/2000\n",
            "3/3 [==============================] - 0s 33ms/step - loss: 1.3557 - accuracy: 0.5900 - val_loss: 1.2492 - val_accuracy: 0.6596\n",
            "Epoch 271/2000\n",
            "3/3 [==============================] - 0s 31ms/step - loss: 1.2219 - accuracy: 0.6700 - val_loss: 1.2651 - val_accuracy: 0.6170\n",
            "Epoch 272/2000\n",
            "3/3 [==============================] - 0s 32ms/step - loss: 1.3236 - accuracy: 0.6100 - val_loss: 1.2614 - val_accuracy: 0.5957\n",
            "Epoch 273/2000\n",
            "3/3 [==============================] - 0s 31ms/step - loss: 1.2927 - accuracy: 0.5200 - val_loss: 1.2328 - val_accuracy: 0.5957\n",
            "Epoch 274/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 1.2748 - accuracy: 0.6000 - val_loss: 1.2221 - val_accuracy: 0.6170\n",
            "Epoch 275/2000\n",
            "3/3 [==============================] - 0s 38ms/step - loss: 1.2060 - accuracy: 0.6400 - val_loss: 1.1996 - val_accuracy: 0.6596\n",
            "Epoch 276/2000\n",
            "3/3 [==============================] - 0s 40ms/step - loss: 1.2500 - accuracy: 0.5500 - val_loss: 1.1952 - val_accuracy: 0.6596\n",
            "Epoch 277/2000\n",
            "3/3 [==============================] - 0s 34ms/step - loss: 1.2106 - accuracy: 0.6300 - val_loss: 1.2080 - val_accuracy: 0.6383\n",
            "Epoch 278/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 1.3156 - accuracy: 0.5800 - val_loss: 1.2007 - val_accuracy: 0.6383\n",
            "Epoch 279/2000\n",
            "3/3 [==============================] - 0s 31ms/step - loss: 1.3242 - accuracy: 0.6200 - val_loss: 1.1681 - val_accuracy: 0.6809\n",
            "Epoch 280/2000\n",
            "3/3 [==============================] - 0s 31ms/step - loss: 1.2818 - accuracy: 0.6000 - val_loss: 1.1861 - val_accuracy: 0.7021\n",
            "Epoch 281/2000\n",
            "3/3 [==============================] - 0s 39ms/step - loss: 1.3164 - accuracy: 0.5600 - val_loss: 1.1967 - val_accuracy: 0.6809\n",
            "Epoch 282/2000\n",
            "3/3 [==============================] - 0s 45ms/step - loss: 1.4172 - accuracy: 0.6100 - val_loss: 1.2000 - val_accuracy: 0.6596\n",
            "Epoch 283/2000\n",
            "3/3 [==============================] - 0s 39ms/step - loss: 1.2907 - accuracy: 0.5200 - val_loss: 1.1924 - val_accuracy: 0.6809\n",
            "Epoch 284/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 1.2286 - accuracy: 0.5700 - val_loss: 1.1867 - val_accuracy: 0.6596\n",
            "Epoch 285/2000\n",
            "3/3 [==============================] - 0s 32ms/step - loss: 1.1999 - accuracy: 0.5900 - val_loss: 1.1756 - val_accuracy: 0.5957\n",
            "Epoch 286/2000\n",
            "3/3 [==============================] - 0s 30ms/step - loss: 1.2959 - accuracy: 0.6400 - val_loss: 1.1619 - val_accuracy: 0.6596\n",
            "Epoch 287/2000\n",
            "3/3 [==============================] - 0s 38ms/step - loss: 1.2599 - accuracy: 0.5900 - val_loss: 1.1377 - val_accuracy: 0.7234\n",
            "Epoch 288/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 1.2476 - accuracy: 0.6300 - val_loss: 1.1373 - val_accuracy: 0.7021\n",
            "Epoch 289/2000\n",
            "3/3 [==============================] - 0s 30ms/step - loss: 1.2497 - accuracy: 0.6400 - val_loss: 1.1429 - val_accuracy: 0.7021\n",
            "Epoch 290/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 1.3103 - accuracy: 0.5700 - val_loss: 1.1328 - val_accuracy: 0.7234\n",
            "Epoch 291/2000\n",
            "3/3 [==============================] - 0s 39ms/step - loss: 1.3428 - accuracy: 0.6000 - val_loss: 1.1153 - val_accuracy: 0.7234\n",
            "Epoch 292/2000\n",
            "3/3 [==============================] - 0s 32ms/step - loss: 1.3247 - accuracy: 0.5800 - val_loss: 1.1115 - val_accuracy: 0.7021\n",
            "Epoch 293/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 1.2238 - accuracy: 0.6400 - val_loss: 1.1636 - val_accuracy: 0.7021\n",
            "Epoch 294/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 1.2823 - accuracy: 0.5800 - val_loss: 1.2690 - val_accuracy: 0.6170\n",
            "Epoch 295/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 1.3581 - accuracy: 0.5400 - val_loss: 1.2473 - val_accuracy: 0.6383\n",
            "Epoch 296/2000\n",
            "3/3 [==============================] - 0s 31ms/step - loss: 1.1547 - accuracy: 0.6500 - val_loss: 1.2030 - val_accuracy: 0.5957\n",
            "Epoch 297/2000\n",
            "3/3 [==============================] - 0s 45ms/step - loss: 1.1738 - accuracy: 0.6400 - val_loss: 1.1927 - val_accuracy: 0.6170\n",
            "Epoch 298/2000\n",
            "3/3 [==============================] - 0s 40ms/step - loss: 1.1916 - accuracy: 0.6200 - val_loss: 1.2015 - val_accuracy: 0.6383\n",
            "Epoch 299/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 1.1207 - accuracy: 0.6700 - val_loss: 1.2046 - val_accuracy: 0.6170\n",
            "Epoch 300/2000\n",
            "3/3 [==============================] - 0s 45ms/step - loss: 1.1444 - accuracy: 0.6600 - val_loss: 1.2146 - val_accuracy: 0.7234\n",
            "Epoch 301/2000\n",
            "3/3 [==============================] - 0s 49ms/step - loss: 1.2495 - accuracy: 0.6400 - val_loss: 1.2253 - val_accuracy: 0.6809\n",
            "Epoch 302/2000\n",
            "3/3 [==============================] - 0s 49ms/step - loss: 1.4073 - accuracy: 0.5600 - val_loss: 1.2164 - val_accuracy: 0.7660\n",
            "Epoch 303/2000\n",
            "3/3 [==============================] - 0s 56ms/step - loss: 1.2609 - accuracy: 0.6900 - val_loss: 1.1784 - val_accuracy: 0.7660\n",
            "Epoch 304/2000\n",
            "3/3 [==============================] - 0s 48ms/step - loss: 1.2746 - accuracy: 0.5900 - val_loss: 1.1567 - val_accuracy: 0.7447\n",
            "Epoch 305/2000\n",
            "3/3 [==============================] - 0s 53ms/step - loss: 1.2441 - accuracy: 0.6100 - val_loss: 1.1751 - val_accuracy: 0.7021\n",
            "Epoch 306/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 1.2911 - accuracy: 0.6000 - val_loss: 1.1855 - val_accuracy: 0.6809\n",
            "Epoch 307/2000\n",
            "3/3 [==============================] - 0s 44ms/step - loss: 1.2309 - accuracy: 0.5600 - val_loss: 1.1703 - val_accuracy: 0.7234\n",
            "Epoch 308/2000\n",
            "3/3 [==============================] - 0s 59ms/step - loss: 1.1036 - accuracy: 0.6800 - val_loss: 1.1494 - val_accuracy: 0.7234\n",
            "Epoch 309/2000\n",
            "3/3 [==============================] - 0s 58ms/step - loss: 1.2017 - accuracy: 0.6000 - val_loss: 1.1482 - val_accuracy: 0.7021\n",
            "Epoch 310/2000\n",
            "3/3 [==============================] - 0s 58ms/step - loss: 1.3237 - accuracy: 0.5700 - val_loss: 1.1538 - val_accuracy: 0.7234\n",
            "Epoch 311/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 1.2907 - accuracy: 0.5800 - val_loss: 1.1722 - val_accuracy: 0.7234\n",
            "Epoch 312/2000\n",
            "3/3 [==============================] - 0s 40ms/step - loss: 1.0889 - accuracy: 0.6700 - val_loss: 1.1727 - val_accuracy: 0.7234\n",
            "Epoch 313/2000\n",
            "3/3 [==============================] - 0s 40ms/step - loss: 1.1552 - accuracy: 0.6100 - val_loss: 1.1581 - val_accuracy: 0.7021\n",
            "Epoch 314/2000\n",
            "3/3 [==============================] - 0s 30ms/step - loss: 1.2680 - accuracy: 0.6100 - val_loss: 1.1444 - val_accuracy: 0.7021\n",
            "Epoch 315/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 1.2228 - accuracy: 0.6000 - val_loss: 1.1441 - val_accuracy: 0.6809\n",
            "Epoch 316/2000\n",
            "3/3 [==============================] - 0s 31ms/step - loss: 1.2321 - accuracy: 0.5700 - val_loss: 1.1602 - val_accuracy: 0.6809\n",
            "Epoch 317/2000\n",
            "3/3 [==============================] - 0s 33ms/step - loss: 1.1194 - accuracy: 0.6500 - val_loss: 1.1667 - val_accuracy: 0.6383\n",
            "Epoch 318/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 1.1305 - accuracy: 0.6400 - val_loss: 1.1616 - val_accuracy: 0.7021\n",
            "Epoch 319/2000\n",
            "3/3 [==============================] - 0s 32ms/step - loss: 1.1171 - accuracy: 0.6800 - val_loss: 1.1433 - val_accuracy: 0.7234\n",
            "Epoch 320/2000\n",
            "3/3 [==============================] - 0s 51ms/step - loss: 1.2384 - accuracy: 0.5900 - val_loss: 1.1338 - val_accuracy: 0.7447\n",
            "Epoch 321/2000\n",
            "3/3 [==============================] - 0s 33ms/step - loss: 1.2968 - accuracy: 0.6000 - val_loss: 1.1214 - val_accuracy: 0.7660\n",
            "Epoch 322/2000\n",
            "3/3 [==============================] - 0s 32ms/step - loss: 1.1752 - accuracy: 0.5900 - val_loss: 1.1192 - val_accuracy: 0.7234\n",
            "Epoch 323/2000\n",
            "3/3 [==============================] - 0s 32ms/step - loss: 1.3256 - accuracy: 0.6100 - val_loss: 1.1170 - val_accuracy: 0.6809\n",
            "Epoch 324/2000\n",
            "3/3 [==============================] - 0s 40ms/step - loss: 1.2278 - accuracy: 0.6000 - val_loss: 1.0980 - val_accuracy: 0.7234\n",
            "Epoch 325/2000\n",
            "3/3 [==============================] - 0s 32ms/step - loss: 1.1288 - accuracy: 0.6500 - val_loss: 1.0861 - val_accuracy: 0.7234\n",
            "Epoch 326/2000\n",
            "3/3 [==============================] - 0s 50ms/step - loss: 1.1831 - accuracy: 0.6100 - val_loss: 1.0820 - val_accuracy: 0.6809\n",
            "Epoch 327/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 1.1902 - accuracy: 0.6300 - val_loss: 1.0493 - val_accuracy: 0.7021\n",
            "Epoch 328/2000\n",
            "3/3 [==============================] - 0s 44ms/step - loss: 1.2398 - accuracy: 0.5900 - val_loss: 1.0505 - val_accuracy: 0.7021\n",
            "Epoch 329/2000\n",
            "3/3 [==============================] - 0s 49ms/step - loss: 1.0566 - accuracy: 0.6800 - val_loss: 1.0360 - val_accuracy: 0.7447\n",
            "Epoch 330/2000\n",
            "3/3 [==============================] - 0s 30ms/step - loss: 1.4154 - accuracy: 0.6000 - val_loss: 1.0392 - val_accuracy: 0.7660\n",
            "Epoch 331/2000\n",
            "3/3 [==============================] - 0s 30ms/step - loss: 1.3222 - accuracy: 0.6100 - val_loss: 1.0447 - val_accuracy: 0.7447\n",
            "Epoch 332/2000\n",
            "3/3 [==============================] - 0s 46ms/step - loss: 1.3257 - accuracy: 0.5600 - val_loss: 1.0455 - val_accuracy: 0.7447\n",
            "Epoch 333/2000\n",
            "3/3 [==============================] - 0s 31ms/step - loss: 1.1445 - accuracy: 0.6600 - val_loss: 1.0412 - val_accuracy: 0.7447\n",
            "Epoch 334/2000\n",
            "3/3 [==============================] - 0s 39ms/step - loss: 1.2003 - accuracy: 0.6300 - val_loss: 1.0620 - val_accuracy: 0.7234\n",
            "Epoch 335/2000\n",
            "3/3 [==============================] - 0s 31ms/step - loss: 1.0926 - accuracy: 0.7300 - val_loss: 1.1093 - val_accuracy: 0.7021\n",
            "Epoch 336/2000\n",
            "3/3 [==============================] - 0s 33ms/step - loss: 1.2019 - accuracy: 0.6500 - val_loss: 1.1325 - val_accuracy: 0.6809\n",
            "Epoch 337/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 1.1657 - accuracy: 0.6200 - val_loss: 1.1175 - val_accuracy: 0.7021\n",
            "Epoch 338/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 1.0785 - accuracy: 0.6300 - val_loss: 1.0913 - val_accuracy: 0.6809\n",
            "Epoch 339/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 1.0947 - accuracy: 0.6400 - val_loss: 1.1100 - val_accuracy: 0.7021\n",
            "Epoch 340/2000\n",
            "3/3 [==============================] - 0s 50ms/step - loss: 1.0217 - accuracy: 0.6600 - val_loss: 1.0985 - val_accuracy: 0.7234\n",
            "Epoch 341/2000\n",
            "3/3 [==============================] - 0s 33ms/step - loss: 1.2172 - accuracy: 0.6700 - val_loss: 1.0841 - val_accuracy: 0.7447\n",
            "Epoch 342/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 1.0889 - accuracy: 0.6100 - val_loss: 1.0686 - val_accuracy: 0.7234\n",
            "Epoch 343/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 1.3327 - accuracy: 0.6000 - val_loss: 1.0657 - val_accuracy: 0.6809\n",
            "Epoch 344/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 1.1790 - accuracy: 0.5900 - val_loss: 1.0561 - val_accuracy: 0.7021\n",
            "Epoch 345/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 1.1164 - accuracy: 0.5900 - val_loss: 1.0511 - val_accuracy: 0.7447\n",
            "Epoch 346/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 1.0033 - accuracy: 0.6700 - val_loss: 1.0680 - val_accuracy: 0.7021\n",
            "Epoch 347/2000\n",
            "3/3 [==============================] - 0s 52ms/step - loss: 1.1833 - accuracy: 0.6200 - val_loss: 1.0693 - val_accuracy: 0.7447\n",
            "Epoch 348/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 1.3402 - accuracy: 0.5200 - val_loss: 1.0862 - val_accuracy: 0.7234\n",
            "Epoch 349/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 1.0813 - accuracy: 0.6800 - val_loss: 1.0767 - val_accuracy: 0.7021\n",
            "Epoch 350/2000\n",
            "3/3 [==============================] - 0s 34ms/step - loss: 1.0728 - accuracy: 0.7200 - val_loss: 1.0827 - val_accuracy: 0.6809\n",
            "Epoch 351/2000\n",
            "3/3 [==============================] - 0s 32ms/step - loss: 1.1211 - accuracy: 0.6400 - val_loss: 1.0764 - val_accuracy: 0.6596\n",
            "Epoch 352/2000\n",
            "3/3 [==============================] - 0s 40ms/step - loss: 1.0615 - accuracy: 0.6500 - val_loss: 1.0685 - val_accuracy: 0.6809\n",
            "Epoch 353/2000\n",
            "3/3 [==============================] - 0s 38ms/step - loss: 1.1058 - accuracy: 0.6400 - val_loss: 1.0733 - val_accuracy: 0.6596\n",
            "Epoch 354/2000\n",
            "3/3 [==============================] - 0s 30ms/step - loss: 1.0977 - accuracy: 0.6400 - val_loss: 1.0676 - val_accuracy: 0.7234\n",
            "Epoch 355/2000\n",
            "3/3 [==============================] - 0s 30ms/step - loss: 1.1905 - accuracy: 0.5500 - val_loss: 1.0613 - val_accuracy: 0.7447\n",
            "Epoch 356/2000\n",
            "3/3 [==============================] - 0s 31ms/step - loss: 1.1433 - accuracy: 0.6100 - val_loss: 1.0585 - val_accuracy: 0.7447\n",
            "Epoch 357/2000\n",
            "3/3 [==============================] - 0s 31ms/step - loss: 1.1452 - accuracy: 0.6300 - val_loss: 1.0562 - val_accuracy: 0.7447\n",
            "Epoch 358/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 1.1123 - accuracy: 0.6200 - val_loss: 1.0465 - val_accuracy: 0.7447\n",
            "Epoch 359/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 1.1267 - accuracy: 0.7200 - val_loss: 1.0499 - val_accuracy: 0.7234\n",
            "Epoch 360/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 1.1908 - accuracy: 0.5800 - val_loss: 1.0595 - val_accuracy: 0.7234\n",
            "Epoch 361/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 1.2318 - accuracy: 0.6200 - val_loss: 1.0855 - val_accuracy: 0.7447\n",
            "Epoch 362/2000\n",
            "3/3 [==============================] - 0s 45ms/step - loss: 1.1017 - accuracy: 0.6900 - val_loss: 1.1155 - val_accuracy: 0.7021\n",
            "Epoch 363/2000\n",
            "3/3 [==============================] - 0s 34ms/step - loss: 1.0461 - accuracy: 0.7000 - val_loss: 1.0782 - val_accuracy: 0.6809\n",
            "Epoch 364/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 1.1807 - accuracy: 0.6500 - val_loss: 1.0534 - val_accuracy: 0.7234\n",
            "Epoch 365/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 1.2382 - accuracy: 0.6100 - val_loss: 1.0342 - val_accuracy: 0.7234\n",
            "Epoch 366/2000\n",
            "3/3 [==============================] - 0s 47ms/step - loss: 1.2760 - accuracy: 0.6300 - val_loss: 1.0713 - val_accuracy: 0.7234\n",
            "Epoch 367/2000\n",
            "3/3 [==============================] - 0s 40ms/step - loss: 1.1587 - accuracy: 0.6400 - val_loss: 1.1045 - val_accuracy: 0.6809\n",
            "Epoch 368/2000\n",
            "3/3 [==============================] - 0s 31ms/step - loss: 1.0380 - accuracy: 0.6600 - val_loss: 1.1140 - val_accuracy: 0.6809\n",
            "Epoch 369/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 1.0991 - accuracy: 0.6100 - val_loss: 1.0710 - val_accuracy: 0.7234\n",
            "Epoch 370/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 0.9617 - accuracy: 0.6700 - val_loss: 1.0295 - val_accuracy: 0.7234\n",
            "Epoch 371/2000\n",
            "3/3 [==============================] - 0s 38ms/step - loss: 1.0373 - accuracy: 0.6300 - val_loss: 1.0146 - val_accuracy: 0.6809\n",
            "Epoch 372/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 1.0739 - accuracy: 0.6600 - val_loss: 1.0110 - val_accuracy: 0.6383\n",
            "Epoch 373/2000\n",
            "3/3 [==============================] - 0s 34ms/step - loss: 1.0675 - accuracy: 0.6500 - val_loss: 0.9990 - val_accuracy: 0.6809\n",
            "Epoch 374/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 1.0658 - accuracy: 0.6600 - val_loss: 1.0051 - val_accuracy: 0.6809\n",
            "Epoch 375/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 1.1204 - accuracy: 0.6400 - val_loss: 1.0102 - val_accuracy: 0.7234\n",
            "Epoch 376/2000\n",
            "3/3 [==============================] - 0s 57ms/step - loss: 1.1544 - accuracy: 0.6300 - val_loss: 1.0149 - val_accuracy: 0.7447\n",
            "Epoch 377/2000\n",
            "3/3 [==============================] - 0s 39ms/step - loss: 1.1009 - accuracy: 0.5800 - val_loss: 0.9774 - val_accuracy: 0.7872\n",
            "Epoch 378/2000\n",
            "3/3 [==============================] - 0s 46ms/step - loss: 1.1735 - accuracy: 0.5900 - val_loss: 0.9571 - val_accuracy: 0.7660\n",
            "Epoch 379/2000\n",
            "3/3 [==============================] - 0s 60ms/step - loss: 1.0911 - accuracy: 0.5900 - val_loss: 0.9387 - val_accuracy: 0.7660\n",
            "Epoch 380/2000\n",
            "3/3 [==============================] - 0s 59ms/step - loss: 1.1382 - accuracy: 0.6500 - val_loss: 0.9289 - val_accuracy: 0.7447\n",
            "Epoch 381/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 1.0516 - accuracy: 0.6600 - val_loss: 0.9289 - val_accuracy: 0.7447\n",
            "Epoch 382/2000\n",
            "3/3 [==============================] - 0s 61ms/step - loss: 1.0701 - accuracy: 0.5900 - val_loss: 0.9347 - val_accuracy: 0.7660\n",
            "Epoch 383/2000\n",
            "3/3 [==============================] - 0s 40ms/step - loss: 1.1239 - accuracy: 0.7300 - val_loss: 0.9493 - val_accuracy: 0.7447\n",
            "Epoch 384/2000\n",
            "3/3 [==============================] - 0s 64ms/step - loss: 1.0715 - accuracy: 0.6700 - val_loss: 0.9592 - val_accuracy: 0.7660\n",
            "Epoch 385/2000\n",
            "3/3 [==============================] - 0s 60ms/step - loss: 1.0624 - accuracy: 0.6400 - val_loss: 0.9643 - val_accuracy: 0.7872\n",
            "Epoch 386/2000\n",
            "3/3 [==============================] - 0s 63ms/step - loss: 1.1103 - accuracy: 0.6100 - val_loss: 0.9554 - val_accuracy: 0.7660\n",
            "Epoch 387/2000\n",
            "3/3 [==============================] - 0s 63ms/step - loss: 1.2843 - accuracy: 0.6800 - val_loss: 0.9403 - val_accuracy: 0.7447\n",
            "Epoch 388/2000\n",
            "3/3 [==============================] - 0s 58ms/step - loss: 0.9901 - accuracy: 0.6900 - val_loss: 0.9316 - val_accuracy: 0.7021\n",
            "Epoch 389/2000\n",
            "3/3 [==============================] - 0s 57ms/step - loss: 1.1016 - accuracy: 0.5900 - val_loss: 0.9334 - val_accuracy: 0.7234\n",
            "Epoch 390/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 1.1344 - accuracy: 0.6500 - val_loss: 0.9438 - val_accuracy: 0.7021\n",
            "Epoch 391/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 1.0096 - accuracy: 0.6200 - val_loss: 0.9660 - val_accuracy: 0.7021\n",
            "Epoch 392/2000\n",
            "3/3 [==============================] - 0s 39ms/step - loss: 1.0719 - accuracy: 0.6700 - val_loss: 0.9845 - val_accuracy: 0.7447\n",
            "Epoch 393/2000\n",
            "3/3 [==============================] - 0s 45ms/step - loss: 0.9898 - accuracy: 0.7000 - val_loss: 0.9715 - val_accuracy: 0.7660\n",
            "Epoch 394/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 0.9560 - accuracy: 0.7200 - val_loss: 0.9528 - val_accuracy: 0.8085\n",
            "Epoch 395/2000\n",
            "3/3 [==============================] - 0s 38ms/step - loss: 1.1137 - accuracy: 0.6500 - val_loss: 0.9414 - val_accuracy: 0.8085\n",
            "Epoch 396/2000\n",
            "3/3 [==============================] - 0s 47ms/step - loss: 1.0892 - accuracy: 0.6300 - val_loss: 0.9421 - val_accuracy: 0.8085\n",
            "Epoch 397/2000\n",
            "3/3 [==============================] - 0s 32ms/step - loss: 1.0557 - accuracy: 0.6500 - val_loss: 0.9327 - val_accuracy: 0.7660\n",
            "Epoch 398/2000\n",
            "3/3 [==============================] - 0s 44ms/step - loss: 1.0233 - accuracy: 0.6800 - val_loss: 0.9411 - val_accuracy: 0.7447\n",
            "Epoch 399/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 1.0711 - accuracy: 0.7000 - val_loss: 0.9698 - val_accuracy: 0.7660\n",
            "Epoch 400/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 1.1129 - accuracy: 0.6400 - val_loss: 0.9726 - val_accuracy: 0.7234\n",
            "Epoch 401/2000\n",
            "3/3 [==============================] - 0s 32ms/step - loss: 0.9232 - accuracy: 0.7100 - val_loss: 0.9588 - val_accuracy: 0.7872\n",
            "Epoch 402/2000\n",
            "3/3 [==============================] - 0s 45ms/step - loss: 0.9630 - accuracy: 0.6800 - val_loss: 0.9575 - val_accuracy: 0.7660\n",
            "Epoch 403/2000\n",
            "3/3 [==============================] - 0s 32ms/step - loss: 1.0940 - accuracy: 0.6300 - val_loss: 0.9582 - val_accuracy: 0.7660\n",
            "Epoch 404/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 1.0757 - accuracy: 0.6700 - val_loss: 0.9322 - val_accuracy: 0.7872\n",
            "Epoch 405/2000\n",
            "3/3 [==============================] - 0s 31ms/step - loss: 1.1126 - accuracy: 0.6200 - val_loss: 0.9172 - val_accuracy: 0.7872\n",
            "Epoch 406/2000\n",
            "3/3 [==============================] - 0s 45ms/step - loss: 1.0926 - accuracy: 0.6600 - val_loss: 0.9461 - val_accuracy: 0.7447\n",
            "Epoch 407/2000\n",
            "3/3 [==============================] - 0s 40ms/step - loss: 1.0405 - accuracy: 0.6800 - val_loss: 0.9787 - val_accuracy: 0.6809\n",
            "Epoch 408/2000\n",
            "3/3 [==============================] - 0s 30ms/step - loss: 1.0671 - accuracy: 0.7000 - val_loss: 1.0001 - val_accuracy: 0.7021\n",
            "Epoch 409/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 1.0520 - accuracy: 0.7300 - val_loss: 1.0121 - val_accuracy: 0.6596\n",
            "Epoch 410/2000\n",
            "3/3 [==============================] - 0s 34ms/step - loss: 1.0667 - accuracy: 0.6600 - val_loss: 0.9760 - val_accuracy: 0.7660\n",
            "Epoch 411/2000\n",
            "3/3 [==============================] - 0s 39ms/step - loss: 1.0274 - accuracy: 0.6500 - val_loss: 0.9391 - val_accuracy: 0.7447\n",
            "Epoch 412/2000\n",
            "3/3 [==============================] - 0s 39ms/step - loss: 1.0979 - accuracy: 0.6400 - val_loss: 0.9198 - val_accuracy: 0.7872\n",
            "Epoch 413/2000\n",
            "3/3 [==============================] - 0s 33ms/step - loss: 1.1249 - accuracy: 0.6000 - val_loss: 0.8990 - val_accuracy: 0.7872\n",
            "Epoch 414/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 0.9800 - accuracy: 0.7100 - val_loss: 0.8793 - val_accuracy: 0.7872\n",
            "Epoch 415/2000\n",
            "3/3 [==============================] - 0s 38ms/step - loss: 0.9963 - accuracy: 0.6900 - val_loss: 0.8758 - val_accuracy: 0.8085\n",
            "Epoch 416/2000\n",
            "3/3 [==============================] - 0s 33ms/step - loss: 1.1463 - accuracy: 0.6600 - val_loss: 0.8842 - val_accuracy: 0.7872\n",
            "Epoch 417/2000\n",
            "3/3 [==============================] - 0s 38ms/step - loss: 0.9361 - accuracy: 0.7300 - val_loss: 0.8714 - val_accuracy: 0.8085\n",
            "Epoch 418/2000\n",
            "3/3 [==============================] - 0s 38ms/step - loss: 0.9804 - accuracy: 0.6900 - val_loss: 0.8630 - val_accuracy: 0.8511\n",
            "Epoch 419/2000\n",
            "3/3 [==============================] - 0s 39ms/step - loss: 0.9640 - accuracy: 0.6600 - val_loss: 0.8540 - val_accuracy: 0.8085\n",
            "Epoch 420/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 1.0923 - accuracy: 0.7000 - val_loss: 0.8587 - val_accuracy: 0.8085\n",
            "Epoch 421/2000\n",
            "3/3 [==============================] - 0s 34ms/step - loss: 0.9828 - accuracy: 0.6600 - val_loss: 0.8593 - val_accuracy: 0.8298\n",
            "Epoch 422/2000\n",
            "3/3 [==============================] - 0s 33ms/step - loss: 0.8831 - accuracy: 0.7600 - val_loss: 0.8464 - val_accuracy: 0.8298\n",
            "Epoch 423/2000\n",
            "3/3 [==============================] - 0s 34ms/step - loss: 1.0655 - accuracy: 0.6700 - val_loss: 0.8416 - val_accuracy: 0.7872\n",
            "Epoch 424/2000\n",
            "3/3 [==============================] - 0s 38ms/step - loss: 0.9655 - accuracy: 0.7600 - val_loss: 0.8559 - val_accuracy: 0.8085\n",
            "Epoch 425/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 0.9128 - accuracy: 0.7300 - val_loss: 0.8572 - val_accuracy: 0.8085\n",
            "Epoch 426/2000\n",
            "3/3 [==============================] - 0s 46ms/step - loss: 1.0613 - accuracy: 0.7100 - val_loss: 0.8511 - val_accuracy: 0.7872\n",
            "Epoch 427/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 0.8878 - accuracy: 0.7800 - val_loss: 0.8540 - val_accuracy: 0.7660\n",
            "Epoch 428/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 1.0511 - accuracy: 0.6400 - val_loss: 0.8518 - val_accuracy: 0.7872\n",
            "Epoch 429/2000\n",
            "3/3 [==============================] - 0s 45ms/step - loss: 1.0749 - accuracy: 0.7000 - val_loss: 0.8587 - val_accuracy: 0.7872\n",
            "Epoch 430/2000\n",
            "3/3 [==============================] - 0s 31ms/step - loss: 0.9845 - accuracy: 0.7100 - val_loss: 0.8726 - val_accuracy: 0.7872\n",
            "Epoch 431/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 0.9742 - accuracy: 0.6900 - val_loss: 0.8556 - val_accuracy: 0.7872\n",
            "Epoch 432/2000\n",
            "3/3 [==============================] - 0s 38ms/step - loss: 0.9994 - accuracy: 0.6400 - val_loss: 0.8437 - val_accuracy: 0.7660\n",
            "Epoch 433/2000\n",
            "3/3 [==============================] - 0s 45ms/step - loss: 1.0007 - accuracy: 0.6700 - val_loss: 0.8541 - val_accuracy: 0.7660\n",
            "Epoch 434/2000\n",
            "3/3 [==============================] - 0s 30ms/step - loss: 1.0413 - accuracy: 0.7000 - val_loss: 0.8637 - val_accuracy: 0.7447\n",
            "Epoch 435/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 1.2332 - accuracy: 0.6100 - val_loss: 0.8684 - val_accuracy: 0.7447\n",
            "Epoch 436/2000\n",
            "3/3 [==============================] - 0s 32ms/step - loss: 1.0100 - accuracy: 0.6600 - val_loss: 0.8686 - val_accuracy: 0.7447\n",
            "Epoch 437/2000\n",
            "3/3 [==============================] - 0s 33ms/step - loss: 1.0264 - accuracy: 0.6700 - val_loss: 0.8704 - val_accuracy: 0.7660\n",
            "Epoch 438/2000\n",
            "3/3 [==============================] - 0s 40ms/step - loss: 0.9702 - accuracy: 0.7500 - val_loss: 0.8691 - val_accuracy: 0.7660\n",
            "Epoch 439/2000\n",
            "3/3 [==============================] - 0s 39ms/step - loss: 0.9908 - accuracy: 0.7200 - val_loss: 0.8794 - val_accuracy: 0.7660\n",
            "Epoch 440/2000\n",
            "3/3 [==============================] - 0s 53ms/step - loss: 1.1473 - accuracy: 0.7000 - val_loss: 0.9253 - val_accuracy: 0.7447\n",
            "Epoch 441/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 0.9737 - accuracy: 0.7000 - val_loss: 0.9089 - val_accuracy: 0.7234\n",
            "Epoch 442/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 0.9767 - accuracy: 0.6600 - val_loss: 0.8865 - val_accuracy: 0.7447\n",
            "Epoch 443/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 0.8939 - accuracy: 0.7300 - val_loss: 0.8648 - val_accuracy: 0.8085\n",
            "Epoch 444/2000\n",
            "3/3 [==============================] - 0s 39ms/step - loss: 1.0061 - accuracy: 0.6800 - val_loss: 0.8591 - val_accuracy: 0.8085\n",
            "Epoch 445/2000\n",
            "3/3 [==============================] - 0s 33ms/step - loss: 0.9211 - accuracy: 0.7700 - val_loss: 0.8534 - val_accuracy: 0.8085\n",
            "Epoch 446/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 0.9422 - accuracy: 0.6900 - val_loss: 0.8611 - val_accuracy: 0.7660\n",
            "Epoch 447/2000\n",
            "3/3 [==============================] - 0s 31ms/step - loss: 0.9027 - accuracy: 0.7400 - val_loss: 0.8701 - val_accuracy: 0.7872\n",
            "Epoch 448/2000\n",
            "3/3 [==============================] - 0s 38ms/step - loss: 1.0544 - accuracy: 0.7500 - val_loss: 0.8771 - val_accuracy: 0.7660\n",
            "Epoch 449/2000\n",
            "3/3 [==============================] - 0s 39ms/step - loss: 0.9594 - accuracy: 0.7300 - val_loss: 0.8778 - val_accuracy: 0.7660\n",
            "Epoch 450/2000\n",
            "3/3 [==============================] - 0s 38ms/step - loss: 0.9452 - accuracy: 0.7000 - val_loss: 0.8851 - val_accuracy: 0.7447\n",
            "Epoch 451/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 0.9577 - accuracy: 0.7000 - val_loss: 0.8923 - val_accuracy: 0.7447\n",
            "Epoch 452/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 1.1691 - accuracy: 0.6100 - val_loss: 0.9162 - val_accuracy: 0.7447\n",
            "Epoch 453/2000\n",
            "3/3 [==============================] - 0s 40ms/step - loss: 0.9510 - accuracy: 0.6800 - val_loss: 0.9575 - val_accuracy: 0.7660\n",
            "Epoch 454/2000\n",
            "3/3 [==============================] - 0s 31ms/step - loss: 1.0523 - accuracy: 0.6500 - val_loss: 1.0019 - val_accuracy: 0.7234\n",
            "Epoch 455/2000\n",
            "3/3 [==============================] - 0s 49ms/step - loss: 1.0079 - accuracy: 0.7100 - val_loss: 1.0515 - val_accuracy: 0.7021\n",
            "Epoch 456/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 1.0645 - accuracy: 0.6900 - val_loss: 1.0013 - val_accuracy: 0.7021\n",
            "Epoch 457/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 0.8386 - accuracy: 0.7500 - val_loss: 0.9198 - val_accuracy: 0.7447\n",
            "Epoch 458/2000\n",
            "3/3 [==============================] - 0s 45ms/step - loss: 0.9986 - accuracy: 0.6600 - val_loss: 0.8746 - val_accuracy: 0.7872\n",
            "Epoch 459/2000\n",
            "3/3 [==============================] - 0s 59ms/step - loss: 1.0782 - accuracy: 0.6700 - val_loss: 0.8529 - val_accuracy: 0.7660\n",
            "Epoch 460/2000\n",
            "3/3 [==============================] - 0s 54ms/step - loss: 1.1216 - accuracy: 0.6100 - val_loss: 0.8451 - val_accuracy: 0.7660\n",
            "Epoch 461/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 0.8716 - accuracy: 0.7400 - val_loss: 0.8389 - val_accuracy: 0.7660\n",
            "Epoch 462/2000\n",
            "3/3 [==============================] - 0s 61ms/step - loss: 0.9512 - accuracy: 0.6600 - val_loss: 0.8121 - val_accuracy: 0.7660\n",
            "Epoch 463/2000\n",
            "3/3 [==============================] - 0s 65ms/step - loss: 0.9542 - accuracy: 0.6500 - val_loss: 0.7946 - val_accuracy: 0.7872\n",
            "Epoch 464/2000\n",
            "3/3 [==============================] - 0s 56ms/step - loss: 1.0691 - accuracy: 0.7500 - val_loss: 0.8082 - val_accuracy: 0.7447\n",
            "Epoch 465/2000\n",
            "3/3 [==============================] - 0s 67ms/step - loss: 0.9501 - accuracy: 0.7100 - val_loss: 0.8458 - val_accuracy: 0.7447\n",
            "Epoch 466/2000\n",
            "3/3 [==============================] - 0s 49ms/step - loss: 1.0190 - accuracy: 0.6700 - val_loss: 0.8696 - val_accuracy: 0.7660\n",
            "Epoch 467/2000\n",
            "3/3 [==============================] - 0s 39ms/step - loss: 0.8796 - accuracy: 0.7000 - val_loss: 0.8663 - val_accuracy: 0.7447\n",
            "Epoch 468/2000\n",
            "3/3 [==============================] - 0s 32ms/step - loss: 1.1554 - accuracy: 0.7200 - val_loss: 0.8745 - val_accuracy: 0.7660\n",
            "Epoch 469/2000\n",
            "3/3 [==============================] - 0s 50ms/step - loss: 1.1607 - accuracy: 0.6600 - val_loss: 0.8412 - val_accuracy: 0.7660\n",
            "Epoch 470/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 1.1235 - accuracy: 0.6500 - val_loss: 0.8335 - val_accuracy: 0.7872\n",
            "Epoch 471/2000\n",
            "3/3 [==============================] - 0s 44ms/step - loss: 0.8978 - accuracy: 0.7400 - val_loss: 0.8093 - val_accuracy: 0.8085\n",
            "Epoch 472/2000\n",
            "3/3 [==============================] - 0s 53ms/step - loss: 0.9829 - accuracy: 0.7200 - val_loss: 0.8129 - val_accuracy: 0.7872\n",
            "Epoch 473/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 0.8564 - accuracy: 0.7500 - val_loss: 0.7995 - val_accuracy: 0.7660\n",
            "Epoch 474/2000\n",
            "3/3 [==============================] - 0s 40ms/step - loss: 1.2665 - accuracy: 0.6600 - val_loss: 0.7890 - val_accuracy: 0.7872\n",
            "Epoch 475/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 0.9522 - accuracy: 0.7200 - val_loss: 0.7994 - val_accuracy: 0.7872\n",
            "Epoch 476/2000\n",
            "3/3 [==============================] - 0s 33ms/step - loss: 1.0336 - accuracy: 0.6500 - val_loss: 0.8196 - val_accuracy: 0.8298\n",
            "Epoch 477/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 0.9732 - accuracy: 0.7000 - val_loss: 0.8266 - val_accuracy: 0.8085\n",
            "Epoch 478/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 0.9276 - accuracy: 0.6500 - val_loss: 0.7915 - val_accuracy: 0.8085\n",
            "Epoch 479/2000\n",
            "3/3 [==============================] - 0s 38ms/step - loss: 0.9200 - accuracy: 0.7200 - val_loss: 0.7726 - val_accuracy: 0.7872\n",
            "Epoch 480/2000\n",
            "3/3 [==============================] - 0s 29ms/step - loss: 0.8354 - accuracy: 0.8000 - val_loss: 0.7576 - val_accuracy: 0.7872\n",
            "Epoch 481/2000\n",
            "3/3 [==============================] - 0s 38ms/step - loss: 0.9423 - accuracy: 0.7500 - val_loss: 0.7513 - val_accuracy: 0.7872\n",
            "Epoch 482/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 0.9747 - accuracy: 0.7400 - val_loss: 0.7578 - val_accuracy: 0.7660\n",
            "Epoch 483/2000\n",
            "3/3 [==============================] - 0s 32ms/step - loss: 0.9214 - accuracy: 0.7300 - val_loss: 0.7724 - val_accuracy: 0.8085\n",
            "Epoch 484/2000\n",
            "3/3 [==============================] - 0s 38ms/step - loss: 0.9334 - accuracy: 0.6900 - val_loss: 0.7784 - val_accuracy: 0.8085\n",
            "Epoch 485/2000\n",
            "3/3 [==============================] - 0s 39ms/step - loss: 1.0042 - accuracy: 0.7000 - val_loss: 0.7827 - val_accuracy: 0.8085\n",
            "Epoch 486/2000\n",
            "3/3 [==============================] - 0s 39ms/step - loss: 1.0519 - accuracy: 0.6900 - val_loss: 0.7904 - val_accuracy: 0.7872\n",
            "Epoch 487/2000\n",
            "3/3 [==============================] - 0s 39ms/step - loss: 1.0912 - accuracy: 0.6600 - val_loss: 0.7906 - val_accuracy: 0.7872\n",
            "Epoch 488/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 0.9570 - accuracy: 0.6900 - val_loss: 0.8006 - val_accuracy: 0.7872\n",
            "Epoch 489/2000\n",
            "3/3 [==============================] - 0s 34ms/step - loss: 0.8920 - accuracy: 0.7000 - val_loss: 0.7874 - val_accuracy: 0.8085\n",
            "Epoch 490/2000\n",
            "3/3 [==============================] - 0s 50ms/step - loss: 0.9677 - accuracy: 0.6800 - val_loss: 0.7822 - val_accuracy: 0.8298\n",
            "Epoch 491/2000\n",
            "3/3 [==============================] - 0s 39ms/step - loss: 1.1162 - accuracy: 0.6600 - val_loss: 0.7946 - val_accuracy: 0.7872\n",
            "Epoch 492/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 0.9310 - accuracy: 0.7400 - val_loss: 0.8140 - val_accuracy: 0.7447\n",
            "Epoch 493/2000\n",
            "3/3 [==============================] - 0s 40ms/step - loss: 0.9777 - accuracy: 0.7400 - val_loss: 0.8227 - val_accuracy: 0.7447\n",
            "Epoch 494/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 0.8828 - accuracy: 0.7900 - val_loss: 0.8302 - val_accuracy: 0.7447\n",
            "Epoch 495/2000\n",
            "3/3 [==============================] - 0s 32ms/step - loss: 0.9722 - accuracy: 0.6800 - val_loss: 0.8030 - val_accuracy: 0.7660\n",
            "Epoch 496/2000\n",
            "3/3 [==============================] - 0s 38ms/step - loss: 0.9606 - accuracy: 0.7100 - val_loss: 0.7811 - val_accuracy: 0.7872\n",
            "Epoch 497/2000\n",
            "3/3 [==============================] - 0s 44ms/step - loss: 0.9498 - accuracy: 0.7400 - val_loss: 0.7673 - val_accuracy: 0.8085\n",
            "Epoch 498/2000\n",
            "3/3 [==============================] - 0s 34ms/step - loss: 0.9608 - accuracy: 0.7400 - val_loss: 0.7676 - val_accuracy: 0.7872\n",
            "Epoch 499/2000\n",
            "3/3 [==============================] - 0s 40ms/step - loss: 0.9886 - accuracy: 0.6700 - val_loss: 0.7870 - val_accuracy: 0.7660\n",
            "Epoch 500/2000\n",
            "3/3 [==============================] - 0s 34ms/step - loss: 0.8827 - accuracy: 0.7100 - val_loss: 0.7832 - val_accuracy: 0.7872\n",
            "Epoch 501/2000\n",
            "3/3 [==============================] - 0s 33ms/step - loss: 0.8284 - accuracy: 0.7100 - val_loss: 0.7811 - val_accuracy: 0.7872\n",
            "Epoch 502/2000\n",
            "3/3 [==============================] - 0s 31ms/step - loss: 0.8949 - accuracy: 0.7400 - val_loss: 0.7742 - val_accuracy: 0.7660\n",
            "Epoch 503/2000\n",
            "3/3 [==============================] - 0s 40ms/step - loss: 0.7658 - accuracy: 0.7800 - val_loss: 0.7720 - val_accuracy: 0.7447\n",
            "Epoch 504/2000\n",
            "3/3 [==============================] - 0s 51ms/step - loss: 0.8579 - accuracy: 0.6900 - val_loss: 0.7735 - val_accuracy: 0.7872\n",
            "Epoch 505/2000\n",
            "3/3 [==============================] - 0s 51ms/step - loss: 0.9199 - accuracy: 0.6800 - val_loss: 0.7713 - val_accuracy: 0.7660\n",
            "Epoch 506/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 0.8163 - accuracy: 0.7200 - val_loss: 0.7771 - val_accuracy: 0.7660\n",
            "Epoch 507/2000\n",
            "3/3 [==============================] - 0s 38ms/step - loss: 0.9141 - accuracy: 0.7400 - val_loss: 0.7740 - val_accuracy: 0.7872\n",
            "Epoch 508/2000\n",
            "3/3 [==============================] - 0s 46ms/step - loss: 0.8386 - accuracy: 0.7200 - val_loss: 0.8302 - val_accuracy: 0.7447\n",
            "Epoch 509/2000\n",
            "3/3 [==============================] - 0s 48ms/step - loss: 0.9017 - accuracy: 0.7200 - val_loss: 0.8848 - val_accuracy: 0.7234\n",
            "Epoch 510/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 0.9446 - accuracy: 0.7000 - val_loss: 0.8724 - val_accuracy: 0.7447\n",
            "Epoch 511/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 0.9182 - accuracy: 0.7100 - val_loss: 0.8413 - val_accuracy: 0.7234\n",
            "Epoch 512/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 0.8294 - accuracy: 0.7100 - val_loss: 0.8323 - val_accuracy: 0.7447\n",
            "Epoch 513/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 0.9928 - accuracy: 0.6700 - val_loss: 0.8015 - val_accuracy: 0.7872\n",
            "Epoch 514/2000\n",
            "3/3 [==============================] - 0s 32ms/step - loss: 1.0244 - accuracy: 0.6600 - val_loss: 0.7880 - val_accuracy: 0.8085\n",
            "Epoch 515/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 0.9685 - accuracy: 0.6600 - val_loss: 0.7698 - val_accuracy: 0.8085\n",
            "Epoch 516/2000\n",
            "3/3 [==============================] - 0s 40ms/step - loss: 0.8282 - accuracy: 0.7300 - val_loss: 0.7919 - val_accuracy: 0.7660\n",
            "Epoch 517/2000\n",
            "3/3 [==============================] - 0s 34ms/step - loss: 0.8849 - accuracy: 0.6900 - val_loss: 0.7982 - val_accuracy: 0.7660\n",
            "Epoch 518/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 0.9393 - accuracy: 0.7000 - val_loss: 0.7767 - val_accuracy: 0.7447\n",
            "Epoch 519/2000\n",
            "3/3 [==============================] - 0s 31ms/step - loss: 0.9090 - accuracy: 0.6900 - val_loss: 0.7630 - val_accuracy: 0.7660\n",
            "Epoch 520/2000\n",
            "3/3 [==============================] - 0s 40ms/step - loss: 0.7674 - accuracy: 0.7600 - val_loss: 0.7929 - val_accuracy: 0.7021\n",
            "Epoch 521/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 0.9367 - accuracy: 0.7100 - val_loss: 0.8110 - val_accuracy: 0.7021\n",
            "Epoch 522/2000\n",
            "3/3 [==============================] - 0s 45ms/step - loss: 0.9689 - accuracy: 0.7100 - val_loss: 0.7831 - val_accuracy: 0.7021\n",
            "Epoch 523/2000\n",
            "3/3 [==============================] - 0s 32ms/step - loss: 0.8224 - accuracy: 0.7300 - val_loss: 0.7428 - val_accuracy: 0.7872\n",
            "Epoch 524/2000\n",
            "3/3 [==============================] - 0s 47ms/step - loss: 1.0563 - accuracy: 0.6900 - val_loss: 0.7566 - val_accuracy: 0.7872\n",
            "Epoch 525/2000\n",
            "3/3 [==============================] - 0s 33ms/step - loss: 1.0493 - accuracy: 0.6500 - val_loss: 0.7777 - val_accuracy: 0.7447\n",
            "Epoch 526/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 1.0489 - accuracy: 0.6700 - val_loss: 0.7731 - val_accuracy: 0.7447\n",
            "Epoch 527/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 0.7717 - accuracy: 0.7800 - val_loss: 0.7624 - val_accuracy: 0.7447\n",
            "Epoch 528/2000\n",
            "3/3 [==============================] - 0s 30ms/step - loss: 0.9308 - accuracy: 0.6500 - val_loss: 0.7586 - val_accuracy: 0.7447\n",
            "Epoch 529/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 0.8825 - accuracy: 0.7000 - val_loss: 0.7820 - val_accuracy: 0.7021\n",
            "Epoch 530/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 0.8972 - accuracy: 0.6500 - val_loss: 0.7950 - val_accuracy: 0.7234\n",
            "Epoch 531/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 0.8603 - accuracy: 0.7500 - val_loss: 0.8018 - val_accuracy: 0.7447\n",
            "Epoch 532/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 0.9138 - accuracy: 0.6900 - val_loss: 0.8039 - val_accuracy: 0.7872\n",
            "Epoch 533/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 0.7494 - accuracy: 0.7000 - val_loss: 0.8144 - val_accuracy: 0.7660\n",
            "Epoch 534/2000\n",
            "3/3 [==============================] - 0s 51ms/step - loss: 1.0418 - accuracy: 0.7100 - val_loss: 0.8286 - val_accuracy: 0.7660\n",
            "Epoch 535/2000\n",
            "3/3 [==============================] - 0s 55ms/step - loss: 0.9150 - accuracy: 0.7100 - val_loss: 0.8526 - val_accuracy: 0.7234\n",
            "Epoch 536/2000\n",
            "3/3 [==============================] - 0s 56ms/step - loss: 0.9081 - accuracy: 0.7100 - val_loss: 0.9033 - val_accuracy: 0.6596\n",
            "Epoch 537/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 0.9694 - accuracy: 0.6900 - val_loss: 0.9017 - val_accuracy: 0.6596\n",
            "Epoch 538/2000\n",
            "3/3 [==============================] - 0s 46ms/step - loss: 0.7624 - accuracy: 0.7400 - val_loss: 0.8693 - val_accuracy: 0.6596\n",
            "Epoch 539/2000\n",
            "3/3 [==============================] - 0s 54ms/step - loss: 0.9323 - accuracy: 0.6700 - val_loss: 0.8495 - val_accuracy: 0.6383\n",
            "Epoch 540/2000\n",
            "3/3 [==============================] - 0s 68ms/step - loss: 0.8637 - accuracy: 0.7200 - val_loss: 0.8479 - val_accuracy: 0.6383\n",
            "Epoch 541/2000\n",
            "3/3 [==============================] - 0s 63ms/step - loss: 0.8490 - accuracy: 0.7600 - val_loss: 0.8334 - val_accuracy: 0.6809\n",
            "Epoch 542/2000\n",
            "3/3 [==============================] - 0s 51ms/step - loss: 0.8930 - accuracy: 0.7500 - val_loss: 0.8177 - val_accuracy: 0.7234\n",
            "Epoch 543/2000\n",
            "3/3 [==============================] - 0s 56ms/step - loss: 0.7844 - accuracy: 0.8200 - val_loss: 0.8075 - val_accuracy: 0.7234\n",
            "Epoch 544/2000\n",
            "3/3 [==============================] - 0s 34ms/step - loss: 0.8595 - accuracy: 0.7400 - val_loss: 0.8079 - val_accuracy: 0.6809\n",
            "Epoch 545/2000\n",
            "3/3 [==============================] - 0s 39ms/step - loss: 0.7937 - accuracy: 0.7400 - val_loss: 0.8124 - val_accuracy: 0.7234\n",
            "Epoch 546/2000\n",
            "3/3 [==============================] - 0s 40ms/step - loss: 0.9588 - accuracy: 0.7000 - val_loss: 0.8126 - val_accuracy: 0.7021\n",
            "Epoch 547/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 0.8811 - accuracy: 0.7200 - val_loss: 0.8131 - val_accuracy: 0.7021\n",
            "Epoch 548/2000\n",
            "3/3 [==============================] - 0s 40ms/step - loss: 0.8345 - accuracy: 0.7300 - val_loss: 0.8077 - val_accuracy: 0.7447\n",
            "Epoch 549/2000\n",
            "3/3 [==============================] - 0s 30ms/step - loss: 0.7079 - accuracy: 0.7800 - val_loss: 0.8139 - val_accuracy: 0.7234\n",
            "Epoch 550/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 0.9764 - accuracy: 0.7000 - val_loss: 0.8028 - val_accuracy: 0.7234\n",
            "Epoch 551/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 0.8245 - accuracy: 0.7600 - val_loss: 0.7982 - val_accuracy: 0.7660\n",
            "Epoch 552/2000\n",
            "3/3 [==============================] - 0s 50ms/step - loss: 0.8759 - accuracy: 0.7500 - val_loss: 0.8132 - val_accuracy: 0.7234\n",
            "Epoch 553/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 0.7740 - accuracy: 0.7500 - val_loss: 0.8636 - val_accuracy: 0.7021\n",
            "Epoch 554/2000\n",
            "3/3 [==============================] - 0s 31ms/step - loss: 0.8741 - accuracy: 0.7600 - val_loss: 0.8904 - val_accuracy: 0.6596\n",
            "Epoch 555/2000\n",
            "3/3 [==============================] - 0s 32ms/step - loss: 0.7869 - accuracy: 0.7500 - val_loss: 0.8946 - val_accuracy: 0.7234\n",
            "Epoch 556/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 0.9021 - accuracy: 0.6900 - val_loss: 0.9146 - val_accuracy: 0.7234\n",
            "Epoch 557/2000\n",
            "3/3 [==============================] - 0s 38ms/step - loss: 0.8294 - accuracy: 0.7600 - val_loss: 0.9059 - val_accuracy: 0.7234\n",
            "Epoch 558/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 0.9112 - accuracy: 0.6900 - val_loss: 0.8571 - val_accuracy: 0.7234\n",
            "Epoch 559/2000\n",
            "3/3 [==============================] - 0s 44ms/step - loss: 0.8351 - accuracy: 0.7600 - val_loss: 0.8262 - val_accuracy: 0.7234\n",
            "Epoch 560/2000\n",
            "3/3 [==============================] - 0s 34ms/step - loss: 0.8531 - accuracy: 0.7800 - val_loss: 0.8225 - val_accuracy: 0.7021\n",
            "Epoch 561/2000\n",
            "3/3 [==============================] - 0s 44ms/step - loss: 0.8086 - accuracy: 0.7500 - val_loss: 0.8175 - val_accuracy: 0.7234\n",
            "Epoch 562/2000\n",
            "3/3 [==============================] - 0s 33ms/step - loss: 0.9254 - accuracy: 0.7600 - val_loss: 0.8074 - val_accuracy: 0.7447\n",
            "Epoch 563/2000\n",
            "3/3 [==============================] - 0s 39ms/step - loss: 0.7922 - accuracy: 0.7000 - val_loss: 0.8028 - val_accuracy: 0.7021\n",
            "Epoch 564/2000\n",
            "3/3 [==============================] - 0s 33ms/step - loss: 1.0708 - accuracy: 0.7000 - val_loss: 0.7924 - val_accuracy: 0.7447\n",
            "Epoch 565/2000\n",
            "3/3 [==============================] - 0s 38ms/step - loss: 1.0872 - accuracy: 0.6800 - val_loss: 0.7972 - val_accuracy: 0.7447\n",
            "Epoch 566/2000\n",
            "3/3 [==============================] - 0s 54ms/step - loss: 0.8759 - accuracy: 0.6500 - val_loss: 0.8230 - val_accuracy: 0.7447\n",
            "Epoch 567/2000\n",
            "3/3 [==============================] - 0s 32ms/step - loss: 0.8026 - accuracy: 0.7300 - val_loss: 0.8315 - val_accuracy: 0.7447\n",
            "Epoch 568/2000\n",
            "3/3 [==============================] - 0s 40ms/step - loss: 0.8574 - accuracy: 0.7700 - val_loss: 0.8276 - val_accuracy: 0.7234\n",
            "Epoch 569/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 0.8657 - accuracy: 0.7100 - val_loss: 0.8272 - val_accuracy: 0.7234\n",
            "Epoch 570/2000\n",
            "3/3 [==============================] - 0s 34ms/step - loss: 0.8941 - accuracy: 0.7200 - val_loss: 0.8306 - val_accuracy: 0.7447\n",
            "Epoch 571/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 0.8562 - accuracy: 0.7400 - val_loss: 0.8305 - val_accuracy: 0.7872\n",
            "Epoch 572/2000\n",
            "3/3 [==============================] - 0s 38ms/step - loss: 0.7276 - accuracy: 0.7500 - val_loss: 0.8180 - val_accuracy: 0.7660\n",
            "Epoch 573/2000\n",
            "3/3 [==============================] - 0s 32ms/step - loss: 0.9131 - accuracy: 0.7500 - val_loss: 0.7971 - val_accuracy: 0.7660\n",
            "Epoch 574/2000\n",
            "3/3 [==============================] - 0s 34ms/step - loss: 1.1003 - accuracy: 0.6500 - val_loss: 0.7906 - val_accuracy: 0.7447\n",
            "Epoch 575/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 1.0242 - accuracy: 0.6600 - val_loss: 0.7795 - val_accuracy: 0.7660\n",
            "Epoch 576/2000\n",
            "3/3 [==============================] - 0s 32ms/step - loss: 0.6306 - accuracy: 0.8200 - val_loss: 0.7765 - val_accuracy: 0.7234\n",
            "Epoch 577/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 0.8746 - accuracy: 0.7200 - val_loss: 0.7804 - val_accuracy: 0.7447\n",
            "Epoch 578/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 0.9676 - accuracy: 0.6900 - val_loss: 0.7677 - val_accuracy: 0.7660\n",
            "Epoch 579/2000\n",
            "3/3 [==============================] - 0s 33ms/step - loss: 0.8579 - accuracy: 0.7400 - val_loss: 0.7510 - val_accuracy: 0.7660\n",
            "Epoch 580/2000\n",
            "3/3 [==============================] - 0s 33ms/step - loss: 1.0503 - accuracy: 0.7400 - val_loss: 0.7347 - val_accuracy: 0.7660\n",
            "Epoch 581/2000\n",
            "3/3 [==============================] - 0s 33ms/step - loss: 0.8530 - accuracy: 0.7200 - val_loss: 0.7265 - val_accuracy: 0.8085\n",
            "Epoch 582/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 0.7918 - accuracy: 0.7500 - val_loss: 0.7253 - val_accuracy: 0.7872\n",
            "Epoch 583/2000\n",
            "3/3 [==============================] - 0s 40ms/step - loss: 0.7602 - accuracy: 0.7600 - val_loss: 0.7279 - val_accuracy: 0.8085\n",
            "Epoch 584/2000\n",
            "3/3 [==============================] - 0s 40ms/step - loss: 0.7883 - accuracy: 0.7500 - val_loss: 0.7213 - val_accuracy: 0.8085\n",
            "Epoch 585/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 0.9705 - accuracy: 0.7600 - val_loss: 0.7212 - val_accuracy: 0.8085\n",
            "Epoch 586/2000\n",
            "3/3 [==============================] - 0s 40ms/step - loss: 0.6832 - accuracy: 0.8000 - val_loss: 0.7409 - val_accuracy: 0.7660\n",
            "Epoch 587/2000\n",
            "3/3 [==============================] - 0s 38ms/step - loss: 0.8630 - accuracy: 0.7000 - val_loss: 0.7518 - val_accuracy: 0.7660\n",
            "Epoch 588/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 0.9811 - accuracy: 0.6400 - val_loss: 0.7529 - val_accuracy: 0.7234\n",
            "Epoch 589/2000\n",
            "3/3 [==============================] - 0s 32ms/step - loss: 0.8604 - accuracy: 0.7400 - val_loss: 0.7459 - val_accuracy: 0.7234\n",
            "Epoch 590/2000\n",
            "3/3 [==============================] - 0s 33ms/step - loss: 0.8755 - accuracy: 0.6700 - val_loss: 0.7373 - val_accuracy: 0.7447\n",
            "Epoch 591/2000\n",
            "3/3 [==============================] - 0s 32ms/step - loss: 0.8080 - accuracy: 0.7400 - val_loss: 0.7265 - val_accuracy: 0.7447\n",
            "Epoch 592/2000\n",
            "3/3 [==============================] - 0s 39ms/step - loss: 0.9203 - accuracy: 0.7800 - val_loss: 0.7154 - val_accuracy: 0.7660\n",
            "Epoch 593/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 0.8684 - accuracy: 0.7200 - val_loss: 0.7120 - val_accuracy: 0.7872\n",
            "Epoch 594/2000\n",
            "3/3 [==============================] - 0s 34ms/step - loss: 0.8323 - accuracy: 0.7400 - val_loss: 0.7186 - val_accuracy: 0.7872\n",
            "Epoch 595/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 0.8333 - accuracy: 0.7800 - val_loss: 0.7465 - val_accuracy: 0.7447\n",
            "Epoch 596/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 1.0479 - accuracy: 0.6500 - val_loss: 0.7848 - val_accuracy: 0.7021\n",
            "Epoch 597/2000\n",
            "3/3 [==============================] - 0s 34ms/step - loss: 0.9473 - accuracy: 0.7300 - val_loss: 0.8114 - val_accuracy: 0.7021\n",
            "Epoch 598/2000\n",
            "3/3 [==============================] - 0s 40ms/step - loss: 0.8411 - accuracy: 0.6900 - val_loss: 0.8014 - val_accuracy: 0.7021\n",
            "Epoch 599/2000\n",
            "3/3 [==============================] - 0s 46ms/step - loss: 0.8794 - accuracy: 0.6900 - val_loss: 0.7684 - val_accuracy: 0.7660\n",
            "Epoch 600/2000\n",
            "3/3 [==============================] - 0s 39ms/step - loss: 0.8440 - accuracy: 0.7400 - val_loss: 0.7816 - val_accuracy: 0.7234\n",
            "Epoch 601/2000\n",
            "3/3 [==============================] - 0s 38ms/step - loss: 0.9197 - accuracy: 0.7900 - val_loss: 0.8137 - val_accuracy: 0.7234\n",
            "Epoch 602/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 0.9070 - accuracy: 0.7500 - val_loss: 0.8320 - val_accuracy: 0.7021\n",
            "Epoch 603/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 0.7559 - accuracy: 0.7800 - val_loss: 0.8347 - val_accuracy: 0.6809\n",
            "Epoch 604/2000\n",
            "3/3 [==============================] - 0s 31ms/step - loss: 0.8924 - accuracy: 0.7700 - val_loss: 0.8051 - val_accuracy: 0.7447\n",
            "Epoch 605/2000\n",
            "3/3 [==============================] - 0s 33ms/step - loss: 0.8016 - accuracy: 0.7700 - val_loss: 0.7670 - val_accuracy: 0.7447\n",
            "Epoch 606/2000\n",
            "3/3 [==============================] - 0s 50ms/step - loss: 0.7701 - accuracy: 0.7600 - val_loss: 0.7626 - val_accuracy: 0.7447\n",
            "Epoch 607/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 0.7147 - accuracy: 0.7800 - val_loss: 0.7715 - val_accuracy: 0.7447\n",
            "Epoch 608/2000\n",
            "3/3 [==============================] - 0s 34ms/step - loss: 0.7851 - accuracy: 0.7300 - val_loss: 0.7764 - val_accuracy: 0.7660\n",
            "Epoch 609/2000\n",
            "3/3 [==============================] - 0s 39ms/step - loss: 0.7438 - accuracy: 0.7700 - val_loss: 0.7704 - val_accuracy: 0.7872\n",
            "Epoch 610/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 0.8166 - accuracy: 0.8000 - val_loss: 0.7914 - val_accuracy: 0.7660\n",
            "Epoch 611/2000\n",
            "3/3 [==============================] - 0s 39ms/step - loss: 0.8776 - accuracy: 0.7600 - val_loss: 0.8002 - val_accuracy: 0.7234\n",
            "Epoch 612/2000\n",
            "3/3 [==============================] - 0s 51ms/step - loss: 0.8886 - accuracy: 0.7200 - val_loss: 0.7969 - val_accuracy: 0.7234\n",
            "Epoch 613/2000\n",
            "3/3 [==============================] - 0s 32ms/step - loss: 0.7339 - accuracy: 0.7400 - val_loss: 0.7740 - val_accuracy: 0.7660\n",
            "Epoch 614/2000\n",
            "3/3 [==============================] - 0s 54ms/step - loss: 0.8546 - accuracy: 0.7800 - val_loss: 0.7481 - val_accuracy: 0.7660\n",
            "Epoch 615/2000\n",
            "3/3 [==============================] - 0s 48ms/step - loss: 0.7540 - accuracy: 0.7800 - val_loss: 0.7429 - val_accuracy: 0.7447\n",
            "Epoch 616/2000\n",
            "3/3 [==============================] - 0s 61ms/step - loss: 0.7816 - accuracy: 0.7300 - val_loss: 0.7315 - val_accuracy: 0.8085\n",
            "Epoch 617/2000\n",
            "3/3 [==============================] - 0s 58ms/step - loss: 0.7569 - accuracy: 0.7600 - val_loss: 0.7351 - val_accuracy: 0.7872\n",
            "Epoch 618/2000\n",
            "3/3 [==============================] - 0s 75ms/step - loss: 0.8793 - accuracy: 0.7100 - val_loss: 0.7441 - val_accuracy: 0.7872\n",
            "Epoch 619/2000\n",
            "3/3 [==============================] - 0s 62ms/step - loss: 0.9392 - accuracy: 0.6500 - val_loss: 0.7451 - val_accuracy: 0.7872\n",
            "Epoch 620/2000\n",
            "3/3 [==============================] - 0s 62ms/step - loss: 0.9352 - accuracy: 0.7800 - val_loss: 0.7449 - val_accuracy: 0.7660\n",
            "Epoch 621/2000\n",
            "3/3 [==============================] - 0s 44ms/step - loss: 0.7166 - accuracy: 0.7400 - val_loss: 0.7455 - val_accuracy: 0.7447\n",
            "Epoch 622/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 0.7850 - accuracy: 0.7100 - val_loss: 0.7457 - val_accuracy: 0.7660\n",
            "Epoch 623/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 0.9399 - accuracy: 0.7400 - val_loss: 0.7497 - val_accuracy: 0.7234\n",
            "Epoch 624/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 0.7782 - accuracy: 0.7400 - val_loss: 0.7277 - val_accuracy: 0.7021\n",
            "Epoch 625/2000\n",
            "3/3 [==============================] - 0s 39ms/step - loss: 0.8277 - accuracy: 0.7400 - val_loss: 0.7023 - val_accuracy: 0.7660\n",
            "Epoch 626/2000\n",
            "3/3 [==============================] - 0s 34ms/step - loss: 0.7170 - accuracy: 0.7500 - val_loss: 0.6943 - val_accuracy: 0.7660\n",
            "Epoch 627/2000\n",
            "3/3 [==============================] - 0s 34ms/step - loss: 0.8662 - accuracy: 0.7500 - val_loss: 0.6930 - val_accuracy: 0.7660\n",
            "Epoch 628/2000\n",
            "3/3 [==============================] - 0s 33ms/step - loss: 0.5880 - accuracy: 0.7600 - val_loss: 0.6966 - val_accuracy: 0.8085\n",
            "Epoch 629/2000\n",
            "3/3 [==============================] - 0s 55ms/step - loss: 0.6791 - accuracy: 0.7900 - val_loss: 0.7076 - val_accuracy: 0.7660\n",
            "Epoch 630/2000\n",
            "3/3 [==============================] - 0s 32ms/step - loss: 0.8826 - accuracy: 0.6900 - val_loss: 0.6957 - val_accuracy: 0.7872\n",
            "Epoch 631/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 0.7860 - accuracy: 0.7500 - val_loss: 0.6888 - val_accuracy: 0.7872\n",
            "Epoch 632/2000\n",
            "3/3 [==============================] - 0s 50ms/step - loss: 0.8147 - accuracy: 0.7500 - val_loss: 0.7016 - val_accuracy: 0.8085\n",
            "Epoch 633/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 0.7870 - accuracy: 0.8000 - val_loss: 0.7098 - val_accuracy: 0.7872\n",
            "Epoch 634/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 1.0631 - accuracy: 0.7000 - val_loss: 0.7020 - val_accuracy: 0.8085\n",
            "Epoch 635/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 0.8486 - accuracy: 0.7400 - val_loss: 0.6954 - val_accuracy: 0.8085\n",
            "Epoch 636/2000\n",
            "3/3 [==============================] - 0s 46ms/step - loss: 0.7165 - accuracy: 0.8000 - val_loss: 0.7024 - val_accuracy: 0.7872\n",
            "Epoch 637/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 0.7879 - accuracy: 0.7600 - val_loss: 0.6860 - val_accuracy: 0.7872\n",
            "Epoch 638/2000\n",
            "3/3 [==============================] - 0s 48ms/step - loss: 1.0029 - accuracy: 0.7000 - val_loss: 0.6645 - val_accuracy: 0.7872\n",
            "Epoch 639/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 0.8404 - accuracy: 0.6700 - val_loss: 0.6535 - val_accuracy: 0.7872\n",
            "Epoch 640/2000\n",
            "3/3 [==============================] - 0s 32ms/step - loss: 0.8302 - accuracy: 0.7300 - val_loss: 0.6543 - val_accuracy: 0.7447\n",
            "Epoch 641/2000\n",
            "3/3 [==============================] - 0s 53ms/step - loss: 0.8438 - accuracy: 0.7700 - val_loss: 0.6780 - val_accuracy: 0.7447\n",
            "Epoch 642/2000\n",
            "3/3 [==============================] - 0s 33ms/step - loss: 0.7937 - accuracy: 0.7700 - val_loss: 0.7077 - val_accuracy: 0.7447\n",
            "Epoch 643/2000\n",
            "3/3 [==============================] - 0s 32ms/step - loss: 0.9280 - accuracy: 0.7200 - val_loss: 0.7202 - val_accuracy: 0.7234\n",
            "Epoch 644/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 0.9347 - accuracy: 0.6600 - val_loss: 0.7194 - val_accuracy: 0.7021\n",
            "Epoch 645/2000\n",
            "3/3 [==============================] - 0s 32ms/step - loss: 0.8817 - accuracy: 0.7000 - val_loss: 0.7156 - val_accuracy: 0.7234\n",
            "Epoch 646/2000\n",
            "3/3 [==============================] - 0s 40ms/step - loss: 0.8706 - accuracy: 0.7300 - val_loss: 0.7115 - val_accuracy: 0.7447\n",
            "Epoch 647/2000\n",
            "3/3 [==============================] - 0s 40ms/step - loss: 0.8126 - accuracy: 0.7400 - val_loss: 0.6930 - val_accuracy: 0.7660\n",
            "Epoch 648/2000\n",
            "3/3 [==============================] - 0s 39ms/step - loss: 0.6450 - accuracy: 0.8200 - val_loss: 0.6743 - val_accuracy: 0.7660\n",
            "Epoch 649/2000\n",
            "3/3 [==============================] - 0s 44ms/step - loss: 0.8748 - accuracy: 0.7400 - val_loss: 0.6374 - val_accuracy: 0.7660\n",
            "Epoch 650/2000\n",
            "3/3 [==============================] - 0s 44ms/step - loss: 0.7782 - accuracy: 0.7300 - val_loss: 0.6282 - val_accuracy: 0.7872\n",
            "Epoch 651/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 0.7685 - accuracy: 0.7300 - val_loss: 0.6639 - val_accuracy: 0.7447\n",
            "Epoch 652/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 0.8466 - accuracy: 0.7600 - val_loss: 0.6811 - val_accuracy: 0.7234\n",
            "Epoch 653/2000\n",
            "3/3 [==============================] - 0s 47ms/step - loss: 0.8650 - accuracy: 0.7200 - val_loss: 0.6833 - val_accuracy: 0.7660\n",
            "Epoch 654/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 0.7654 - accuracy: 0.7800 - val_loss: 0.6616 - val_accuracy: 0.7660\n",
            "Epoch 655/2000\n",
            "3/3 [==============================] - 0s 44ms/step - loss: 0.8283 - accuracy: 0.7500 - val_loss: 0.6516 - val_accuracy: 0.7447\n",
            "Epoch 656/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 0.9081 - accuracy: 0.7400 - val_loss: 0.6747 - val_accuracy: 0.7660\n",
            "Epoch 657/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 0.6868 - accuracy: 0.7700 - val_loss: 0.7009 - val_accuracy: 0.7660\n",
            "Epoch 658/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 0.6904 - accuracy: 0.7300 - val_loss: 0.7032 - val_accuracy: 0.7660\n",
            "Epoch 659/2000\n",
            "3/3 [==============================] - 0s 31ms/step - loss: 0.6486 - accuracy: 0.8200 - val_loss: 0.7053 - val_accuracy: 0.7660\n",
            "Epoch 660/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 0.8084 - accuracy: 0.7700 - val_loss: 0.7147 - val_accuracy: 0.7234\n",
            "Epoch 661/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 0.7777 - accuracy: 0.7000 - val_loss: 0.7050 - val_accuracy: 0.7447\n",
            "Epoch 662/2000\n",
            "3/3 [==============================] - 0s 34ms/step - loss: 0.7907 - accuracy: 0.7700 - val_loss: 0.6785 - val_accuracy: 0.7660\n",
            "Epoch 663/2000\n",
            "3/3 [==============================] - 0s 38ms/step - loss: 0.7325 - accuracy: 0.8100 - val_loss: 0.6576 - val_accuracy: 0.7872\n",
            "Epoch 664/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 0.8606 - accuracy: 0.6600 - val_loss: 0.6623 - val_accuracy: 0.7660\n",
            "Epoch 665/2000\n",
            "3/3 [==============================] - 0s 55ms/step - loss: 0.8267 - accuracy: 0.7400 - val_loss: 0.7042 - val_accuracy: 0.7660\n",
            "Epoch 666/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 0.6093 - accuracy: 0.8300 - val_loss: 0.7153 - val_accuracy: 0.7660\n",
            "Epoch 667/2000\n",
            "3/3 [==============================] - 0s 34ms/step - loss: 0.6775 - accuracy: 0.7600 - val_loss: 0.6991 - val_accuracy: 0.7660\n",
            "Epoch 668/2000\n",
            "3/3 [==============================] - 0s 39ms/step - loss: 0.6871 - accuracy: 0.7800 - val_loss: 0.6929 - val_accuracy: 0.8085\n",
            "Epoch 669/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 0.8378 - accuracy: 0.7900 - val_loss: 0.6892 - val_accuracy: 0.8298\n",
            "Epoch 670/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 0.6820 - accuracy: 0.7900 - val_loss: 0.6930 - val_accuracy: 0.8085\n",
            "Epoch 671/2000\n",
            "3/3 [==============================] - 0s 34ms/step - loss: 0.8792 - accuracy: 0.7100 - val_loss: 0.6918 - val_accuracy: 0.8085\n",
            "Epoch 672/2000\n",
            "3/3 [==============================] - 0s 34ms/step - loss: 0.9923 - accuracy: 0.6900 - val_loss: 0.6991 - val_accuracy: 0.8085\n",
            "Epoch 673/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 0.8091 - accuracy: 0.7700 - val_loss: 0.6796 - val_accuracy: 0.8511\n",
            "Epoch 674/2000\n",
            "3/3 [==============================] - 0s 34ms/step - loss: 0.7973 - accuracy: 0.7500 - val_loss: 0.6657 - val_accuracy: 0.8723\n",
            "Epoch 675/2000\n",
            "3/3 [==============================] - 0s 33ms/step - loss: 0.7035 - accuracy: 0.7900 - val_loss: 0.6689 - val_accuracy: 0.7660\n",
            "Epoch 676/2000\n",
            "3/3 [==============================] - 0s 39ms/step - loss: 0.8394 - accuracy: 0.7100 - val_loss: 0.6704 - val_accuracy: 0.7447\n",
            "Epoch 677/2000\n",
            "3/3 [==============================] - 0s 33ms/step - loss: 0.7598 - accuracy: 0.7700 - val_loss: 0.6518 - val_accuracy: 0.7447\n",
            "Epoch 678/2000\n",
            "3/3 [==============================] - 0s 40ms/step - loss: 0.7879 - accuracy: 0.7800 - val_loss: 0.6438 - val_accuracy: 0.8298\n",
            "Epoch 679/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 0.9502 - accuracy: 0.6500 - val_loss: 0.6533 - val_accuracy: 0.8723\n",
            "Epoch 680/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 0.8443 - accuracy: 0.7500 - val_loss: 0.6713 - val_accuracy: 0.8085\n",
            "Epoch 681/2000\n",
            "3/3 [==============================] - 0s 40ms/step - loss: 0.8322 - accuracy: 0.7600 - val_loss: 0.6718 - val_accuracy: 0.8085\n",
            "Epoch 682/2000\n",
            "3/3 [==============================] - 0s 40ms/step - loss: 0.7789 - accuracy: 0.7800 - val_loss: 0.6907 - val_accuracy: 0.7872\n",
            "Epoch 683/2000\n",
            "3/3 [==============================] - 0s 34ms/step - loss: 0.8323 - accuracy: 0.6900 - val_loss: 0.6870 - val_accuracy: 0.7872\n",
            "Epoch 684/2000\n",
            "3/3 [==============================] - 0s 31ms/step - loss: 0.9985 - accuracy: 0.6700 - val_loss: 0.6885 - val_accuracy: 0.7660\n",
            "Epoch 685/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 0.7940 - accuracy: 0.7200 - val_loss: 0.6950 - val_accuracy: 0.7660\n",
            "Epoch 686/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 0.9122 - accuracy: 0.7300 - val_loss: 0.6974 - val_accuracy: 0.7660\n",
            "Epoch 687/2000\n",
            "3/3 [==============================] - 0s 45ms/step - loss: 0.6801 - accuracy: 0.8300 - val_loss: 0.6929 - val_accuracy: 0.7660\n",
            "Epoch 688/2000\n",
            "3/3 [==============================] - 0s 44ms/step - loss: 0.7218 - accuracy: 0.8200 - val_loss: 0.6996 - val_accuracy: 0.7872\n",
            "Epoch 689/2000\n",
            "3/3 [==============================] - 0s 53ms/step - loss: 0.8225 - accuracy: 0.6800 - val_loss: 0.7155 - val_accuracy: 0.7872\n",
            "Epoch 690/2000\n",
            "3/3 [==============================] - 0s 40ms/step - loss: 0.8194 - accuracy: 0.7600 - val_loss: 0.7228 - val_accuracy: 0.7660\n",
            "Epoch 691/2000\n",
            "3/3 [==============================] - 0s 66ms/step - loss: 0.7027 - accuracy: 0.8000 - val_loss: 0.6986 - val_accuracy: 0.7447\n",
            "Epoch 692/2000\n",
            "3/3 [==============================] - 0s 67ms/step - loss: 0.7036 - accuracy: 0.7500 - val_loss: 0.6966 - val_accuracy: 0.7447\n",
            "Epoch 693/2000\n",
            "3/3 [==============================] - 0s 59ms/step - loss: 0.6919 - accuracy: 0.7800 - val_loss: 0.7096 - val_accuracy: 0.7447\n",
            "Epoch 694/2000\n",
            "3/3 [==============================] - 0s 56ms/step - loss: 0.7547 - accuracy: 0.7700 - val_loss: 0.7040 - val_accuracy: 0.7660\n",
            "Epoch 695/2000\n",
            "3/3 [==============================] - 0s 54ms/step - loss: 0.8568 - accuracy: 0.7200 - val_loss: 0.7042 - val_accuracy: 0.7234\n",
            "Epoch 696/2000\n",
            "3/3 [==============================] - 0s 53ms/step - loss: 0.8171 - accuracy: 0.7700 - val_loss: 0.7001 - val_accuracy: 0.7447\n",
            "Epoch 697/2000\n",
            "3/3 [==============================] - 0s 55ms/step - loss: 0.6258 - accuracy: 0.8100 - val_loss: 0.7075 - val_accuracy: 0.7447\n",
            "Epoch 698/2000\n",
            "3/3 [==============================] - 0s 61ms/step - loss: 0.8697 - accuracy: 0.7200 - val_loss: 0.6988 - val_accuracy: 0.7660\n",
            "Epoch 699/2000\n",
            "3/3 [==============================] - 0s 68ms/step - loss: 0.7075 - accuracy: 0.8100 - val_loss: 0.6582 - val_accuracy: 0.7660\n",
            "Epoch 700/2000\n",
            "3/3 [==============================] - 0s 71ms/step - loss: 0.8698 - accuracy: 0.7500 - val_loss: 0.6285 - val_accuracy: 0.7872\n",
            "Epoch 701/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 0.6508 - accuracy: 0.8200 - val_loss: 0.6340 - val_accuracy: 0.7872\n",
            "Epoch 702/2000\n",
            "3/3 [==============================] - 0s 32ms/step - loss: 0.8459 - accuracy: 0.7400 - val_loss: 0.6495 - val_accuracy: 0.7872\n",
            "Epoch 703/2000\n",
            "3/3 [==============================] - 0s 40ms/step - loss: 0.7479 - accuracy: 0.7000 - val_loss: 0.6485 - val_accuracy: 0.7660\n",
            "Epoch 704/2000\n",
            "3/3 [==============================] - 0s 33ms/step - loss: 0.6469 - accuracy: 0.8000 - val_loss: 0.6763 - val_accuracy: 0.7872\n",
            "Epoch 705/2000\n",
            "3/3 [==============================] - 0s 34ms/step - loss: 0.8182 - accuracy: 0.7900 - val_loss: 0.7198 - val_accuracy: 0.7660\n",
            "Epoch 706/2000\n",
            "3/3 [==============================] - 0s 33ms/step - loss: 0.7279 - accuracy: 0.7000 - val_loss: 0.7437 - val_accuracy: 0.7234\n",
            "Epoch 707/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 0.8042 - accuracy: 0.7600 - val_loss: 0.7541 - val_accuracy: 0.7447\n",
            "Epoch 708/2000\n",
            "3/3 [==============================] - 0s 40ms/step - loss: 0.7464 - accuracy: 0.7200 - val_loss: 0.7358 - val_accuracy: 0.7234\n",
            "Epoch 709/2000\n",
            "3/3 [==============================] - 0s 40ms/step - loss: 0.5939 - accuracy: 0.8200 - val_loss: 0.7136 - val_accuracy: 0.7447\n",
            "Epoch 710/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 0.7348 - accuracy: 0.7500 - val_loss: 0.7035 - val_accuracy: 0.7447\n",
            "Epoch 711/2000\n",
            "3/3 [==============================] - 0s 40ms/step - loss: 0.7089 - accuracy: 0.7700 - val_loss: 0.7222 - val_accuracy: 0.7021\n",
            "Epoch 712/2000\n",
            "3/3 [==============================] - 0s 44ms/step - loss: 0.7429 - accuracy: 0.7800 - val_loss: 0.7396 - val_accuracy: 0.7234\n",
            "Epoch 713/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 0.7866 - accuracy: 0.7700 - val_loss: 0.7274 - val_accuracy: 0.7234\n",
            "Epoch 714/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 0.8080 - accuracy: 0.7200 - val_loss: 0.7157 - val_accuracy: 0.7447\n",
            "Epoch 715/2000\n",
            "3/3 [==============================] - 0s 39ms/step - loss: 0.7334 - accuracy: 0.7600 - val_loss: 0.7215 - val_accuracy: 0.7447\n",
            "Epoch 716/2000\n",
            "3/3 [==============================] - 0s 45ms/step - loss: 0.6591 - accuracy: 0.7800 - val_loss: 0.7385 - val_accuracy: 0.7234\n",
            "Epoch 717/2000\n",
            "3/3 [==============================] - 0s 40ms/step - loss: 0.7692 - accuracy: 0.7700 - val_loss: 0.7418 - val_accuracy: 0.7234\n",
            "Epoch 718/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 0.6666 - accuracy: 0.7700 - val_loss: 0.7249 - val_accuracy: 0.7660\n",
            "Epoch 719/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 0.9496 - accuracy: 0.6900 - val_loss: 0.7097 - val_accuracy: 0.7872\n",
            "Epoch 720/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 0.6853 - accuracy: 0.7900 - val_loss: 0.6886 - val_accuracy: 0.7660\n",
            "Epoch 721/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 0.7994 - accuracy: 0.7700 - val_loss: 0.6626 - val_accuracy: 0.7660\n",
            "Epoch 722/2000\n",
            "3/3 [==============================] - 0s 44ms/step - loss: 0.8286 - accuracy: 0.7500 - val_loss: 0.6557 - val_accuracy: 0.7872\n",
            "Epoch 723/2000\n",
            "3/3 [==============================] - 0s 54ms/step - loss: 0.8832 - accuracy: 0.7400 - val_loss: 0.6544 - val_accuracy: 0.7872\n",
            "Epoch 724/2000\n",
            "3/3 [==============================] - 0s 31ms/step - loss: 0.7562 - accuracy: 0.7300 - val_loss: 0.6535 - val_accuracy: 0.8298\n",
            "Epoch 725/2000\n",
            "3/3 [==============================] - 0s 45ms/step - loss: 0.8576 - accuracy: 0.7300 - val_loss: 0.6497 - val_accuracy: 0.8298\n",
            "Epoch 726/2000\n",
            "3/3 [==============================] - 0s 39ms/step - loss: 0.6749 - accuracy: 0.8000 - val_loss: 0.6536 - val_accuracy: 0.8085\n",
            "Epoch 727/2000\n",
            "3/3 [==============================] - 0s 33ms/step - loss: 0.9045 - accuracy: 0.7400 - val_loss: 0.6572 - val_accuracy: 0.8085\n",
            "Epoch 728/2000\n",
            "3/3 [==============================] - 0s 33ms/step - loss: 0.6465 - accuracy: 0.8000 - val_loss: 0.6699 - val_accuracy: 0.8511\n",
            "Epoch 729/2000\n",
            "3/3 [==============================] - 0s 34ms/step - loss: 0.8894 - accuracy: 0.7400 - val_loss: 0.6808 - val_accuracy: 0.8085\n",
            "Epoch 730/2000\n",
            "3/3 [==============================] - 0s 40ms/step - loss: 0.7122 - accuracy: 0.7500 - val_loss: 0.7013 - val_accuracy: 0.7872\n",
            "Epoch 731/2000\n",
            "3/3 [==============================] - 0s 40ms/step - loss: 0.8286 - accuracy: 0.7600 - val_loss: 0.7306 - val_accuracy: 0.7872\n",
            "Epoch 732/2000\n",
            "3/3 [==============================] - 0s 44ms/step - loss: 1.0879 - accuracy: 0.7500 - val_loss: 0.7353 - val_accuracy: 0.7872\n",
            "Epoch 733/2000\n",
            "3/3 [==============================] - 0s 45ms/step - loss: 0.9205 - accuracy: 0.6700 - val_loss: 0.7912 - val_accuracy: 0.7021\n",
            "Epoch 734/2000\n",
            "3/3 [==============================] - 0s 32ms/step - loss: 0.7334 - accuracy: 0.7700 - val_loss: 0.7848 - val_accuracy: 0.7021\n",
            "Epoch 735/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 0.8229 - accuracy: 0.7800 - val_loss: 0.7589 - val_accuracy: 0.7447\n",
            "Epoch 736/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 0.9865 - accuracy: 0.6800 - val_loss: 0.7335 - val_accuracy: 0.7447\n",
            "Epoch 737/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 0.7920 - accuracy: 0.8000 - val_loss: 0.7220 - val_accuracy: 0.7660\n",
            "Epoch 738/2000\n",
            "3/3 [==============================] - 0s 51ms/step - loss: 0.8352 - accuracy: 0.8000 - val_loss: 0.7173 - val_accuracy: 0.7660\n",
            "Epoch 739/2000\n",
            "3/3 [==============================] - 0s 40ms/step - loss: 0.7106 - accuracy: 0.8100 - val_loss: 0.7352 - val_accuracy: 0.7660\n",
            "Epoch 740/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 0.7895 - accuracy: 0.7700 - val_loss: 0.7131 - val_accuracy: 0.8085\n",
            "Epoch 741/2000\n",
            "3/3 [==============================] - 0s 32ms/step - loss: 0.6840 - accuracy: 0.8000 - val_loss: 0.6902 - val_accuracy: 0.8085\n",
            "Epoch 742/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 0.7301 - accuracy: 0.8100 - val_loss: 0.6821 - val_accuracy: 0.7872\n",
            "Epoch 743/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 0.7446 - accuracy: 0.7300 - val_loss: 0.6788 - val_accuracy: 0.7660\n",
            "Epoch 744/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 0.6491 - accuracy: 0.8200 - val_loss: 0.6802 - val_accuracy: 0.7660\n",
            "Epoch 745/2000\n",
            "3/3 [==============================] - 0s 31ms/step - loss: 0.7918 - accuracy: 0.7300 - val_loss: 0.6805 - val_accuracy: 0.7660\n",
            "Epoch 746/2000\n",
            "3/3 [==============================] - 0s 33ms/step - loss: 0.8557 - accuracy: 0.7200 - val_loss: 0.6955 - val_accuracy: 0.7447\n",
            "Epoch 747/2000\n",
            "3/3 [==============================] - 0s 32ms/step - loss: 0.6320 - accuracy: 0.8400 - val_loss: 0.6907 - val_accuracy: 0.7447\n",
            "Epoch 748/2000\n",
            "3/3 [==============================] - 0s 44ms/step - loss: 0.7034 - accuracy: 0.8200 - val_loss: 0.6680 - val_accuracy: 0.7447\n",
            "Epoch 749/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 0.7878 - accuracy: 0.8100 - val_loss: 0.6522 - val_accuracy: 0.7447\n",
            "Epoch 750/2000\n",
            "3/3 [==============================] - 0s 52ms/step - loss: 0.8578 - accuracy: 0.6900 - val_loss: 0.6426 - val_accuracy: 0.7872\n",
            "Epoch 751/2000\n",
            "3/3 [==============================] - 0s 47ms/step - loss: 0.9957 - accuracy: 0.7200 - val_loss: 0.6405 - val_accuracy: 0.7872\n",
            "Epoch 752/2000\n",
            "3/3 [==============================] - 0s 56ms/step - loss: 0.6594 - accuracy: 0.7800 - val_loss: 0.6594 - val_accuracy: 0.7447\n",
            "Epoch 753/2000\n",
            "3/3 [==============================] - 0s 40ms/step - loss: 0.8211 - accuracy: 0.7300 - val_loss: 0.6451 - val_accuracy: 0.7872\n",
            "Epoch 754/2000\n",
            "3/3 [==============================] - 0s 40ms/step - loss: 0.7880 - accuracy: 0.7000 - val_loss: 0.6191 - val_accuracy: 0.8085\n",
            "Epoch 755/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 0.7764 - accuracy: 0.7400 - val_loss: 0.6524 - val_accuracy: 0.7660\n",
            "Epoch 756/2000\n",
            "3/3 [==============================] - 0s 33ms/step - loss: 0.6376 - accuracy: 0.7900 - val_loss: 0.6684 - val_accuracy: 0.7447\n",
            "Epoch 757/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 0.9103 - accuracy: 0.6700 - val_loss: 0.6890 - val_accuracy: 0.7021\n",
            "Epoch 758/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 0.9759 - accuracy: 0.6500 - val_loss: 0.7110 - val_accuracy: 0.7234\n",
            "Epoch 759/2000\n",
            "3/3 [==============================] - 0s 31ms/step - loss: 0.7740 - accuracy: 0.7400 - val_loss: 0.6800 - val_accuracy: 0.7447\n",
            "Epoch 760/2000\n",
            "3/3 [==============================] - 0s 34ms/step - loss: 0.7556 - accuracy: 0.7700 - val_loss: 0.6198 - val_accuracy: 0.7660\n",
            "Epoch 761/2000\n",
            "3/3 [==============================] - 0s 40ms/step - loss: 0.6054 - accuracy: 0.8400 - val_loss: 0.6110 - val_accuracy: 0.7660\n",
            "Epoch 762/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 0.8828 - accuracy: 0.6800 - val_loss: 0.6491 - val_accuracy: 0.7660\n",
            "Epoch 763/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 0.8191 - accuracy: 0.7300 - val_loss: 0.6499 - val_accuracy: 0.7872\n",
            "Epoch 764/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 0.6881 - accuracy: 0.7800 - val_loss: 0.6599 - val_accuracy: 0.7872\n",
            "Epoch 765/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 0.7874 - accuracy: 0.7400 - val_loss: 0.6611 - val_accuracy: 0.7660\n",
            "Epoch 766/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 0.6545 - accuracy: 0.8200 - val_loss: 0.6637 - val_accuracy: 0.7447\n",
            "Epoch 767/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 0.8240 - accuracy: 0.7500 - val_loss: 0.6956 - val_accuracy: 0.7447\n",
            "Epoch 768/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 0.6860 - accuracy: 0.7400 - val_loss: 0.7258 - val_accuracy: 0.6809\n",
            "Epoch 769/2000\n",
            "3/3 [==============================] - 0s 33ms/step - loss: 0.6811 - accuracy: 0.7900 - val_loss: 0.7198 - val_accuracy: 0.7234\n",
            "Epoch 770/2000\n",
            "3/3 [==============================] - 0s 45ms/step - loss: 0.6446 - accuracy: 0.7600 - val_loss: 0.7164 - val_accuracy: 0.7660\n",
            "Epoch 771/2000\n",
            "3/3 [==============================] - 0s 60ms/step - loss: 0.8107 - accuracy: 0.7300 - val_loss: 0.6935 - val_accuracy: 0.7660\n",
            "Epoch 772/2000\n",
            "3/3 [==============================] - 0s 49ms/step - loss: 0.7296 - accuracy: 0.7900 - val_loss: 0.6677 - val_accuracy: 0.7872\n",
            "Epoch 773/2000\n",
            "3/3 [==============================] - 0s 60ms/step - loss: 0.6147 - accuracy: 0.7700 - val_loss: 0.6522 - val_accuracy: 0.8085\n",
            "Epoch 774/2000\n",
            "3/3 [==============================] - 0s 64ms/step - loss: 0.7326 - accuracy: 0.7500 - val_loss: 0.6240 - val_accuracy: 0.7872\n",
            "Epoch 775/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 0.6933 - accuracy: 0.7700 - val_loss: 0.6203 - val_accuracy: 0.7660\n",
            "Epoch 776/2000\n",
            "3/3 [==============================] - 0s 55ms/step - loss: 0.6446 - accuracy: 0.7700 - val_loss: 0.6456 - val_accuracy: 0.7872\n",
            "Epoch 777/2000\n",
            "3/3 [==============================] - 0s 64ms/step - loss: 0.6761 - accuracy: 0.7800 - val_loss: 0.6505 - val_accuracy: 0.7872\n",
            "Epoch 778/2000\n",
            "3/3 [==============================] - 0s 64ms/step - loss: 0.7107 - accuracy: 0.7800 - val_loss: 0.6399 - val_accuracy: 0.7872\n",
            "Epoch 779/2000\n",
            "3/3 [==============================] - 0s 67ms/step - loss: 0.7181 - accuracy: 0.7500 - val_loss: 0.6326 - val_accuracy: 0.7872\n",
            "Epoch 780/2000\n",
            "3/3 [==============================] - 0s 60ms/step - loss: 0.7010 - accuracy: 0.8100 - val_loss: 0.6166 - val_accuracy: 0.7872\n",
            "Epoch 781/2000\n",
            "3/3 [==============================] - 0s 45ms/step - loss: 0.6887 - accuracy: 0.7700 - val_loss: 0.6031 - val_accuracy: 0.8298\n",
            "Epoch 782/2000\n",
            "3/3 [==============================] - 0s 40ms/step - loss: 0.6193 - accuracy: 0.8200 - val_loss: 0.6007 - val_accuracy: 0.8085\n",
            "Epoch 783/2000\n",
            "3/3 [==============================] - 0s 44ms/step - loss: 0.8447 - accuracy: 0.6900 - val_loss: 0.6248 - val_accuracy: 0.7660\n",
            "Epoch 784/2000\n",
            "3/3 [==============================] - 0s 44ms/step - loss: 0.7878 - accuracy: 0.7800 - val_loss: 0.6157 - val_accuracy: 0.7872\n",
            "Epoch 785/2000\n",
            "3/3 [==============================] - 0s 33ms/step - loss: 0.8467 - accuracy: 0.7700 - val_loss: 0.6116 - val_accuracy: 0.7660\n",
            "Epoch 786/2000\n",
            "3/3 [==============================] - 0s 34ms/step - loss: 0.8542 - accuracy: 0.7500 - val_loss: 0.6033 - val_accuracy: 0.7447\n",
            "Epoch 787/2000\n",
            "3/3 [==============================] - 0s 32ms/step - loss: 0.6966 - accuracy: 0.8200 - val_loss: 0.6054 - val_accuracy: 0.7447\n",
            "Epoch 788/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 0.5788 - accuracy: 0.8400 - val_loss: 0.6165 - val_accuracy: 0.7447\n",
            "Epoch 789/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 0.8808 - accuracy: 0.6800 - val_loss: 0.6331 - val_accuracy: 0.8085\n",
            "Epoch 790/2000\n",
            "3/3 [==============================] - 0s 44ms/step - loss: 0.5988 - accuracy: 0.8100 - val_loss: 0.6474 - val_accuracy: 0.7872\n",
            "Epoch 791/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 0.7535 - accuracy: 0.7600 - val_loss: 0.6634 - val_accuracy: 0.7660\n",
            "Epoch 792/2000\n",
            "3/3 [==============================] - 0s 47ms/step - loss: 0.7267 - accuracy: 0.7700 - val_loss: 0.6701 - val_accuracy: 0.7660\n",
            "Epoch 793/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 0.5381 - accuracy: 0.8000 - val_loss: 0.6652 - val_accuracy: 0.7447\n",
            "Epoch 794/2000\n",
            "3/3 [==============================] - 0s 47ms/step - loss: 0.6959 - accuracy: 0.7500 - val_loss: 0.6665 - val_accuracy: 0.7872\n",
            "Epoch 795/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 0.6778 - accuracy: 0.7700 - val_loss: 0.6649 - val_accuracy: 0.8085\n",
            "Epoch 796/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 0.6146 - accuracy: 0.7600 - val_loss: 0.6734 - val_accuracy: 0.8298\n",
            "Epoch 797/2000\n",
            "3/3 [==============================] - 0s 55ms/step - loss: 0.6932 - accuracy: 0.7600 - val_loss: 0.6811 - val_accuracy: 0.8298\n",
            "Epoch 798/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 0.8120 - accuracy: 0.7800 - val_loss: 0.6825 - val_accuracy: 0.8085\n",
            "Epoch 799/2000\n",
            "3/3 [==============================] - 0s 47ms/step - loss: 0.4976 - accuracy: 0.8700 - val_loss: 0.6760 - val_accuracy: 0.8085\n",
            "Epoch 800/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 0.7859 - accuracy: 0.7900 - val_loss: 0.6855 - val_accuracy: 0.8085\n",
            "Epoch 801/2000\n",
            "3/3 [==============================] - 0s 52ms/step - loss: 0.7070 - accuracy: 0.8400 - val_loss: 0.6972 - val_accuracy: 0.8085\n",
            "Epoch 802/2000\n",
            "3/3 [==============================] - 0s 44ms/step - loss: 0.6955 - accuracy: 0.7900 - val_loss: 0.7322 - val_accuracy: 0.7660\n",
            "Epoch 803/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 0.5851 - accuracy: 0.8200 - val_loss: 0.6971 - val_accuracy: 0.7447\n",
            "Epoch 804/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 0.6900 - accuracy: 0.7800 - val_loss: 0.6686 - val_accuracy: 0.7447\n",
            "Epoch 805/2000\n",
            "3/3 [==============================] - 0s 45ms/step - loss: 0.6497 - accuracy: 0.7900 - val_loss: 0.6561 - val_accuracy: 0.7872\n",
            "Epoch 806/2000\n",
            "3/3 [==============================] - 0s 38ms/step - loss: 0.6120 - accuracy: 0.7900 - val_loss: 0.6456 - val_accuracy: 0.7660\n",
            "Epoch 807/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 0.7509 - accuracy: 0.7500 - val_loss: 0.6427 - val_accuracy: 0.7660\n",
            "Epoch 808/2000\n",
            "3/3 [==============================] - 0s 40ms/step - loss: 0.6244 - accuracy: 0.8100 - val_loss: 0.6499 - val_accuracy: 0.7872\n",
            "Epoch 809/2000\n",
            "3/3 [==============================] - 0s 52ms/step - loss: 0.7947 - accuracy: 0.7500 - val_loss: 0.6700 - val_accuracy: 0.7660\n",
            "Epoch 810/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 0.6156 - accuracy: 0.8000 - val_loss: 0.6727 - val_accuracy: 0.8085\n",
            "Epoch 811/2000\n",
            "3/3 [==============================] - 0s 44ms/step - loss: 0.7320 - accuracy: 0.7500 - val_loss: 0.6925 - val_accuracy: 0.7872\n",
            "Epoch 812/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 0.7117 - accuracy: 0.7900 - val_loss: 0.7074 - val_accuracy: 0.7872\n",
            "Epoch 813/2000\n",
            "3/3 [==============================] - 0s 32ms/step - loss: 0.6703 - accuracy: 0.7400 - val_loss: 0.7219 - val_accuracy: 0.8085\n",
            "Epoch 814/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 0.6550 - accuracy: 0.8000 - val_loss: 0.7135 - val_accuracy: 0.7447\n",
            "Epoch 815/2000\n",
            "3/3 [==============================] - 0s 50ms/step - loss: 0.6982 - accuracy: 0.7700 - val_loss: 0.6910 - val_accuracy: 0.7660\n",
            "Epoch 816/2000\n",
            "3/3 [==============================] - 0s 32ms/step - loss: 0.7170 - accuracy: 0.7500 - val_loss: 0.6539 - val_accuracy: 0.7660\n",
            "Epoch 817/2000\n",
            "3/3 [==============================] - 0s 38ms/step - loss: 0.6362 - accuracy: 0.8100 - val_loss: 0.6203 - val_accuracy: 0.8298\n",
            "Epoch 818/2000\n",
            "3/3 [==============================] - 0s 34ms/step - loss: 0.7362 - accuracy: 0.8100 - val_loss: 0.6070 - val_accuracy: 0.8298\n",
            "Epoch 819/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 0.7511 - accuracy: 0.7700 - val_loss: 0.5993 - val_accuracy: 0.8298\n",
            "Epoch 820/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 0.6059 - accuracy: 0.8300 - val_loss: 0.6000 - val_accuracy: 0.8298\n",
            "Epoch 821/2000\n",
            "3/3 [==============================] - 0s 34ms/step - loss: 0.7670 - accuracy: 0.7600 - val_loss: 0.5863 - val_accuracy: 0.7872\n",
            "Epoch 822/2000\n",
            "3/3 [==============================] - 0s 40ms/step - loss: 0.5863 - accuracy: 0.8000 - val_loss: 0.5855 - val_accuracy: 0.7660\n",
            "Epoch 823/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 0.8648 - accuracy: 0.7200 - val_loss: 0.5931 - val_accuracy: 0.7872\n",
            "Epoch 824/2000\n",
            "3/3 [==============================] - 0s 32ms/step - loss: 0.6979 - accuracy: 0.8100 - val_loss: 0.5926 - val_accuracy: 0.7872\n",
            "Epoch 825/2000\n",
            "3/3 [==============================] - 0s 31ms/step - loss: 0.6727 - accuracy: 0.7500 - val_loss: 0.5945 - val_accuracy: 0.7660\n",
            "Epoch 826/2000\n",
            "3/3 [==============================] - 0s 33ms/step - loss: 0.7260 - accuracy: 0.8100 - val_loss: 0.5894 - val_accuracy: 0.7660\n",
            "Epoch 827/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 0.7181 - accuracy: 0.7500 - val_loss: 0.5947 - val_accuracy: 0.7447\n",
            "Epoch 828/2000\n",
            "3/3 [==============================] - 0s 40ms/step - loss: 0.6660 - accuracy: 0.7500 - val_loss: 0.6135 - val_accuracy: 0.7660\n",
            "Epoch 829/2000\n",
            "3/3 [==============================] - 0s 33ms/step - loss: 0.5724 - accuracy: 0.7900 - val_loss: 0.6490 - val_accuracy: 0.7872\n",
            "Epoch 830/2000\n",
            "3/3 [==============================] - 0s 52ms/step - loss: 0.7283 - accuracy: 0.7500 - val_loss: 0.6866 - val_accuracy: 0.7234\n",
            "Epoch 831/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 0.7650 - accuracy: 0.7700 - val_loss: 0.6975 - val_accuracy: 0.7660\n",
            "Epoch 832/2000\n",
            "3/3 [==============================] - 0s 33ms/step - loss: 0.8559 - accuracy: 0.7200 - val_loss: 0.6835 - val_accuracy: 0.7660\n",
            "Epoch 833/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 0.6601 - accuracy: 0.8000 - val_loss: 0.6784 - val_accuracy: 0.8298\n",
            "Epoch 834/2000\n",
            "3/3 [==============================] - 0s 39ms/step - loss: 0.7786 - accuracy: 0.7300 - val_loss: 0.6870 - val_accuracy: 0.7872\n",
            "Epoch 835/2000\n",
            "3/3 [==============================] - 0s 34ms/step - loss: 0.7614 - accuracy: 0.7200 - val_loss: 0.6801 - val_accuracy: 0.7660\n",
            "Epoch 836/2000\n",
            "3/3 [==============================] - 0s 39ms/step - loss: 0.6253 - accuracy: 0.7700 - val_loss: 0.6406 - val_accuracy: 0.7872\n",
            "Epoch 837/2000\n",
            "3/3 [==============================] - 0s 32ms/step - loss: 0.7668 - accuracy: 0.8000 - val_loss: 0.6144 - val_accuracy: 0.8298\n",
            "Epoch 838/2000\n",
            "3/3 [==============================] - 0s 30ms/step - loss: 0.7484 - accuracy: 0.7300 - val_loss: 0.5970 - val_accuracy: 0.8298\n",
            "Epoch 839/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 0.5305 - accuracy: 0.8000 - val_loss: 0.5971 - val_accuracy: 0.8298\n",
            "Epoch 840/2000\n",
            "3/3 [==============================] - 0s 46ms/step - loss: 0.7253 - accuracy: 0.7500 - val_loss: 0.6304 - val_accuracy: 0.8085\n",
            "Epoch 841/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 0.9525 - accuracy: 0.6900 - val_loss: 0.6547 - val_accuracy: 0.8298\n",
            "Epoch 842/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 0.7419 - accuracy: 0.8000 - val_loss: 0.6315 - val_accuracy: 0.8298\n",
            "Epoch 843/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 0.5417 - accuracy: 0.8200 - val_loss: 0.6056 - val_accuracy: 0.8085\n",
            "Epoch 844/2000\n",
            "3/3 [==============================] - 0s 39ms/step - loss: 0.6384 - accuracy: 0.8100 - val_loss: 0.6074 - val_accuracy: 0.8298\n",
            "Epoch 845/2000\n",
            "3/3 [==============================] - 0s 44ms/step - loss: 0.8210 - accuracy: 0.7400 - val_loss: 0.6243 - val_accuracy: 0.7872\n",
            "Epoch 846/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 0.6254 - accuracy: 0.8000 - val_loss: 0.6373 - val_accuracy: 0.7447\n",
            "Epoch 847/2000\n",
            "3/3 [==============================] - 0s 40ms/step - loss: 0.5723 - accuracy: 0.8100 - val_loss: 0.6493 - val_accuracy: 0.7660\n",
            "Epoch 848/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 0.7566 - accuracy: 0.7000 - val_loss: 0.6540 - val_accuracy: 0.7872\n",
            "Epoch 849/2000\n",
            "3/3 [==============================] - 0s 62ms/step - loss: 0.8219 - accuracy: 0.7300 - val_loss: 0.6652 - val_accuracy: 0.7447\n",
            "Epoch 850/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 0.6816 - accuracy: 0.8100 - val_loss: 0.6610 - val_accuracy: 0.7660\n",
            "Epoch 851/2000\n",
            "3/3 [==============================] - 0s 74ms/step - loss: 0.7365 - accuracy: 0.7400 - val_loss: 0.6605 - val_accuracy: 0.8085\n",
            "Epoch 852/2000\n",
            "3/3 [==============================] - 0s 51ms/step - loss: 0.5885 - accuracy: 0.7800 - val_loss: 0.6528 - val_accuracy: 0.8511\n",
            "Epoch 853/2000\n",
            "3/3 [==============================] - 0s 61ms/step - loss: 0.5469 - accuracy: 0.8200 - val_loss: 0.6481 - val_accuracy: 0.8085\n",
            "Epoch 854/2000\n",
            "3/3 [==============================] - 0s 64ms/step - loss: 0.5513 - accuracy: 0.8200 - val_loss: 0.6555 - val_accuracy: 0.7660\n",
            "Epoch 855/2000\n",
            "3/3 [==============================] - 0s 53ms/step - loss: 0.6280 - accuracy: 0.8000 - val_loss: 0.6632 - val_accuracy: 0.7447\n",
            "Epoch 856/2000\n",
            "3/3 [==============================] - 0s 59ms/step - loss: 0.7893 - accuracy: 0.7200 - val_loss: 0.6158 - val_accuracy: 0.7660\n",
            "Epoch 857/2000\n",
            "3/3 [==============================] - 0s 64ms/step - loss: 0.7777 - accuracy: 0.7900 - val_loss: 0.5861 - val_accuracy: 0.8085\n",
            "Epoch 858/2000\n",
            "3/3 [==============================] - 0s 60ms/step - loss: 0.7342 - accuracy: 0.7900 - val_loss: 0.5730 - val_accuracy: 0.7872\n",
            "Epoch 859/2000\n",
            "3/3 [==============================] - 0s 52ms/step - loss: 0.5955 - accuracy: 0.7700 - val_loss: 0.5424 - val_accuracy: 0.8511\n",
            "Epoch 860/2000\n",
            "3/3 [==============================] - 0s 31ms/step - loss: 0.8135 - accuracy: 0.7600 - val_loss: 0.5337 - val_accuracy: 0.8511\n",
            "Epoch 861/2000\n",
            "3/3 [==============================] - 0s 40ms/step - loss: 0.6591 - accuracy: 0.7700 - val_loss: 0.5417 - val_accuracy: 0.8085\n",
            "Epoch 862/2000\n",
            "3/3 [==============================] - 0s 33ms/step - loss: 0.8316 - accuracy: 0.7500 - val_loss: 0.5541 - val_accuracy: 0.8085\n",
            "Epoch 863/2000\n",
            "3/3 [==============================] - 0s 33ms/step - loss: 0.6774 - accuracy: 0.8000 - val_loss: 0.5667 - val_accuracy: 0.8085\n",
            "Epoch 864/2000\n",
            "3/3 [==============================] - 0s 47ms/step - loss: 0.5088 - accuracy: 0.8500 - val_loss: 0.5762 - val_accuracy: 0.8085\n",
            "Epoch 865/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 0.6618 - accuracy: 0.7600 - val_loss: 0.5796 - val_accuracy: 0.8085\n",
            "Epoch 866/2000\n",
            "3/3 [==============================] - 0s 33ms/step - loss: 0.7544 - accuracy: 0.7900 - val_loss: 0.6036 - val_accuracy: 0.7660\n",
            "Epoch 867/2000\n",
            "3/3 [==============================] - 0s 45ms/step - loss: 0.6774 - accuracy: 0.6900 - val_loss: 0.6169 - val_accuracy: 0.7660\n",
            "Epoch 868/2000\n",
            "3/3 [==============================] - 0s 34ms/step - loss: 0.6806 - accuracy: 0.8000 - val_loss: 0.6094 - val_accuracy: 0.7660\n",
            "Epoch 869/2000\n",
            "3/3 [==============================] - 0s 34ms/step - loss: 0.9083 - accuracy: 0.7400 - val_loss: 0.6013 - val_accuracy: 0.7660\n",
            "Epoch 870/2000\n",
            "3/3 [==============================] - 0s 40ms/step - loss: 0.5522 - accuracy: 0.8100 - val_loss: 0.5981 - val_accuracy: 0.7872\n",
            "Epoch 871/2000\n",
            "3/3 [==============================] - 0s 39ms/step - loss: 0.6908 - accuracy: 0.7700 - val_loss: 0.5909 - val_accuracy: 0.7872\n",
            "Epoch 872/2000\n",
            "3/3 [==============================] - 0s 44ms/step - loss: 0.6749 - accuracy: 0.7800 - val_loss: 0.5981 - val_accuracy: 0.7660\n",
            "Epoch 873/2000\n",
            "3/3 [==============================] - 0s 52ms/step - loss: 0.7205 - accuracy: 0.7600 - val_loss: 0.6325 - val_accuracy: 0.8085\n",
            "Epoch 874/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 0.6325 - accuracy: 0.8100 - val_loss: 0.6647 - val_accuracy: 0.8085\n",
            "Epoch 875/2000\n",
            "3/3 [==============================] - 0s 40ms/step - loss: 0.6838 - accuracy: 0.7600 - val_loss: 0.6748 - val_accuracy: 0.8085\n",
            "Epoch 876/2000\n",
            "3/3 [==============================] - 0s 32ms/step - loss: 0.6212 - accuracy: 0.8100 - val_loss: 0.6717 - val_accuracy: 0.7872\n",
            "Epoch 877/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 0.6413 - accuracy: 0.7800 - val_loss: 0.6618 - val_accuracy: 0.7660\n",
            "Epoch 878/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 0.7821 - accuracy: 0.7600 - val_loss: 0.6413 - val_accuracy: 0.7660\n",
            "Epoch 879/2000\n",
            "3/3 [==============================] - 0s 47ms/step - loss: 0.6547 - accuracy: 0.8100 - val_loss: 0.6056 - val_accuracy: 0.8085\n",
            "Epoch 880/2000\n",
            "3/3 [==============================] - 0s 40ms/step - loss: 0.6465 - accuracy: 0.8300 - val_loss: 0.5903 - val_accuracy: 0.8085\n",
            "Epoch 881/2000\n",
            "3/3 [==============================] - 0s 38ms/step - loss: 0.5666 - accuracy: 0.8200 - val_loss: 0.5894 - val_accuracy: 0.7872\n",
            "Epoch 882/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 0.5952 - accuracy: 0.8000 - val_loss: 0.6051 - val_accuracy: 0.7872\n",
            "Epoch 883/2000\n",
            "3/3 [==============================] - 0s 32ms/step - loss: 0.6366 - accuracy: 0.8100 - val_loss: 0.6248 - val_accuracy: 0.7872\n",
            "Epoch 884/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 0.5832 - accuracy: 0.8300 - val_loss: 0.6241 - val_accuracy: 0.7660\n",
            "Epoch 885/2000\n",
            "3/3 [==============================] - 0s 33ms/step - loss: 0.5183 - accuracy: 0.8600 - val_loss: 0.6191 - val_accuracy: 0.7872\n",
            "Epoch 886/2000\n",
            "3/3 [==============================] - 0s 30ms/step - loss: 0.8551 - accuracy: 0.7400 - val_loss: 0.6166 - val_accuracy: 0.8085\n",
            "Epoch 887/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 0.6915 - accuracy: 0.7700 - val_loss: 0.6056 - val_accuracy: 0.8085\n",
            "Epoch 888/2000\n",
            "3/3 [==============================] - 0s 40ms/step - loss: 0.9443 - accuracy: 0.7500 - val_loss: 0.5948 - val_accuracy: 0.8085\n",
            "Epoch 889/2000\n",
            "3/3 [==============================] - 0s 33ms/step - loss: 0.6499 - accuracy: 0.8000 - val_loss: 0.5896 - val_accuracy: 0.7872\n",
            "Epoch 890/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 0.6450 - accuracy: 0.7700 - val_loss: 0.5868 - val_accuracy: 0.8085\n",
            "Epoch 891/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 0.7578 - accuracy: 0.7000 - val_loss: 0.5812 - val_accuracy: 0.8298\n",
            "Epoch 892/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 0.6387 - accuracy: 0.7600 - val_loss: 0.5945 - val_accuracy: 0.8298\n",
            "Epoch 893/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 0.7129 - accuracy: 0.7500 - val_loss: 0.5911 - val_accuracy: 0.8723\n",
            "Epoch 894/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 0.5914 - accuracy: 0.7900 - val_loss: 0.6020 - val_accuracy: 0.8298\n",
            "Epoch 895/2000\n",
            "3/3 [==============================] - 0s 40ms/step - loss: 0.6495 - accuracy: 0.8100 - val_loss: 0.6056 - val_accuracy: 0.8298\n",
            "Epoch 896/2000\n",
            "3/3 [==============================] - 0s 40ms/step - loss: 0.6373 - accuracy: 0.8400 - val_loss: 0.5848 - val_accuracy: 0.8511\n",
            "Epoch 897/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 0.5266 - accuracy: 0.8100 - val_loss: 0.5825 - val_accuracy: 0.8085\n",
            "Epoch 898/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 0.6396 - accuracy: 0.8200 - val_loss: 0.5822 - val_accuracy: 0.8085\n",
            "Epoch 899/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 0.6185 - accuracy: 0.7800 - val_loss: 0.6105 - val_accuracy: 0.7447\n",
            "Epoch 900/2000\n",
            "3/3 [==============================] - 0s 51ms/step - loss: 0.6514 - accuracy: 0.8000 - val_loss: 0.6059 - val_accuracy: 0.7872\n",
            "Epoch 901/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 0.6746 - accuracy: 0.8100 - val_loss: 0.6185 - val_accuracy: 0.7872\n",
            "Epoch 902/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 0.6038 - accuracy: 0.7600 - val_loss: 0.6236 - val_accuracy: 0.7872\n",
            "Epoch 903/2000\n",
            "3/3 [==============================] - 0s 50ms/step - loss: 0.6909 - accuracy: 0.7800 - val_loss: 0.6333 - val_accuracy: 0.7660\n",
            "Epoch 904/2000\n",
            "3/3 [==============================] - 0s 33ms/step - loss: 0.6435 - accuracy: 0.8200 - val_loss: 0.6434 - val_accuracy: 0.7660\n",
            "Epoch 905/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 0.7715 - accuracy: 0.7500 - val_loss: 0.6434 - val_accuracy: 0.7660\n",
            "Epoch 906/2000\n",
            "3/3 [==============================] - 0s 44ms/step - loss: 0.5622 - accuracy: 0.8100 - val_loss: 0.6394 - val_accuracy: 0.7660\n",
            "Epoch 907/2000\n",
            "3/3 [==============================] - 0s 40ms/step - loss: 0.6304 - accuracy: 0.8300 - val_loss: 0.6335 - val_accuracy: 0.7872\n",
            "Epoch 908/2000\n",
            "3/3 [==============================] - 0s 45ms/step - loss: 0.5649 - accuracy: 0.8300 - val_loss: 0.6360 - val_accuracy: 0.7872\n",
            "Epoch 909/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 0.7950 - accuracy: 0.7800 - val_loss: 0.6509 - val_accuracy: 0.7872\n",
            "Epoch 910/2000\n",
            "3/3 [==============================] - 0s 33ms/step - loss: 0.6196 - accuracy: 0.8300 - val_loss: 0.6546 - val_accuracy: 0.7872\n",
            "Epoch 911/2000\n",
            "3/3 [==============================] - 0s 33ms/step - loss: 0.8324 - accuracy: 0.7100 - val_loss: 0.6439 - val_accuracy: 0.7872\n",
            "Epoch 912/2000\n",
            "3/3 [==============================] - 0s 40ms/step - loss: 0.6410 - accuracy: 0.7900 - val_loss: 0.6137 - val_accuracy: 0.8298\n",
            "Epoch 913/2000\n",
            "3/3 [==============================] - 0s 40ms/step - loss: 0.6446 - accuracy: 0.7600 - val_loss: 0.5958 - val_accuracy: 0.8085\n",
            "Epoch 914/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 0.5957 - accuracy: 0.8300 - val_loss: 0.5960 - val_accuracy: 0.7872\n",
            "Epoch 915/2000\n",
            "3/3 [==============================] - 0s 46ms/step - loss: 0.6501 - accuracy: 0.7600 - val_loss: 0.5986 - val_accuracy: 0.7872\n",
            "Epoch 916/2000\n",
            "3/3 [==============================] - 0s 46ms/step - loss: 0.6433 - accuracy: 0.8200 - val_loss: 0.6041 - val_accuracy: 0.8085\n",
            "Epoch 917/2000\n",
            "3/3 [==============================] - 0s 51ms/step - loss: 0.6336 - accuracy: 0.8200 - val_loss: 0.6024 - val_accuracy: 0.7872\n",
            "Epoch 918/2000\n",
            "3/3 [==============================] - 0s 44ms/step - loss: 0.9287 - accuracy: 0.7200 - val_loss: 0.6001 - val_accuracy: 0.8085\n",
            "Epoch 919/2000\n",
            "3/3 [==============================] - 0s 40ms/step - loss: 0.6054 - accuracy: 0.8400 - val_loss: 0.5944 - val_accuracy: 0.8085\n",
            "Epoch 920/2000\n",
            "3/3 [==============================] - 0s 33ms/step - loss: 0.6404 - accuracy: 0.7800 - val_loss: 0.5909 - val_accuracy: 0.7872\n",
            "Epoch 921/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 0.6767 - accuracy: 0.7900 - val_loss: 0.5978 - val_accuracy: 0.7872\n",
            "Epoch 922/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 0.6356 - accuracy: 0.7900 - val_loss: 0.5934 - val_accuracy: 0.7872\n",
            "Epoch 923/2000\n",
            "3/3 [==============================] - 0s 45ms/step - loss: 0.7142 - accuracy: 0.7900 - val_loss: 0.5810 - val_accuracy: 0.7872\n",
            "Epoch 924/2000\n",
            "3/3 [==============================] - 0s 53ms/step - loss: 0.6412 - accuracy: 0.8000 - val_loss: 0.5783 - val_accuracy: 0.8085\n",
            "Epoch 925/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 0.6902 - accuracy: 0.7500 - val_loss: 0.5756 - val_accuracy: 0.8085\n",
            "Epoch 926/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 0.6084 - accuracy: 0.8300 - val_loss: 0.5658 - val_accuracy: 0.8298\n",
            "Epoch 927/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 0.7104 - accuracy: 0.8300 - val_loss: 0.5594 - val_accuracy: 0.8298\n",
            "Epoch 928/2000\n",
            "3/3 [==============================] - 0s 66ms/step - loss: 0.6690 - accuracy: 0.7900 - val_loss: 0.5551 - val_accuracy: 0.8298\n",
            "Epoch 929/2000\n",
            "3/3 [==============================] - 0s 68ms/step - loss: 0.6126 - accuracy: 0.8100 - val_loss: 0.5514 - val_accuracy: 0.8298\n",
            "Epoch 930/2000\n",
            "3/3 [==============================] - 0s 69ms/step - loss: 0.6196 - accuracy: 0.7900 - val_loss: 0.5634 - val_accuracy: 0.8085\n",
            "Epoch 931/2000\n",
            "3/3 [==============================] - 0s 58ms/step - loss: 0.9224 - accuracy: 0.7100 - val_loss: 0.5981 - val_accuracy: 0.8085\n",
            "Epoch 932/2000\n",
            "3/3 [==============================] - 0s 71ms/step - loss: 0.6808 - accuracy: 0.7900 - val_loss: 0.6117 - val_accuracy: 0.7872\n",
            "Epoch 933/2000\n",
            "3/3 [==============================] - 0s 69ms/step - loss: 0.6397 - accuracy: 0.7900 - val_loss: 0.6045 - val_accuracy: 0.7660\n",
            "Epoch 934/2000\n",
            "3/3 [==============================] - 0s 58ms/step - loss: 0.7393 - accuracy: 0.7800 - val_loss: 0.5929 - val_accuracy: 0.7660\n",
            "Epoch 935/2000\n",
            "3/3 [==============================] - 0s 70ms/step - loss: 0.6373 - accuracy: 0.8000 - val_loss: 0.5848 - val_accuracy: 0.7660\n",
            "Epoch 936/2000\n",
            "3/3 [==============================] - 0s 57ms/step - loss: 0.4846 - accuracy: 0.8600 - val_loss: 0.5738 - val_accuracy: 0.7872\n",
            "Epoch 937/2000\n",
            "3/3 [==============================] - 0s 73ms/step - loss: 0.7050 - accuracy: 0.7300 - val_loss: 0.5695 - val_accuracy: 0.7872\n",
            "Epoch 938/2000\n",
            "3/3 [==============================] - 0s 60ms/step - loss: 0.6587 - accuracy: 0.7600 - val_loss: 0.5797 - val_accuracy: 0.7872\n",
            "Epoch 939/2000\n",
            "3/3 [==============================] - 0s 66ms/step - loss: 0.5835 - accuracy: 0.7900 - val_loss: 0.5975 - val_accuracy: 0.7872\n",
            "Epoch 940/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 0.5422 - accuracy: 0.8100 - val_loss: 0.5954 - val_accuracy: 0.8085\n",
            "Epoch 941/2000\n",
            "3/3 [==============================] - 0s 32ms/step - loss: 0.6620 - accuracy: 0.8100 - val_loss: 0.5829 - val_accuracy: 0.8085\n",
            "Epoch 942/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 0.8005 - accuracy: 0.7100 - val_loss: 0.5665 - val_accuracy: 0.8085\n",
            "Epoch 943/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 0.6637 - accuracy: 0.7800 - val_loss: 0.5709 - val_accuracy: 0.8298\n",
            "Epoch 944/2000\n",
            "3/3 [==============================] - 0s 39ms/step - loss: 0.7368 - accuracy: 0.7700 - val_loss: 0.5745 - val_accuracy: 0.8511\n",
            "Epoch 945/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 0.7536 - accuracy: 0.7100 - val_loss: 0.5917 - val_accuracy: 0.8085\n",
            "Epoch 946/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 0.5730 - accuracy: 0.8000 - val_loss: 0.6182 - val_accuracy: 0.7447\n",
            "Epoch 947/2000\n",
            "3/3 [==============================] - 0s 33ms/step - loss: 0.4680 - accuracy: 0.8300 - val_loss: 0.6369 - val_accuracy: 0.7660\n",
            "Epoch 948/2000\n",
            "3/3 [==============================] - 0s 29ms/step - loss: 0.7016 - accuracy: 0.7500 - val_loss: 0.6254 - val_accuracy: 0.7660\n",
            "Epoch 949/2000\n",
            "3/3 [==============================] - 0s 45ms/step - loss: 0.6298 - accuracy: 0.7900 - val_loss: 0.5925 - val_accuracy: 0.7872\n",
            "Epoch 950/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 0.5349 - accuracy: 0.8400 - val_loss: 0.5629 - val_accuracy: 0.7660\n",
            "Epoch 951/2000\n",
            "3/3 [==============================] - 0s 32ms/step - loss: 0.5932 - accuracy: 0.7900 - val_loss: 0.5623 - val_accuracy: 0.8085\n",
            "Epoch 952/2000\n",
            "3/3 [==============================] - 0s 40ms/step - loss: 0.6667 - accuracy: 0.7900 - val_loss: 0.5649 - val_accuracy: 0.8298\n",
            "Epoch 953/2000\n",
            "3/3 [==============================] - 0s 51ms/step - loss: 0.8137 - accuracy: 0.7700 - val_loss: 0.5762 - val_accuracy: 0.8085\n",
            "Epoch 954/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 0.6593 - accuracy: 0.7900 - val_loss: 0.6036 - val_accuracy: 0.7660\n",
            "Epoch 955/2000\n",
            "3/3 [==============================] - 0s 33ms/step - loss: 0.6899 - accuracy: 0.7800 - val_loss: 0.6375 - val_accuracy: 0.7660\n",
            "Epoch 956/2000\n",
            "3/3 [==============================] - 0s 34ms/step - loss: 0.5813 - accuracy: 0.8500 - val_loss: 0.6414 - val_accuracy: 0.7660\n",
            "Epoch 957/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 0.6029 - accuracy: 0.7900 - val_loss: 0.6193 - val_accuracy: 0.7660\n",
            "Epoch 958/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 0.6682 - accuracy: 0.8000 - val_loss: 0.5931 - val_accuracy: 0.7872\n",
            "Epoch 959/2000\n",
            "3/3 [==============================] - 0s 40ms/step - loss: 0.6176 - accuracy: 0.8000 - val_loss: 0.5644 - val_accuracy: 0.8085\n",
            "Epoch 960/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 0.5819 - accuracy: 0.8100 - val_loss: 0.5481 - val_accuracy: 0.8085\n",
            "Epoch 961/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 0.4802 - accuracy: 0.8400 - val_loss: 0.5432 - val_accuracy: 0.8085\n",
            "Epoch 962/2000\n",
            "3/3 [==============================] - 0s 33ms/step - loss: 0.6561 - accuracy: 0.7900 - val_loss: 0.5510 - val_accuracy: 0.8085\n",
            "Epoch 963/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 0.6589 - accuracy: 0.8200 - val_loss: 0.5661 - val_accuracy: 0.8085\n",
            "Epoch 964/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 0.5550 - accuracy: 0.8400 - val_loss: 0.5839 - val_accuracy: 0.7872\n",
            "Epoch 965/2000\n",
            "3/3 [==============================] - 0s 51ms/step - loss: 0.5476 - accuracy: 0.8400 - val_loss: 0.5935 - val_accuracy: 0.7872\n",
            "Epoch 966/2000\n",
            "3/3 [==============================] - 0s 44ms/step - loss: 0.5958 - accuracy: 0.8100 - val_loss: 0.6070 - val_accuracy: 0.7660\n",
            "Epoch 967/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 0.6931 - accuracy: 0.7800 - val_loss: 0.6181 - val_accuracy: 0.7872\n",
            "Epoch 968/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 0.6836 - accuracy: 0.8100 - val_loss: 0.6130 - val_accuracy: 0.8085\n",
            "Epoch 969/2000\n",
            "3/3 [==============================] - 0s 45ms/step - loss: 0.6366 - accuracy: 0.7700 - val_loss: 0.5971 - val_accuracy: 0.7660\n",
            "Epoch 970/2000\n",
            "3/3 [==============================] - 0s 33ms/step - loss: 0.7039 - accuracy: 0.7800 - val_loss: 0.5801 - val_accuracy: 0.7872\n",
            "Epoch 971/2000\n",
            "3/3 [==============================] - 0s 33ms/step - loss: 0.5663 - accuracy: 0.8200 - val_loss: 0.5761 - val_accuracy: 0.7447\n",
            "Epoch 972/2000\n",
            "3/3 [==============================] - 0s 46ms/step - loss: 0.6830 - accuracy: 0.7700 - val_loss: 0.5784 - val_accuracy: 0.7660\n",
            "Epoch 973/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 0.6283 - accuracy: 0.8300 - val_loss: 0.5936 - val_accuracy: 0.7660\n",
            "Epoch 974/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 0.5778 - accuracy: 0.8400 - val_loss: 0.5925 - val_accuracy: 0.7872\n",
            "Epoch 975/2000\n",
            "3/3 [==============================] - 0s 34ms/step - loss: 0.4740 - accuracy: 0.9000 - val_loss: 0.5703 - val_accuracy: 0.7872\n",
            "Epoch 976/2000\n",
            "3/3 [==============================] - 0s 33ms/step - loss: 0.6130 - accuracy: 0.7900 - val_loss: 0.5382 - val_accuracy: 0.7872\n",
            "Epoch 977/2000\n",
            "3/3 [==============================] - 0s 34ms/step - loss: 0.4764 - accuracy: 0.8600 - val_loss: 0.5234 - val_accuracy: 0.8511\n",
            "Epoch 978/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 0.5418 - accuracy: 0.8000 - val_loss: 0.5215 - val_accuracy: 0.8723\n",
            "Epoch 979/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 0.7443 - accuracy: 0.7600 - val_loss: 0.5364 - val_accuracy: 0.8511\n",
            "Epoch 980/2000\n",
            "3/3 [==============================] - 0s 40ms/step - loss: 0.5788 - accuracy: 0.7900 - val_loss: 0.5538 - val_accuracy: 0.8511\n",
            "Epoch 981/2000\n",
            "3/3 [==============================] - 0s 53ms/step - loss: 0.5885 - accuracy: 0.8100 - val_loss: 0.5698 - val_accuracy: 0.8723\n",
            "Epoch 982/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 0.5765 - accuracy: 0.8100 - val_loss: 0.5730 - val_accuracy: 0.8511\n",
            "Epoch 983/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 0.7008 - accuracy: 0.7100 - val_loss: 0.5769 - val_accuracy: 0.8511\n",
            "Epoch 984/2000\n",
            "3/3 [==============================] - 0s 40ms/step - loss: 0.6414 - accuracy: 0.8100 - val_loss: 0.5865 - val_accuracy: 0.8085\n",
            "Epoch 985/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 0.5928 - accuracy: 0.8100 - val_loss: 0.5713 - val_accuracy: 0.8298\n",
            "Epoch 986/2000\n",
            "3/3 [==============================] - 0s 33ms/step - loss: 0.6850 - accuracy: 0.7400 - val_loss: 0.5546 - val_accuracy: 0.8511\n",
            "Epoch 987/2000\n",
            "3/3 [==============================] - 0s 34ms/step - loss: 0.6685 - accuracy: 0.7900 - val_loss: 0.5371 - val_accuracy: 0.8298\n",
            "Epoch 988/2000\n",
            "3/3 [==============================] - 0s 44ms/step - loss: 0.5979 - accuracy: 0.8400 - val_loss: 0.5295 - val_accuracy: 0.8298\n",
            "Epoch 989/2000\n",
            "3/3 [==============================] - 0s 33ms/step - loss: 0.5235 - accuracy: 0.8100 - val_loss: 0.5269 - val_accuracy: 0.8298\n",
            "Epoch 990/2000\n",
            "3/3 [==============================] - 0s 34ms/step - loss: 0.7024 - accuracy: 0.7900 - val_loss: 0.5377 - val_accuracy: 0.8085\n",
            "Epoch 991/2000\n",
            "3/3 [==============================] - 0s 38ms/step - loss: 0.6605 - accuracy: 0.7700 - val_loss: 0.5729 - val_accuracy: 0.8085\n",
            "Epoch 992/2000\n",
            "3/3 [==============================] - 0s 53ms/step - loss: 0.5540 - accuracy: 0.7900 - val_loss: 0.6056 - val_accuracy: 0.8085\n",
            "Epoch 993/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 0.8412 - accuracy: 0.7500 - val_loss: 0.5939 - val_accuracy: 0.8085\n",
            "Epoch 994/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 0.4868 - accuracy: 0.8500 - val_loss: 0.5647 - val_accuracy: 0.7872\n",
            "Epoch 995/2000\n",
            "3/3 [==============================] - 0s 34ms/step - loss: 0.5549 - accuracy: 0.8600 - val_loss: 0.5499 - val_accuracy: 0.7872\n",
            "Epoch 996/2000\n",
            "3/3 [==============================] - 0s 40ms/step - loss: 0.6996 - accuracy: 0.7900 - val_loss: 0.5460 - val_accuracy: 0.7872\n",
            "Epoch 997/2000\n",
            "3/3 [==============================] - 0s 44ms/step - loss: 0.5237 - accuracy: 0.8700 - val_loss: 0.5564 - val_accuracy: 0.8298\n",
            "Epoch 998/2000\n",
            "3/3 [==============================] - 0s 45ms/step - loss: 0.5392 - accuracy: 0.8500 - val_loss: 0.5618 - val_accuracy: 0.8298\n",
            "Epoch 999/2000\n",
            "3/3 [==============================] - 0s 47ms/step - loss: 0.5620 - accuracy: 0.8400 - val_loss: 0.5542 - val_accuracy: 0.8298\n",
            "Epoch 1000/2000\n",
            "3/3 [==============================] - 0s 38ms/step - loss: 0.5882 - accuracy: 0.8100 - val_loss: 0.5373 - val_accuracy: 0.8298\n",
            "Epoch 1001/2000\n",
            "3/3 [==============================] - 0s 57ms/step - loss: 0.6638 - accuracy: 0.7900 - val_loss: 0.5396 - val_accuracy: 0.8298\n",
            "Epoch 1002/2000\n",
            "3/3 [==============================] - 0s 47ms/step - loss: 0.5884 - accuracy: 0.7700 - val_loss: 0.5622 - val_accuracy: 0.8085\n",
            "Epoch 1003/2000\n",
            "3/3 [==============================] - 0s 45ms/step - loss: 0.6607 - accuracy: 0.8500 - val_loss: 0.5511 - val_accuracy: 0.8085\n",
            "Epoch 1004/2000\n",
            "3/3 [==============================] - 0s 45ms/step - loss: 0.5473 - accuracy: 0.8300 - val_loss: 0.5380 - val_accuracy: 0.8298\n",
            "Epoch 1005/2000\n",
            "3/3 [==============================] - 0s 62ms/step - loss: 0.4717 - accuracy: 0.8500 - val_loss: 0.5439 - val_accuracy: 0.8085\n",
            "Epoch 1006/2000\n",
            "3/3 [==============================] - 0s 75ms/step - loss: 0.6247 - accuracy: 0.7700 - val_loss: 0.5598 - val_accuracy: 0.8511\n",
            "Epoch 1007/2000\n",
            "3/3 [==============================] - 0s 55ms/step - loss: 0.8009 - accuracy: 0.7500 - val_loss: 0.5717 - val_accuracy: 0.8511\n",
            "Epoch 1008/2000\n",
            "3/3 [==============================] - 0s 47ms/step - loss: 0.5443 - accuracy: 0.8300 - val_loss: 0.5910 - val_accuracy: 0.8085\n",
            "Epoch 1009/2000\n",
            "3/3 [==============================] - 0s 51ms/step - loss: 0.6005 - accuracy: 0.7900 - val_loss: 0.6055 - val_accuracy: 0.7872\n",
            "Epoch 1010/2000\n",
            "3/3 [==============================] - 0s 64ms/step - loss: 0.5324 - accuracy: 0.8300 - val_loss: 0.6050 - val_accuracy: 0.7447\n",
            "Epoch 1011/2000\n",
            "3/3 [==============================] - 0s 71ms/step - loss: 0.6778 - accuracy: 0.7900 - val_loss: 0.5897 - val_accuracy: 0.7872\n",
            "Epoch 1012/2000\n",
            "3/3 [==============================] - 0s 77ms/step - loss: 0.5309 - accuracy: 0.8200 - val_loss: 0.5781 - val_accuracy: 0.7660\n",
            "Epoch 1013/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 0.5872 - accuracy: 0.8100 - val_loss: 0.5625 - val_accuracy: 0.8085\n",
            "Epoch 1014/2000\n",
            "3/3 [==============================] - 0s 44ms/step - loss: 0.6714 - accuracy: 0.7800 - val_loss: 0.5465 - val_accuracy: 0.8298\n",
            "Epoch 1015/2000\n",
            "3/3 [==============================] - 0s 38ms/step - loss: 0.6279 - accuracy: 0.7900 - val_loss: 0.5509 - val_accuracy: 0.8085\n",
            "Epoch 1016/2000\n",
            "3/3 [==============================] - 0s 50ms/step - loss: 0.4903 - accuracy: 0.8600 - val_loss: 0.5608 - val_accuracy: 0.8085\n",
            "Epoch 1017/2000\n",
            "3/3 [==============================] - 0s 55ms/step - loss: 0.6173 - accuracy: 0.7600 - val_loss: 0.5600 - val_accuracy: 0.8298\n",
            "Epoch 1018/2000\n",
            "3/3 [==============================] - 0s 45ms/step - loss: 0.7274 - accuracy: 0.7500 - val_loss: 0.5381 - val_accuracy: 0.8511\n",
            "Epoch 1019/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 0.8035 - accuracy: 0.8100 - val_loss: 0.5203 - val_accuracy: 0.8298\n",
            "Epoch 1020/2000\n",
            "3/3 [==============================] - 0s 44ms/step - loss: 0.5773 - accuracy: 0.8400 - val_loss: 0.5308 - val_accuracy: 0.8085\n",
            "Epoch 1021/2000\n",
            "3/3 [==============================] - 0s 57ms/step - loss: 0.6563 - accuracy: 0.7600 - val_loss: 0.5415 - val_accuracy: 0.7872\n",
            "Epoch 1022/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 0.7531 - accuracy: 0.7500 - val_loss: 0.5538 - val_accuracy: 0.7660\n",
            "Epoch 1023/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 0.5769 - accuracy: 0.7900 - val_loss: 0.5712 - val_accuracy: 0.8085\n",
            "Epoch 1024/2000\n",
            "3/3 [==============================] - 0s 49ms/step - loss: 0.4248 - accuracy: 0.8900 - val_loss: 0.5727 - val_accuracy: 0.7872\n",
            "Epoch 1025/2000\n",
            "3/3 [==============================] - 0s 44ms/step - loss: 0.4946 - accuracy: 0.8900 - val_loss: 0.5995 - val_accuracy: 0.7872\n",
            "Epoch 1026/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 0.6710 - accuracy: 0.7100 - val_loss: 0.6152 - val_accuracy: 0.8085\n",
            "Epoch 1027/2000\n",
            "3/3 [==============================] - 0s 40ms/step - loss: 0.7517 - accuracy: 0.7700 - val_loss: 0.6073 - val_accuracy: 0.8085\n",
            "Epoch 1028/2000\n",
            "3/3 [==============================] - 0s 55ms/step - loss: 0.8006 - accuracy: 0.7400 - val_loss: 0.5908 - val_accuracy: 0.7872\n",
            "Epoch 1029/2000\n",
            "3/3 [==============================] - 0s 33ms/step - loss: 0.5689 - accuracy: 0.8300 - val_loss: 0.5625 - val_accuracy: 0.8085\n",
            "Epoch 1030/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 0.6252 - accuracy: 0.8400 - val_loss: 0.5451 - val_accuracy: 0.8085\n",
            "Epoch 1031/2000\n",
            "3/3 [==============================] - 0s 52ms/step - loss: 1.0556 - accuracy: 0.7300 - val_loss: 0.5528 - val_accuracy: 0.7872\n",
            "Epoch 1032/2000\n",
            "3/3 [==============================] - 0s 40ms/step - loss: 0.7068 - accuracy: 0.8100 - val_loss: 0.5523 - val_accuracy: 0.8085\n",
            "Epoch 1033/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 0.4673 - accuracy: 0.8400 - val_loss: 0.5433 - val_accuracy: 0.7872\n",
            "Epoch 1034/2000\n",
            "3/3 [==============================] - 0s 56ms/step - loss: 0.7173 - accuracy: 0.7700 - val_loss: 0.5369 - val_accuracy: 0.8298\n",
            "Epoch 1035/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 0.7068 - accuracy: 0.7700 - val_loss: 0.5330 - val_accuracy: 0.8511\n",
            "Epoch 1036/2000\n",
            "3/3 [==============================] - 0s 53ms/step - loss: 0.5328 - accuracy: 0.8300 - val_loss: 0.5441 - val_accuracy: 0.8085\n",
            "Epoch 1037/2000\n",
            "3/3 [==============================] - 0s 45ms/step - loss: 0.7110 - accuracy: 0.7700 - val_loss: 0.5386 - val_accuracy: 0.8511\n",
            "Epoch 1038/2000\n",
            "3/3 [==============================] - 0s 40ms/step - loss: 0.8565 - accuracy: 0.8000 - val_loss: 0.5429 - val_accuracy: 0.8298\n",
            "Epoch 1039/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 0.5476 - accuracy: 0.8400 - val_loss: 0.5467 - val_accuracy: 0.8298\n",
            "Epoch 1040/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 0.6436 - accuracy: 0.8300 - val_loss: 0.5491 - val_accuracy: 0.8085\n",
            "Epoch 1041/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 0.6890 - accuracy: 0.8200 - val_loss: 0.5601 - val_accuracy: 0.8085\n",
            "Epoch 1042/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 0.8108 - accuracy: 0.7700 - val_loss: 0.5648 - val_accuracy: 0.8085\n",
            "Epoch 1043/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 0.5656 - accuracy: 0.8400 - val_loss: 0.5674 - val_accuracy: 0.8085\n",
            "Epoch 1044/2000\n",
            "3/3 [==============================] - 0s 45ms/step - loss: 0.6020 - accuracy: 0.8100 - val_loss: 0.5871 - val_accuracy: 0.8085\n",
            "Epoch 1045/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 0.6516 - accuracy: 0.7600 - val_loss: 0.5738 - val_accuracy: 0.8085\n",
            "Epoch 1046/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 0.5995 - accuracy: 0.7700 - val_loss: 0.5539 - val_accuracy: 0.8298\n",
            "Epoch 1047/2000\n",
            "3/3 [==============================] - 0s 47ms/step - loss: 0.5575 - accuracy: 0.7700 - val_loss: 0.5767 - val_accuracy: 0.7234\n",
            "Epoch 1048/2000\n",
            "3/3 [==============================] - 0s 33ms/step - loss: 0.6786 - accuracy: 0.7800 - val_loss: 0.6129 - val_accuracy: 0.7234\n",
            "Epoch 1049/2000\n",
            "3/3 [==============================] - 0s 33ms/step - loss: 0.6278 - accuracy: 0.7900 - val_loss: 0.6118 - val_accuracy: 0.7447\n",
            "Epoch 1050/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 0.6219 - accuracy: 0.8300 - val_loss: 0.5550 - val_accuracy: 0.7872\n",
            "Epoch 1051/2000\n",
            "3/3 [==============================] - 0s 45ms/step - loss: 0.5107 - accuracy: 0.8600 - val_loss: 0.5369 - val_accuracy: 0.8085\n",
            "Epoch 1052/2000\n",
            "3/3 [==============================] - 0s 33ms/step - loss: 0.5291 - accuracy: 0.8700 - val_loss: 0.5278 - val_accuracy: 0.8085\n",
            "Epoch 1053/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 0.5482 - accuracy: 0.8200 - val_loss: 0.5279 - val_accuracy: 0.8298\n",
            "Epoch 1054/2000\n",
            "3/3 [==============================] - 0s 54ms/step - loss: 0.5386 - accuracy: 0.8200 - val_loss: 0.5324 - val_accuracy: 0.7872\n",
            "Epoch 1055/2000\n",
            "3/3 [==============================] - 0s 52ms/step - loss: 0.6323 - accuracy: 0.7800 - val_loss: 0.5358 - val_accuracy: 0.7660\n",
            "Epoch 1056/2000\n",
            "3/3 [==============================] - 0s 34ms/step - loss: 0.5675 - accuracy: 0.8200 - val_loss: 0.5359 - val_accuracy: 0.8085\n",
            "Epoch 1057/2000\n",
            "3/3 [==============================] - 0s 44ms/step - loss: 0.6519 - accuracy: 0.8200 - val_loss: 0.5510 - val_accuracy: 0.8298\n",
            "Epoch 1058/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 0.5453 - accuracy: 0.8300 - val_loss: 0.5547 - val_accuracy: 0.8298\n",
            "Epoch 1059/2000\n",
            "3/3 [==============================] - 0s 45ms/step - loss: 0.7181 - accuracy: 0.7700 - val_loss: 0.5505 - val_accuracy: 0.7872\n",
            "Epoch 1060/2000\n",
            "3/3 [==============================] - 0s 47ms/step - loss: 0.7011 - accuracy: 0.8100 - val_loss: 0.5509 - val_accuracy: 0.7872\n",
            "Epoch 1061/2000\n",
            "3/3 [==============================] - 0s 38ms/step - loss: 0.5466 - accuracy: 0.8200 - val_loss: 0.5441 - val_accuracy: 0.8085\n",
            "Epoch 1062/2000\n",
            "3/3 [==============================] - 0s 45ms/step - loss: 0.8407 - accuracy: 0.7600 - val_loss: 0.5803 - val_accuracy: 0.7660\n",
            "Epoch 1063/2000\n",
            "3/3 [==============================] - 0s 52ms/step - loss: 0.5168 - accuracy: 0.8600 - val_loss: 0.6009 - val_accuracy: 0.7234\n",
            "Epoch 1064/2000\n",
            "3/3 [==============================] - 0s 34ms/step - loss: 0.6922 - accuracy: 0.7700 - val_loss: 0.5924 - val_accuracy: 0.7447\n",
            "Epoch 1065/2000\n",
            "3/3 [==============================] - 0s 46ms/step - loss: 0.5583 - accuracy: 0.7900 - val_loss: 0.6065 - val_accuracy: 0.7447\n",
            "Epoch 1066/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 0.6688 - accuracy: 0.7700 - val_loss: 0.6030 - val_accuracy: 0.7872\n",
            "Epoch 1067/2000\n",
            "3/3 [==============================] - 0s 57ms/step - loss: 0.4582 - accuracy: 0.8300 - val_loss: 0.5856 - val_accuracy: 0.7872\n",
            "Epoch 1068/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 0.6099 - accuracy: 0.8100 - val_loss: 0.5942 - val_accuracy: 0.7660\n",
            "Epoch 1069/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 0.6584 - accuracy: 0.6900 - val_loss: 0.6046 - val_accuracy: 0.7447\n",
            "Epoch 1070/2000\n",
            "3/3 [==============================] - 0s 34ms/step - loss: 0.5827 - accuracy: 0.8200 - val_loss: 0.6062 - val_accuracy: 0.7660\n",
            "Epoch 1071/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 0.6818 - accuracy: 0.7600 - val_loss: 0.6072 - val_accuracy: 0.7872\n",
            "Epoch 1072/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 0.5430 - accuracy: 0.8200 - val_loss: 0.6011 - val_accuracy: 0.8085\n",
            "Epoch 1073/2000\n",
            "3/3 [==============================] - 0s 45ms/step - loss: 0.5701 - accuracy: 0.8400 - val_loss: 0.5884 - val_accuracy: 0.7872\n",
            "Epoch 1074/2000\n",
            "3/3 [==============================] - 0s 75ms/step - loss: 0.5591 - accuracy: 0.8600 - val_loss: 0.5762 - val_accuracy: 0.8085\n",
            "Epoch 1075/2000\n",
            "3/3 [==============================] - 0s 68ms/step - loss: 0.6826 - accuracy: 0.8000 - val_loss: 0.5634 - val_accuracy: 0.8085\n",
            "Epoch 1076/2000\n",
            "3/3 [==============================] - 0s 54ms/step - loss: 0.8235 - accuracy: 0.7400 - val_loss: 0.5758 - val_accuracy: 0.8085\n",
            "Epoch 1077/2000\n",
            "3/3 [==============================] - 0s 77ms/step - loss: 0.5336 - accuracy: 0.8500 - val_loss: 0.5959 - val_accuracy: 0.8085\n",
            "Epoch 1078/2000\n",
            "3/3 [==============================] - 0s 76ms/step - loss: 0.5207 - accuracy: 0.9200 - val_loss: 0.5991 - val_accuracy: 0.8085\n",
            "Epoch 1079/2000\n",
            "3/3 [==============================] - 0s 54ms/step - loss: 0.6742 - accuracy: 0.8100 - val_loss: 0.5949 - val_accuracy: 0.7447\n",
            "Epoch 1080/2000\n",
            "3/3 [==============================] - 0s 73ms/step - loss: 0.6725 - accuracy: 0.7900 - val_loss: 0.5923 - val_accuracy: 0.7872\n",
            "Epoch 1081/2000\n",
            "3/3 [==============================] - 0s 63ms/step - loss: 0.8223 - accuracy: 0.6800 - val_loss: 0.6003 - val_accuracy: 0.7447\n",
            "Epoch 1082/2000\n",
            "3/3 [==============================] - 0s 53ms/step - loss: 0.4601 - accuracy: 0.8700 - val_loss: 0.6171 - val_accuracy: 0.7447\n",
            "Epoch 1083/2000\n",
            "3/3 [==============================] - 0s 54ms/step - loss: 0.6808 - accuracy: 0.7300 - val_loss: 0.6264 - val_accuracy: 0.7660\n",
            "Epoch 1084/2000\n",
            "3/3 [==============================] - 0s 50ms/step - loss: 0.6266 - accuracy: 0.8400 - val_loss: 0.6122 - val_accuracy: 0.7872\n",
            "Epoch 1085/2000\n",
            "3/3 [==============================] - 0s 33ms/step - loss: 0.5237 - accuracy: 0.8800 - val_loss: 0.6016 - val_accuracy: 0.7872\n",
            "Epoch 1086/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 0.7727 - accuracy: 0.7300 - val_loss: 0.5907 - val_accuracy: 0.7872\n",
            "Epoch 1087/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 0.5239 - accuracy: 0.8200 - val_loss: 0.5813 - val_accuracy: 0.7660\n",
            "Epoch 1088/2000\n",
            "3/3 [==============================] - 0s 44ms/step - loss: 0.6568 - accuracy: 0.8100 - val_loss: 0.5681 - val_accuracy: 0.7872\n",
            "Epoch 1089/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 0.5629 - accuracy: 0.8000 - val_loss: 0.5558 - val_accuracy: 0.8085\n",
            "Epoch 1090/2000\n",
            "3/3 [==============================] - 0s 47ms/step - loss: 0.7422 - accuracy: 0.7800 - val_loss: 0.5555 - val_accuracy: 0.7872\n",
            "Epoch 1091/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 0.7418 - accuracy: 0.7600 - val_loss: 0.5701 - val_accuracy: 0.7872\n",
            "Epoch 1092/2000\n",
            "3/3 [==============================] - 0s 33ms/step - loss: 0.5561 - accuracy: 0.8200 - val_loss: 0.5923 - val_accuracy: 0.7660\n",
            "Epoch 1093/2000\n",
            "3/3 [==============================] - 0s 45ms/step - loss: 0.5360 - accuracy: 0.7900 - val_loss: 0.6003 - val_accuracy: 0.8085\n",
            "Epoch 1094/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 0.6820 - accuracy: 0.7700 - val_loss: 0.6021 - val_accuracy: 0.7660\n",
            "Epoch 1095/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 0.6018 - accuracy: 0.8100 - val_loss: 0.6275 - val_accuracy: 0.7660\n",
            "Epoch 1096/2000\n",
            "3/3 [==============================] - 0s 34ms/step - loss: 0.6242 - accuracy: 0.7900 - val_loss: 0.6553 - val_accuracy: 0.7660\n",
            "Epoch 1097/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 0.6204 - accuracy: 0.8100 - val_loss: 0.6403 - val_accuracy: 0.7872\n",
            "Epoch 1098/2000\n",
            "3/3 [==============================] - 0s 33ms/step - loss: 0.5486 - accuracy: 0.8500 - val_loss: 0.5808 - val_accuracy: 0.7660\n",
            "Epoch 1099/2000\n",
            "3/3 [==============================] - 0s 33ms/step - loss: 0.5524 - accuracy: 0.8300 - val_loss: 0.5446 - val_accuracy: 0.8085\n",
            "Epoch 1100/2000\n",
            "3/3 [==============================] - 0s 60ms/step - loss: 0.7308 - accuracy: 0.7900 - val_loss: 0.5438 - val_accuracy: 0.8298\n",
            "Epoch 1101/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 0.5751 - accuracy: 0.8200 - val_loss: 0.5456 - val_accuracy: 0.8298\n",
            "Epoch 1102/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 0.4947 - accuracy: 0.8300 - val_loss: 0.5476 - val_accuracy: 0.8298\n",
            "Epoch 1103/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 0.4606 - accuracy: 0.8800 - val_loss: 0.5373 - val_accuracy: 0.8298\n",
            "Epoch 1104/2000\n",
            "3/3 [==============================] - 0s 40ms/step - loss: 0.4664 - accuracy: 0.8900 - val_loss: 0.5348 - val_accuracy: 0.8298\n",
            "Epoch 1105/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 0.5092 - accuracy: 0.8600 - val_loss: 0.5175 - val_accuracy: 0.8298\n",
            "Epoch 1106/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 0.6443 - accuracy: 0.8000 - val_loss: 0.4915 - val_accuracy: 0.8298\n",
            "Epoch 1107/2000\n",
            "3/3 [==============================] - 0s 48ms/step - loss: 0.6120 - accuracy: 0.7800 - val_loss: 0.4853 - val_accuracy: 0.8723\n",
            "Epoch 1108/2000\n",
            "3/3 [==============================] - 0s 33ms/step - loss: 0.6864 - accuracy: 0.7400 - val_loss: 0.4839 - val_accuracy: 0.8936\n",
            "Epoch 1109/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 0.6512 - accuracy: 0.8400 - val_loss: 0.4915 - val_accuracy: 0.8723\n",
            "Epoch 1110/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 0.6450 - accuracy: 0.7700 - val_loss: 0.4940 - val_accuracy: 0.8298\n",
            "Epoch 1111/2000\n",
            "3/3 [==============================] - 0s 51ms/step - loss: 0.5796 - accuracy: 0.8300 - val_loss: 0.5221 - val_accuracy: 0.7872\n",
            "Epoch 1112/2000\n",
            "3/3 [==============================] - 0s 53ms/step - loss: 0.4887 - accuracy: 0.8500 - val_loss: 0.5551 - val_accuracy: 0.7872\n",
            "Epoch 1113/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 0.6732 - accuracy: 0.7900 - val_loss: 0.5620 - val_accuracy: 0.7660\n",
            "Epoch 1114/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 0.7668 - accuracy: 0.7400 - val_loss: 0.5273 - val_accuracy: 0.8085\n",
            "Epoch 1115/2000\n",
            "3/3 [==============================] - 0s 34ms/step - loss: 0.5549 - accuracy: 0.8500 - val_loss: 0.5041 - val_accuracy: 0.7872\n",
            "Epoch 1116/2000\n",
            "3/3 [==============================] - 0s 53ms/step - loss: 0.5779 - accuracy: 0.7900 - val_loss: 0.5006 - val_accuracy: 0.8511\n",
            "Epoch 1117/2000\n",
            "3/3 [==============================] - 0s 45ms/step - loss: 0.6321 - accuracy: 0.8000 - val_loss: 0.5028 - val_accuracy: 0.8511\n",
            "Epoch 1118/2000\n",
            "3/3 [==============================] - 0s 40ms/step - loss: 0.5872 - accuracy: 0.7600 - val_loss: 0.5029 - val_accuracy: 0.8298\n",
            "Epoch 1119/2000\n",
            "3/3 [==============================] - 0s 55ms/step - loss: 0.5667 - accuracy: 0.8200 - val_loss: 0.5252 - val_accuracy: 0.8085\n",
            "Epoch 1120/2000\n",
            "3/3 [==============================] - 0s 34ms/step - loss: 0.7154 - accuracy: 0.7600 - val_loss: 0.5382 - val_accuracy: 0.7872\n",
            "Epoch 1121/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 0.6817 - accuracy: 0.8100 - val_loss: 0.5600 - val_accuracy: 0.7872\n",
            "Epoch 1122/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 0.6241 - accuracy: 0.7900 - val_loss: 0.5939 - val_accuracy: 0.8085\n",
            "Epoch 1123/2000\n",
            "3/3 [==============================] - 0s 39ms/step - loss: 0.7261 - accuracy: 0.7200 - val_loss: 0.6126 - val_accuracy: 0.8085\n",
            "Epoch 1124/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 0.5270 - accuracy: 0.8500 - val_loss: 0.6071 - val_accuracy: 0.8085\n",
            "Epoch 1125/2000\n",
            "3/3 [==============================] - 0s 48ms/step - loss: 0.6757 - accuracy: 0.8000 - val_loss: 0.5897 - val_accuracy: 0.8085\n",
            "Epoch 1126/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 0.5429 - accuracy: 0.8200 - val_loss: 0.5521 - val_accuracy: 0.7872\n",
            "Epoch 1127/2000\n",
            "3/3 [==============================] - 0s 44ms/step - loss: 0.5134 - accuracy: 0.8400 - val_loss: 0.5265 - val_accuracy: 0.8298\n",
            "Epoch 1128/2000\n",
            "3/3 [==============================] - 0s 33ms/step - loss: 0.7509 - accuracy: 0.7800 - val_loss: 0.5115 - val_accuracy: 0.8298\n",
            "Epoch 1129/2000\n",
            "3/3 [==============================] - 0s 57ms/step - loss: 0.5878 - accuracy: 0.7800 - val_loss: 0.5050 - val_accuracy: 0.8085\n",
            "Epoch 1130/2000\n",
            "3/3 [==============================] - 0s 40ms/step - loss: 0.5768 - accuracy: 0.8200 - val_loss: 0.4935 - val_accuracy: 0.8085\n",
            "Epoch 1131/2000\n",
            "3/3 [==============================] - 0s 51ms/step - loss: 0.5198 - accuracy: 0.8300 - val_loss: 0.5378 - val_accuracy: 0.8511\n",
            "Epoch 1132/2000\n",
            "3/3 [==============================] - 0s 46ms/step - loss: 0.8312 - accuracy: 0.8100 - val_loss: 0.5727 - val_accuracy: 0.8298\n",
            "Epoch 1133/2000\n",
            "3/3 [==============================] - 0s 49ms/step - loss: 0.6439 - accuracy: 0.7800 - val_loss: 0.5759 - val_accuracy: 0.8085\n",
            "Epoch 1134/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 0.4853 - accuracy: 0.8400 - val_loss: 0.5620 - val_accuracy: 0.8085\n",
            "Epoch 1135/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 0.6618 - accuracy: 0.8000 - val_loss: 0.5593 - val_accuracy: 0.8298\n",
            "Epoch 1136/2000\n",
            "3/3 [==============================] - 0s 33ms/step - loss: 0.5583 - accuracy: 0.7500 - val_loss: 0.6110 - val_accuracy: 0.7872\n",
            "Epoch 1137/2000\n",
            "3/3 [==============================] - 0s 33ms/step - loss: 0.6774 - accuracy: 0.7800 - val_loss: 0.6416 - val_accuracy: 0.7872\n",
            "Epoch 1138/2000\n",
            "3/3 [==============================] - 0s 34ms/step - loss: 0.6590 - accuracy: 0.7700 - val_loss: 0.6077 - val_accuracy: 0.7872\n",
            "Epoch 1139/2000\n",
            "3/3 [==============================] - 0s 34ms/step - loss: 0.7164 - accuracy: 0.8300 - val_loss: 0.6021 - val_accuracy: 0.7660\n",
            "Epoch 1140/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 0.6733 - accuracy: 0.8100 - val_loss: 0.6132 - val_accuracy: 0.7660\n",
            "Epoch 1141/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 0.5555 - accuracy: 0.8300 - val_loss: 0.6093 - val_accuracy: 0.8085\n",
            "Epoch 1142/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 0.6444 - accuracy: 0.7900 - val_loss: 0.5868 - val_accuracy: 0.8085\n",
            "Epoch 1143/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 0.5870 - accuracy: 0.7800 - val_loss: 0.5636 - val_accuracy: 0.8723\n",
            "Epoch 1144/2000\n",
            "3/3 [==============================] - 0s 38ms/step - loss: 0.5563 - accuracy: 0.8500 - val_loss: 0.5476 - val_accuracy: 0.8298\n",
            "Epoch 1145/2000\n",
            "3/3 [==============================] - 0s 46ms/step - loss: 0.6837 - accuracy: 0.7900 - val_loss: 0.5409 - val_accuracy: 0.8511\n",
            "Epoch 1146/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 0.6576 - accuracy: 0.8200 - val_loss: 0.5483 - val_accuracy: 0.8298\n",
            "Epoch 1147/2000\n",
            "3/3 [==============================] - 0s 44ms/step - loss: 0.5468 - accuracy: 0.8500 - val_loss: 0.5505 - val_accuracy: 0.8085\n",
            "Epoch 1148/2000\n",
            "3/3 [==============================] - 0s 49ms/step - loss: 0.4999 - accuracy: 0.8300 - val_loss: 0.5558 - val_accuracy: 0.7872\n",
            "Epoch 1149/2000\n",
            "3/3 [==============================] - 0s 55ms/step - loss: 0.6570 - accuracy: 0.7700 - val_loss: 0.5545 - val_accuracy: 0.7872\n",
            "Epoch 1150/2000\n",
            "3/3 [==============================] - 0s 66ms/step - loss: 0.6239 - accuracy: 0.8200 - val_loss: 0.5512 - val_accuracy: 0.7872\n",
            "Epoch 1151/2000\n",
            "3/3 [==============================] - 0s 61ms/step - loss: 0.4597 - accuracy: 0.8200 - val_loss: 0.5414 - val_accuracy: 0.7872\n",
            "Epoch 1152/2000\n",
            "3/3 [==============================] - 0s 66ms/step - loss: 0.5552 - accuracy: 0.8300 - val_loss: 0.5397 - val_accuracy: 0.7447\n",
            "Epoch 1153/2000\n",
            "3/3 [==============================] - 0s 75ms/step - loss: 0.5322 - accuracy: 0.8300 - val_loss: 0.5585 - val_accuracy: 0.7660\n",
            "Epoch 1154/2000\n",
            "3/3 [==============================] - 0s 71ms/step - loss: 0.5850 - accuracy: 0.8500 - val_loss: 0.5827 - val_accuracy: 0.7660\n",
            "Epoch 1155/2000\n",
            "3/3 [==============================] - 0s 56ms/step - loss: 0.5043 - accuracy: 0.8100 - val_loss: 0.5902 - val_accuracy: 0.7872\n",
            "Epoch 1156/2000\n",
            "3/3 [==============================] - 0s 75ms/step - loss: 0.5695 - accuracy: 0.7900 - val_loss: 0.5579 - val_accuracy: 0.8085\n",
            "Epoch 1157/2000\n",
            "3/3 [==============================] - 0s 74ms/step - loss: 0.6859 - accuracy: 0.7800 - val_loss: 0.5282 - val_accuracy: 0.8085\n",
            "Epoch 1158/2000\n",
            "3/3 [==============================] - 0s 52ms/step - loss: 0.6082 - accuracy: 0.8200 - val_loss: 0.5056 - val_accuracy: 0.8085\n",
            "Epoch 1159/2000\n",
            "3/3 [==============================] - 0s 68ms/step - loss: 0.6292 - accuracy: 0.7600 - val_loss: 0.4982 - val_accuracy: 0.8085\n",
            "Epoch 1160/2000\n",
            "3/3 [==============================] - 0s 58ms/step - loss: 0.6332 - accuracy: 0.8600 - val_loss: 0.4947 - val_accuracy: 0.8085\n",
            "Epoch 1161/2000\n",
            "3/3 [==============================] - 0s 44ms/step - loss: 0.5356 - accuracy: 0.8000 - val_loss: 0.5047 - val_accuracy: 0.8085\n",
            "Epoch 1162/2000\n",
            "3/3 [==============================] - 0s 34ms/step - loss: 0.5327 - accuracy: 0.8200 - val_loss: 0.5270 - val_accuracy: 0.8085\n",
            "Epoch 1163/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 0.5080 - accuracy: 0.8400 - val_loss: 0.5574 - val_accuracy: 0.7872\n",
            "Epoch 1164/2000\n",
            "3/3 [==============================] - 0s 46ms/step - loss: 0.9237 - accuracy: 0.6800 - val_loss: 0.5694 - val_accuracy: 0.8085\n",
            "Epoch 1165/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 0.5011 - accuracy: 0.8400 - val_loss: 0.5412 - val_accuracy: 0.8085\n",
            "Epoch 1166/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 0.6417 - accuracy: 0.8000 - val_loss: 0.5220 - val_accuracy: 0.8298\n",
            "Epoch 1167/2000\n",
            "3/3 [==============================] - 0s 54ms/step - loss: 0.6028 - accuracy: 0.8300 - val_loss: 0.5310 - val_accuracy: 0.8085\n",
            "Epoch 1168/2000\n",
            "3/3 [==============================] - 0s 47ms/step - loss: 0.5404 - accuracy: 0.8100 - val_loss: 0.5366 - val_accuracy: 0.7872\n",
            "Epoch 1169/2000\n",
            "3/3 [==============================] - 0s 44ms/step - loss: 0.4345 - accuracy: 0.8900 - val_loss: 0.5416 - val_accuracy: 0.7872\n",
            "Epoch 1170/2000\n",
            "3/3 [==============================] - 0s 59ms/step - loss: 0.8705 - accuracy: 0.7800 - val_loss: 0.5461 - val_accuracy: 0.8085\n",
            "Epoch 1171/2000\n",
            "3/3 [==============================] - 0s 44ms/step - loss: 0.4452 - accuracy: 0.9000 - val_loss: 0.5495 - val_accuracy: 0.8298\n",
            "Epoch 1172/2000\n",
            "3/3 [==============================] - 0s 51ms/step - loss: 0.5331 - accuracy: 0.7900 - val_loss: 0.5498 - val_accuracy: 0.8298\n",
            "Epoch 1173/2000\n",
            "3/3 [==============================] - 0s 44ms/step - loss: 0.5193 - accuracy: 0.8100 - val_loss: 0.5427 - val_accuracy: 0.8298\n",
            "Epoch 1174/2000\n",
            "3/3 [==============================] - 0s 44ms/step - loss: 0.6553 - accuracy: 0.7600 - val_loss: 0.5224 - val_accuracy: 0.8511\n",
            "Epoch 1175/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 0.5937 - accuracy: 0.7700 - val_loss: 0.4995 - val_accuracy: 0.8511\n",
            "Epoch 1176/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 0.4336 - accuracy: 0.9100 - val_loss: 0.4984 - val_accuracy: 0.8298\n",
            "Epoch 1177/2000\n",
            "3/3 [==============================] - 0s 40ms/step - loss: 0.4888 - accuracy: 0.8400 - val_loss: 0.5116 - val_accuracy: 0.8298\n",
            "Epoch 1178/2000\n",
            "3/3 [==============================] - 0s 55ms/step - loss: 0.8182 - accuracy: 0.7500 - val_loss: 0.5208 - val_accuracy: 0.8511\n",
            "Epoch 1179/2000\n",
            "3/3 [==============================] - 0s 52ms/step - loss: 0.5054 - accuracy: 0.8300 - val_loss: 0.5232 - val_accuracy: 0.8511\n",
            "Epoch 1180/2000\n",
            "3/3 [==============================] - 0s 34ms/step - loss: 0.6219 - accuracy: 0.7900 - val_loss: 0.5190 - val_accuracy: 0.8298\n",
            "Epoch 1181/2000\n",
            "3/3 [==============================] - 0s 54ms/step - loss: 0.5376 - accuracy: 0.8100 - val_loss: 0.5190 - val_accuracy: 0.8511\n",
            "Epoch 1182/2000\n",
            "3/3 [==============================] - 0s 34ms/step - loss: 0.6341 - accuracy: 0.7900 - val_loss: 0.4971 - val_accuracy: 0.8723\n",
            "Epoch 1183/2000\n",
            "3/3 [==============================] - 0s 44ms/step - loss: 0.4989 - accuracy: 0.8300 - val_loss: 0.4805 - val_accuracy: 0.8511\n",
            "Epoch 1184/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 0.5856 - accuracy: 0.8000 - val_loss: 0.4805 - val_accuracy: 0.8511\n",
            "Epoch 1185/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 0.7212 - accuracy: 0.7900 - val_loss: 0.4853 - val_accuracy: 0.8085\n",
            "Epoch 1186/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 0.4678 - accuracy: 0.8500 - val_loss: 0.4966 - val_accuracy: 0.8298\n",
            "Epoch 1187/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 0.5253 - accuracy: 0.8400 - val_loss: 0.4974 - val_accuracy: 0.8298\n",
            "Epoch 1188/2000\n",
            "3/3 [==============================] - 0s 62ms/step - loss: 0.7513 - accuracy: 0.7900 - val_loss: 0.4935 - val_accuracy: 0.8298\n",
            "Epoch 1189/2000\n",
            "3/3 [==============================] - 0s 51ms/step - loss: 0.5424 - accuracy: 0.8500 - val_loss: 0.4852 - val_accuracy: 0.8298\n",
            "Epoch 1190/2000\n",
            "3/3 [==============================] - 0s 52ms/step - loss: 0.5037 - accuracy: 0.8500 - val_loss: 0.4772 - val_accuracy: 0.8511\n",
            "Epoch 1191/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 0.6196 - accuracy: 0.8100 - val_loss: 0.4810 - val_accuracy: 0.8723\n",
            "Epoch 1192/2000\n",
            "3/3 [==============================] - 0s 34ms/step - loss: 0.4700 - accuracy: 0.8700 - val_loss: 0.4981 - val_accuracy: 0.8298\n",
            "Epoch 1193/2000\n",
            "3/3 [==============================] - 0s 56ms/step - loss: 0.5157 - accuracy: 0.8400 - val_loss: 0.5061 - val_accuracy: 0.8511\n",
            "Epoch 1194/2000\n",
            "3/3 [==============================] - 0s 34ms/step - loss: 0.5900 - accuracy: 0.7900 - val_loss: 0.5207 - val_accuracy: 0.8723\n",
            "Epoch 1195/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 0.4983 - accuracy: 0.8300 - val_loss: 0.5463 - val_accuracy: 0.8298\n",
            "Epoch 1196/2000\n",
            "3/3 [==============================] - 0s 45ms/step - loss: 0.6248 - accuracy: 0.8200 - val_loss: 0.5364 - val_accuracy: 0.8298\n",
            "Epoch 1197/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 0.7501 - accuracy: 0.7600 - val_loss: 0.5487 - val_accuracy: 0.8298\n",
            "Epoch 1198/2000\n",
            "3/3 [==============================] - 0s 48ms/step - loss: 0.4992 - accuracy: 0.8400 - val_loss: 0.5689 - val_accuracy: 0.8298\n",
            "Epoch 1199/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 0.5795 - accuracy: 0.7900 - val_loss: 0.5801 - val_accuracy: 0.8085\n",
            "Epoch 1200/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 0.6692 - accuracy: 0.8100 - val_loss: 0.5724 - val_accuracy: 0.8085\n",
            "Epoch 1201/2000\n",
            "3/3 [==============================] - 0s 47ms/step - loss: 0.6200 - accuracy: 0.8200 - val_loss: 0.5734 - val_accuracy: 0.7872\n",
            "Epoch 1202/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 0.5456 - accuracy: 0.8000 - val_loss: 0.5608 - val_accuracy: 0.8298\n",
            "Epoch 1203/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 0.5806 - accuracy: 0.8100 - val_loss: 0.5546 - val_accuracy: 0.8298\n",
            "Epoch 1204/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 0.5398 - accuracy: 0.8300 - val_loss: 0.5540 - val_accuracy: 0.8085\n",
            "Epoch 1205/2000\n",
            "3/3 [==============================] - 0s 45ms/step - loss: 0.5090 - accuracy: 0.8900 - val_loss: 0.5425 - val_accuracy: 0.8298\n",
            "Epoch 1206/2000\n",
            "3/3 [==============================] - 0s 55ms/step - loss: 0.4753 - accuracy: 0.8900 - val_loss: 0.5275 - val_accuracy: 0.8298\n",
            "Epoch 1207/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 0.5477 - accuracy: 0.8300 - val_loss: 0.5238 - val_accuracy: 0.8085\n",
            "Epoch 1208/2000\n",
            "3/3 [==============================] - 0s 54ms/step - loss: 0.5338 - accuracy: 0.8200 - val_loss: 0.5359 - val_accuracy: 0.8085\n",
            "Epoch 1209/2000\n",
            "3/3 [==============================] - 0s 45ms/step - loss: 0.5029 - accuracy: 0.8700 - val_loss: 0.5687 - val_accuracy: 0.8085\n",
            "Epoch 1210/2000\n",
            "3/3 [==============================] - 0s 56ms/step - loss: 0.5699 - accuracy: 0.8400 - val_loss: 0.5818 - val_accuracy: 0.8085\n",
            "Epoch 1211/2000\n",
            "3/3 [==============================] - 0s 34ms/step - loss: 0.6401 - accuracy: 0.7900 - val_loss: 0.5722 - val_accuracy: 0.8085\n",
            "Epoch 1212/2000\n",
            "3/3 [==============================] - 0s 54ms/step - loss: 0.5809 - accuracy: 0.8000 - val_loss: 0.5373 - val_accuracy: 0.8298\n",
            "Epoch 1213/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 0.6043 - accuracy: 0.8000 - val_loss: 0.5208 - val_accuracy: 0.8298\n",
            "Epoch 1214/2000\n",
            "3/3 [==============================] - 0s 49ms/step - loss: 0.6920 - accuracy: 0.8300 - val_loss: 0.5191 - val_accuracy: 0.8085\n",
            "Epoch 1215/2000\n",
            "3/3 [==============================] - 0s 55ms/step - loss: 0.4972 - accuracy: 0.8200 - val_loss: 0.5093 - val_accuracy: 0.8085\n",
            "Epoch 1216/2000\n",
            "3/3 [==============================] - 0s 48ms/step - loss: 0.7298 - accuracy: 0.7700 - val_loss: 0.4917 - val_accuracy: 0.8511\n",
            "Epoch 1217/2000\n",
            "3/3 [==============================] - 0s 40ms/step - loss: 0.6163 - accuracy: 0.7700 - val_loss: 0.4885 - val_accuracy: 0.8511\n",
            "Epoch 1218/2000\n",
            "3/3 [==============================] - 0s 73ms/step - loss: 0.4442 - accuracy: 0.8800 - val_loss: 0.5061 - val_accuracy: 0.8298\n",
            "Epoch 1219/2000\n",
            "3/3 [==============================] - 0s 75ms/step - loss: 0.5206 - accuracy: 0.8600 - val_loss: 0.5292 - val_accuracy: 0.7872\n",
            "Epoch 1220/2000\n",
            "3/3 [==============================] - 0s 74ms/step - loss: 0.6389 - accuracy: 0.7800 - val_loss: 0.5527 - val_accuracy: 0.8085\n",
            "Epoch 1221/2000\n",
            "3/3 [==============================] - 0s 46ms/step - loss: 0.5162 - accuracy: 0.8400 - val_loss: 0.5704 - val_accuracy: 0.8085\n",
            "Epoch 1222/2000\n",
            "3/3 [==============================] - 0s 62ms/step - loss: 0.7383 - accuracy: 0.7300 - val_loss: 0.5580 - val_accuracy: 0.7872\n",
            "Epoch 1223/2000\n",
            "3/3 [==============================] - 0s 59ms/step - loss: 0.6728 - accuracy: 0.8000 - val_loss: 0.5604 - val_accuracy: 0.7872\n",
            "Epoch 1224/2000\n",
            "3/3 [==============================] - 0s 89ms/step - loss: 0.5313 - accuracy: 0.8000 - val_loss: 0.5551 - val_accuracy: 0.8085\n",
            "Epoch 1225/2000\n",
            "3/3 [==============================] - 0s 54ms/step - loss: 0.5953 - accuracy: 0.7500 - val_loss: 0.5452 - val_accuracy: 0.8085\n",
            "Epoch 1226/2000\n",
            "3/3 [==============================] - 0s 60ms/step - loss: 0.5830 - accuracy: 0.8200 - val_loss: 0.5506 - val_accuracy: 0.8298\n",
            "Epoch 1227/2000\n",
            "3/3 [==============================] - 0s 70ms/step - loss: 0.6486 - accuracy: 0.8200 - val_loss: 0.5573 - val_accuracy: 0.8085\n",
            "Epoch 1228/2000\n",
            "3/3 [==============================] - 0s 74ms/step - loss: 0.4833 - accuracy: 0.8500 - val_loss: 0.5533 - val_accuracy: 0.8085\n",
            "Epoch 1229/2000\n",
            "3/3 [==============================] - 0s 54ms/step - loss: 0.4557 - accuracy: 0.8800 - val_loss: 0.5395 - val_accuracy: 0.8085\n",
            "Epoch 1230/2000\n",
            "3/3 [==============================] - 0s 72ms/step - loss: 0.6412 - accuracy: 0.8000 - val_loss: 0.5299 - val_accuracy: 0.8085\n",
            "Epoch 1231/2000\n",
            "3/3 [==============================] - 0s 40ms/step - loss: 0.5270 - accuracy: 0.8200 - val_loss: 0.5178 - val_accuracy: 0.8085\n",
            "Epoch 1232/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 0.5298 - accuracy: 0.8500 - val_loss: 0.5281 - val_accuracy: 0.7872\n",
            "Epoch 1233/2000\n",
            "3/3 [==============================] - 0s 48ms/step - loss: 0.6444 - accuracy: 0.8200 - val_loss: 0.5462 - val_accuracy: 0.7660\n",
            "Epoch 1234/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 0.5498 - accuracy: 0.8100 - val_loss: 0.5532 - val_accuracy: 0.7660\n",
            "Epoch 1235/2000\n",
            "3/3 [==============================] - 0s 38ms/step - loss: 0.5471 - accuracy: 0.8400 - val_loss: 0.5740 - val_accuracy: 0.7447\n",
            "Epoch 1236/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 0.5444 - accuracy: 0.8100 - val_loss: 0.5668 - val_accuracy: 0.7872\n",
            "Epoch 1237/2000\n",
            "3/3 [==============================] - 0s 47ms/step - loss: 0.4260 - accuracy: 0.8700 - val_loss: 0.5714 - val_accuracy: 0.7872\n",
            "Epoch 1238/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 0.5473 - accuracy: 0.8400 - val_loss: 0.5731 - val_accuracy: 0.8085\n",
            "Epoch 1239/2000\n",
            "3/3 [==============================] - 0s 48ms/step - loss: 0.5779 - accuracy: 0.7600 - val_loss: 0.5446 - val_accuracy: 0.7872\n",
            "Epoch 1240/2000\n",
            "3/3 [==============================] - 0s 34ms/step - loss: 0.6217 - accuracy: 0.8000 - val_loss: 0.5257 - val_accuracy: 0.8085\n",
            "Epoch 1241/2000\n",
            "3/3 [==============================] - 0s 40ms/step - loss: 0.6587 - accuracy: 0.7800 - val_loss: 0.5281 - val_accuracy: 0.7872\n",
            "Epoch 1242/2000\n",
            "3/3 [==============================] - 0s 52ms/step - loss: 0.6520 - accuracy: 0.7600 - val_loss: 0.5885 - val_accuracy: 0.7660\n",
            "Epoch 1243/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 0.5416 - accuracy: 0.7600 - val_loss: 0.6004 - val_accuracy: 0.7872\n",
            "Epoch 1244/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 0.6827 - accuracy: 0.8000 - val_loss: 0.5457 - val_accuracy: 0.7660\n",
            "Epoch 1245/2000\n",
            "3/3 [==============================] - 0s 44ms/step - loss: 0.5457 - accuracy: 0.8000 - val_loss: 0.5057 - val_accuracy: 0.8085\n",
            "Epoch 1246/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 0.5750 - accuracy: 0.8600 - val_loss: 0.5105 - val_accuracy: 0.8085\n",
            "Epoch 1247/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 0.4719 - accuracy: 0.8400 - val_loss: 0.4960 - val_accuracy: 0.7660\n",
            "Epoch 1248/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 0.6809 - accuracy: 0.8100 - val_loss: 0.4733 - val_accuracy: 0.7872\n",
            "Epoch 1249/2000\n",
            "3/3 [==============================] - 0s 52ms/step - loss: 0.4701 - accuracy: 0.8500 - val_loss: 0.4700 - val_accuracy: 0.8085\n",
            "Epoch 1250/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 0.4718 - accuracy: 0.8300 - val_loss: 0.4811 - val_accuracy: 0.8085\n",
            "Epoch 1251/2000\n",
            "3/3 [==============================] - 0s 33ms/step - loss: 0.5616 - accuracy: 0.8100 - val_loss: 0.4916 - val_accuracy: 0.7660\n",
            "Epoch 1252/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 0.5348 - accuracy: 0.8400 - val_loss: 0.4924 - val_accuracy: 0.7660\n",
            "Epoch 1253/2000\n",
            "3/3 [==============================] - 0s 45ms/step - loss: 0.6409 - accuracy: 0.8100 - val_loss: 0.5012 - val_accuracy: 0.8085\n",
            "Epoch 1254/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 0.5207 - accuracy: 0.8400 - val_loss: 0.5115 - val_accuracy: 0.8298\n",
            "Epoch 1255/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 0.4886 - accuracy: 0.8300 - val_loss: 0.5276 - val_accuracy: 0.8085\n",
            "Epoch 1256/2000\n",
            "3/3 [==============================] - 0s 49ms/step - loss: 0.6341 - accuracy: 0.8200 - val_loss: 0.5371 - val_accuracy: 0.7872\n",
            "Epoch 1257/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 0.5249 - accuracy: 0.8600 - val_loss: 0.5597 - val_accuracy: 0.7660\n",
            "Epoch 1258/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 0.5973 - accuracy: 0.7900 - val_loss: 0.6084 - val_accuracy: 0.7660\n",
            "Epoch 1259/2000\n",
            "3/3 [==============================] - 0s 34ms/step - loss: 0.5770 - accuracy: 0.8200 - val_loss: 0.6175 - val_accuracy: 0.7872\n",
            "Epoch 1260/2000\n",
            "3/3 [==============================] - 0s 56ms/step - loss: 0.6090 - accuracy: 0.8100 - val_loss: 0.5986 - val_accuracy: 0.7872\n",
            "Epoch 1261/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 0.5693 - accuracy: 0.8100 - val_loss: 0.5722 - val_accuracy: 0.7872\n",
            "Epoch 1262/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 0.5818 - accuracy: 0.8000 - val_loss: 0.5593 - val_accuracy: 0.7872\n",
            "Epoch 1263/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 0.4195 - accuracy: 0.8800 - val_loss: 0.5779 - val_accuracy: 0.7660\n",
            "Epoch 1264/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 0.6003 - accuracy: 0.8000 - val_loss: 0.5909 - val_accuracy: 0.7872\n",
            "Epoch 1265/2000\n",
            "3/3 [==============================] - 0s 53ms/step - loss: 0.5120 - accuracy: 0.8600 - val_loss: 0.5867 - val_accuracy: 0.7872\n",
            "Epoch 1266/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 0.6037 - accuracy: 0.8300 - val_loss: 0.5798 - val_accuracy: 0.8298\n",
            "Epoch 1267/2000\n",
            "3/3 [==============================] - 0s 53ms/step - loss: 0.6242 - accuracy: 0.8600 - val_loss: 0.6010 - val_accuracy: 0.8085\n",
            "Epoch 1268/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 0.4878 - accuracy: 0.8200 - val_loss: 0.6073 - val_accuracy: 0.8085\n",
            "Epoch 1269/2000\n",
            "3/3 [==============================] - 0s 34ms/step - loss: 0.6449 - accuracy: 0.8000 - val_loss: 0.5870 - val_accuracy: 0.8298\n",
            "Epoch 1270/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 0.5082 - accuracy: 0.8300 - val_loss: 0.5518 - val_accuracy: 0.8298\n",
            "Epoch 1271/2000\n",
            "3/3 [==============================] - 0s 57ms/step - loss: 0.4856 - accuracy: 0.8500 - val_loss: 0.5339 - val_accuracy: 0.8511\n",
            "Epoch 1272/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 0.5030 - accuracy: 0.8300 - val_loss: 0.5146 - val_accuracy: 0.8511\n",
            "Epoch 1273/2000\n",
            "3/3 [==============================] - 0s 45ms/step - loss: 0.6480 - accuracy: 0.8000 - val_loss: 0.5018 - val_accuracy: 0.8511\n",
            "Epoch 1274/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 0.5744 - accuracy: 0.8100 - val_loss: 0.5023 - val_accuracy: 0.7872\n",
            "Epoch 1275/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 0.8719 - accuracy: 0.7200 - val_loss: 0.5115 - val_accuracy: 0.7872\n",
            "Epoch 1276/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 0.6036 - accuracy: 0.8200 - val_loss: 0.5233 - val_accuracy: 0.7447\n",
            "Epoch 1277/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 0.6233 - accuracy: 0.8100 - val_loss: 0.5350 - val_accuracy: 0.7872\n",
            "Epoch 1278/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 0.5167 - accuracy: 0.8500 - val_loss: 0.5373 - val_accuracy: 0.8085\n",
            "Epoch 1279/2000\n",
            "3/3 [==============================] - 0s 54ms/step - loss: 0.5142 - accuracy: 0.8600 - val_loss: 0.5423 - val_accuracy: 0.8085\n",
            "Epoch 1280/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 0.5275 - accuracy: 0.8500 - val_loss: 0.5582 - val_accuracy: 0.8085\n",
            "Epoch 1281/2000\n",
            "3/3 [==============================] - 0s 50ms/step - loss: 0.7185 - accuracy: 0.7600 - val_loss: 0.5842 - val_accuracy: 0.7872\n",
            "Epoch 1282/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 0.6050 - accuracy: 0.8200 - val_loss: 0.5998 - val_accuracy: 0.7447\n",
            "Epoch 1283/2000\n",
            "3/3 [==============================] - 0s 44ms/step - loss: 0.5060 - accuracy: 0.8500 - val_loss: 0.6302 - val_accuracy: 0.7447\n",
            "Epoch 1284/2000\n",
            "3/3 [==============================] - 0s 55ms/step - loss: 0.4792 - accuracy: 0.8100 - val_loss: 0.6533 - val_accuracy: 0.7447\n",
            "Epoch 1285/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 0.7045 - accuracy: 0.7500 - val_loss: 0.6393 - val_accuracy: 0.7447\n",
            "Epoch 1286/2000\n",
            "3/3 [==============================] - 0s 45ms/step - loss: 0.5702 - accuracy: 0.7800 - val_loss: 0.6296 - val_accuracy: 0.7234\n",
            "Epoch 1287/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 0.4482 - accuracy: 0.8500 - val_loss: 0.6184 - val_accuracy: 0.7021\n",
            "Epoch 1288/2000\n",
            "3/3 [==============================] - 0s 34ms/step - loss: 0.4428 - accuracy: 0.8400 - val_loss: 0.6284 - val_accuracy: 0.7660\n",
            "Epoch 1289/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 0.6964 - accuracy: 0.8100 - val_loss: 0.6224 - val_accuracy: 0.7660\n",
            "Epoch 1290/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 0.5110 - accuracy: 0.8200 - val_loss: 0.6069 - val_accuracy: 0.7872\n",
            "Epoch 1291/2000\n",
            "3/3 [==============================] - 0s 46ms/step - loss: 0.7653 - accuracy: 0.7300 - val_loss: 0.5924 - val_accuracy: 0.7660\n",
            "Epoch 1292/2000\n",
            "3/3 [==============================] - 0s 44ms/step - loss: 0.4647 - accuracy: 0.8700 - val_loss: 0.5783 - val_accuracy: 0.7660\n",
            "Epoch 1293/2000\n",
            "3/3 [==============================] - 0s 34ms/step - loss: 0.8259 - accuracy: 0.8300 - val_loss: 0.5806 - val_accuracy: 0.7447\n",
            "Epoch 1294/2000\n",
            "3/3 [==============================] - 0s 44ms/step - loss: 0.5399 - accuracy: 0.8100 - val_loss: 0.6097 - val_accuracy: 0.7660\n",
            "Epoch 1295/2000\n",
            "3/3 [==============================] - 0s 62ms/step - loss: 0.5673 - accuracy: 0.8200 - val_loss: 0.6266 - val_accuracy: 0.7234\n",
            "Epoch 1296/2000\n",
            "3/3 [==============================] - 0s 66ms/step - loss: 0.4912 - accuracy: 0.8300 - val_loss: 0.6325 - val_accuracy: 0.7660\n",
            "Epoch 1297/2000\n",
            "3/3 [==============================] - 0s 68ms/step - loss: 0.4520 - accuracy: 0.8500 - val_loss: 0.6323 - val_accuracy: 0.7660\n",
            "Epoch 1298/2000\n",
            "3/3 [==============================] - 0s 77ms/step - loss: 0.4660 - accuracy: 0.8200 - val_loss: 0.6177 - val_accuracy: 0.7660\n",
            "Epoch 1299/2000\n",
            "3/3 [==============================] - 0s 75ms/step - loss: 0.5753 - accuracy: 0.8000 - val_loss: 0.6036 - val_accuracy: 0.7660\n",
            "Epoch 1300/2000\n",
            "3/3 [==============================] - 0s 66ms/step - loss: 0.6043 - accuracy: 0.8200 - val_loss: 0.5876 - val_accuracy: 0.7660\n",
            "Epoch 1301/2000\n",
            "3/3 [==============================] - 0s 63ms/step - loss: 0.6037 - accuracy: 0.8000 - val_loss: 0.5869 - val_accuracy: 0.7872\n",
            "Epoch 1302/2000\n",
            "3/3 [==============================] - 0s 68ms/step - loss: 0.5959 - accuracy: 0.8600 - val_loss: 0.5944 - val_accuracy: 0.8085\n",
            "Epoch 1303/2000\n",
            "3/3 [==============================] - 0s 69ms/step - loss: 0.5290 - accuracy: 0.8400 - val_loss: 0.6045 - val_accuracy: 0.8085\n",
            "Epoch 1304/2000\n",
            "3/3 [==============================] - 0s 70ms/step - loss: 0.5299 - accuracy: 0.8300 - val_loss: 0.6318 - val_accuracy: 0.7447\n",
            "Epoch 1305/2000\n",
            "3/3 [==============================] - 0s 60ms/step - loss: 0.4627 - accuracy: 0.8600 - val_loss: 0.6538 - val_accuracy: 0.7660\n",
            "Epoch 1306/2000\n",
            "3/3 [==============================] - 0s 38ms/step - loss: 0.5295 - accuracy: 0.7800 - val_loss: 0.6355 - val_accuracy: 0.7660\n",
            "Epoch 1307/2000\n",
            "3/3 [==============================] - 0s 52ms/step - loss: 0.5592 - accuracy: 0.8200 - val_loss: 0.6182 - val_accuracy: 0.7872\n",
            "Epoch 1308/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 0.5127 - accuracy: 0.8500 - val_loss: 0.5910 - val_accuracy: 0.7872\n",
            "Epoch 1309/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 0.5704 - accuracy: 0.8300 - val_loss: 0.5748 - val_accuracy: 0.7872\n",
            "Epoch 1310/2000\n",
            "3/3 [==============================] - 0s 34ms/step - loss: 0.5733 - accuracy: 0.8500 - val_loss: 0.5664 - val_accuracy: 0.8085\n",
            "Epoch 1311/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 0.5915 - accuracy: 0.7900 - val_loss: 0.5702 - val_accuracy: 0.7872\n",
            "Epoch 1312/2000\n",
            "3/3 [==============================] - 0s 51ms/step - loss: 0.5594 - accuracy: 0.8100 - val_loss: 0.5746 - val_accuracy: 0.8298\n",
            "Epoch 1313/2000\n",
            "3/3 [==============================] - 0s 46ms/step - loss: 0.5734 - accuracy: 0.8100 - val_loss: 0.5850 - val_accuracy: 0.8085\n",
            "Epoch 1314/2000\n",
            "3/3 [==============================] - 0s 34ms/step - loss: 0.6056 - accuracy: 0.7500 - val_loss: 0.5816 - val_accuracy: 0.8085\n",
            "Epoch 1315/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 0.4395 - accuracy: 0.8400 - val_loss: 0.5716 - val_accuracy: 0.8085\n",
            "Epoch 1316/2000\n",
            "3/3 [==============================] - 0s 44ms/step - loss: 0.6331 - accuracy: 0.8400 - val_loss: 0.5711 - val_accuracy: 0.8085\n",
            "Epoch 1317/2000\n",
            "3/3 [==============================] - 0s 55ms/step - loss: 0.5636 - accuracy: 0.8100 - val_loss: 0.6019 - val_accuracy: 0.7872\n",
            "Epoch 1318/2000\n",
            "3/3 [==============================] - 0s 53ms/step - loss: 0.3919 - accuracy: 0.9200 - val_loss: 0.7198 - val_accuracy: 0.7447\n",
            "Epoch 1319/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 0.6348 - accuracy: 0.8200 - val_loss: 0.7306 - val_accuracy: 0.7660\n",
            "Epoch 1320/2000\n",
            "3/3 [==============================] - 0s 53ms/step - loss: 0.6166 - accuracy: 0.8100 - val_loss: 0.6867 - val_accuracy: 0.8511\n",
            "Epoch 1321/2000\n",
            "3/3 [==============================] - 0s 34ms/step - loss: 0.4268 - accuracy: 0.8500 - val_loss: 0.6361 - val_accuracy: 0.8085\n",
            "Epoch 1322/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 0.5184 - accuracy: 0.8100 - val_loss: 0.5912 - val_accuracy: 0.8298\n",
            "Epoch 1323/2000\n",
            "3/3 [==============================] - 0s 47ms/step - loss: 0.5933 - accuracy: 0.7500 - val_loss: 0.5622 - val_accuracy: 0.8298\n",
            "Epoch 1324/2000\n",
            "3/3 [==============================] - 0s 45ms/step - loss: 0.6058 - accuracy: 0.8400 - val_loss: 0.5546 - val_accuracy: 0.8298\n",
            "Epoch 1325/2000\n",
            "3/3 [==============================] - 0s 51ms/step - loss: 0.6452 - accuracy: 0.8000 - val_loss: 0.5621 - val_accuracy: 0.8085\n",
            "Epoch 1326/2000\n",
            "3/3 [==============================] - 0s 39ms/step - loss: 0.6207 - accuracy: 0.8000 - val_loss: 0.5854 - val_accuracy: 0.8085\n",
            "Epoch 1327/2000\n",
            "3/3 [==============================] - 0s 34ms/step - loss: 0.6970 - accuracy: 0.7900 - val_loss: 0.5965 - val_accuracy: 0.7872\n",
            "Epoch 1328/2000\n",
            "3/3 [==============================] - 0s 46ms/step - loss: 0.5820 - accuracy: 0.8200 - val_loss: 0.6263 - val_accuracy: 0.7872\n",
            "Epoch 1329/2000\n",
            "3/3 [==============================] - 0s 33ms/step - loss: 0.5386 - accuracy: 0.8400 - val_loss: 0.6552 - val_accuracy: 0.7872\n",
            "Epoch 1330/2000\n",
            "3/3 [==============================] - 0s 55ms/step - loss: 0.4983 - accuracy: 0.8400 - val_loss: 0.6730 - val_accuracy: 0.7872\n",
            "Epoch 1331/2000\n",
            "3/3 [==============================] - 0s 34ms/step - loss: 0.7530 - accuracy: 0.7900 - val_loss: 0.6637 - val_accuracy: 0.7234\n",
            "Epoch 1332/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 0.5526 - accuracy: 0.8200 - val_loss: 0.6298 - val_accuracy: 0.7234\n",
            "Epoch 1333/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 0.6597 - accuracy: 0.7800 - val_loss: 0.5930 - val_accuracy: 0.7660\n",
            "Epoch 1334/2000\n",
            "3/3 [==============================] - 0s 34ms/step - loss: 0.4506 - accuracy: 0.8500 - val_loss: 0.5746 - val_accuracy: 0.7660\n",
            "Epoch 1335/2000\n",
            "3/3 [==============================] - 0s 34ms/step - loss: 0.7499 - accuracy: 0.7600 - val_loss: 0.5781 - val_accuracy: 0.7872\n",
            "Epoch 1336/2000\n",
            "3/3 [==============================] - 0s 40ms/step - loss: 0.6462 - accuracy: 0.7900 - val_loss: 0.5652 - val_accuracy: 0.7660\n",
            "Epoch 1337/2000\n",
            "3/3 [==============================] - 0s 46ms/step - loss: 0.3965 - accuracy: 0.8600 - val_loss: 0.5708 - val_accuracy: 0.7447\n",
            "Epoch 1338/2000\n",
            "3/3 [==============================] - 0s 47ms/step - loss: 0.4353 - accuracy: 0.8400 - val_loss: 0.5573 - val_accuracy: 0.7234\n",
            "Epoch 1339/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 0.6266 - accuracy: 0.7700 - val_loss: 0.5402 - val_accuracy: 0.7447\n",
            "Epoch 1340/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 0.6900 - accuracy: 0.7500 - val_loss: 0.5305 - val_accuracy: 0.7872\n",
            "Epoch 1341/2000\n",
            "3/3 [==============================] - 0s 57ms/step - loss: 0.5916 - accuracy: 0.8600 - val_loss: 0.5590 - val_accuracy: 0.7447\n",
            "Epoch 1342/2000\n",
            "3/3 [==============================] - 0s 50ms/step - loss: 0.5175 - accuracy: 0.7900 - val_loss: 0.5843 - val_accuracy: 0.7660\n",
            "Epoch 1343/2000\n",
            "3/3 [==============================] - 0s 47ms/step - loss: 0.4735 - accuracy: 0.8600 - val_loss: 0.5906 - val_accuracy: 0.7872\n",
            "Epoch 1344/2000\n",
            "3/3 [==============================] - 0s 58ms/step - loss: 0.6038 - accuracy: 0.8000 - val_loss: 0.5837 - val_accuracy: 0.8085\n",
            "Epoch 1345/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 0.5083 - accuracy: 0.8000 - val_loss: 0.5882 - val_accuracy: 0.7660\n",
            "Epoch 1346/2000\n",
            "3/3 [==============================] - 0s 44ms/step - loss: 0.5603 - accuracy: 0.8000 - val_loss: 0.5741 - val_accuracy: 0.7660\n",
            "Epoch 1347/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 0.5462 - accuracy: 0.8200 - val_loss: 0.5492 - val_accuracy: 0.8085\n",
            "Epoch 1348/2000\n",
            "3/3 [==============================] - 0s 51ms/step - loss: 0.5777 - accuracy: 0.8400 - val_loss: 0.5307 - val_accuracy: 0.8085\n",
            "Epoch 1349/2000\n",
            "3/3 [==============================] - 0s 40ms/step - loss: 0.5021 - accuracy: 0.8700 - val_loss: 0.5197 - val_accuracy: 0.8085\n",
            "Epoch 1350/2000\n",
            "3/3 [==============================] - 0s 53ms/step - loss: 0.5700 - accuracy: 0.7700 - val_loss: 0.5425 - val_accuracy: 0.8085\n",
            "Epoch 1351/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 0.6878 - accuracy: 0.8200 - val_loss: 0.5537 - val_accuracy: 0.8298\n",
            "Epoch 1352/2000\n",
            "3/3 [==============================] - 0s 46ms/step - loss: 0.6878 - accuracy: 0.7700 - val_loss: 0.5481 - val_accuracy: 0.8298\n",
            "Epoch 1353/2000\n",
            "3/3 [==============================] - 0s 38ms/step - loss: 0.5538 - accuracy: 0.8400 - val_loss: 0.5430 - val_accuracy: 0.8511\n",
            "Epoch 1354/2000\n",
            "3/3 [==============================] - 0s 44ms/step - loss: 0.5629 - accuracy: 0.8500 - val_loss: 0.5516 - val_accuracy: 0.8298\n",
            "Epoch 1355/2000\n",
            "3/3 [==============================] - 0s 45ms/step - loss: 0.5600 - accuracy: 0.8500 - val_loss: 0.5774 - val_accuracy: 0.8085\n",
            "Epoch 1356/2000\n",
            "3/3 [==============================] - 0s 52ms/step - loss: 0.6604 - accuracy: 0.7900 - val_loss: 0.6469 - val_accuracy: 0.7872\n",
            "Epoch 1357/2000\n",
            "3/3 [==============================] - 0s 45ms/step - loss: 0.7007 - accuracy: 0.7700 - val_loss: 0.6777 - val_accuracy: 0.7660\n",
            "Epoch 1358/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 0.5682 - accuracy: 0.8200 - val_loss: 0.6357 - val_accuracy: 0.7660\n",
            "Epoch 1359/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 0.5938 - accuracy: 0.8400 - val_loss: 0.5873 - val_accuracy: 0.7447\n",
            "Epoch 1360/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 0.4765 - accuracy: 0.8500 - val_loss: 0.5814 - val_accuracy: 0.7660\n",
            "Epoch 1361/2000\n",
            "3/3 [==============================] - 0s 48ms/step - loss: 0.5067 - accuracy: 0.8600 - val_loss: 0.5669 - val_accuracy: 0.7872\n",
            "Epoch 1362/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 0.5095 - accuracy: 0.8900 - val_loss: 0.5474 - val_accuracy: 0.7872\n",
            "Epoch 1363/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 0.4997 - accuracy: 0.8500 - val_loss: 0.5334 - val_accuracy: 0.7660\n",
            "Epoch 1364/2000\n",
            "3/3 [==============================] - 0s 52ms/step - loss: 0.5137 - accuracy: 0.8500 - val_loss: 0.5490 - val_accuracy: 0.7660\n",
            "Epoch 1365/2000\n",
            "3/3 [==============================] - 0s 45ms/step - loss: 0.6458 - accuracy: 0.8100 - val_loss: 0.5544 - val_accuracy: 0.7872\n",
            "Epoch 1366/2000\n",
            "3/3 [==============================] - 0s 61ms/step - loss: 0.6254 - accuracy: 0.8100 - val_loss: 0.5831 - val_accuracy: 0.7447\n",
            "Epoch 1367/2000\n",
            "3/3 [==============================] - 0s 56ms/step - loss: 0.6390 - accuracy: 0.8400 - val_loss: 0.6188 - val_accuracy: 0.7660\n",
            "Epoch 1368/2000\n",
            "3/3 [==============================] - 0s 73ms/step - loss: 0.6223 - accuracy: 0.8100 - val_loss: 0.6339 - val_accuracy: 0.7447\n",
            "Epoch 1369/2000\n",
            "3/3 [==============================] - 0s 65ms/step - loss: 0.4739 - accuracy: 0.8500 - val_loss: 0.6352 - val_accuracy: 0.7660\n",
            "Epoch 1370/2000\n",
            "3/3 [==============================] - 0s 60ms/step - loss: 0.5865 - accuracy: 0.8200 - val_loss: 0.6343 - val_accuracy: 0.7660\n",
            "Epoch 1371/2000\n",
            "3/3 [==============================] - 0s 58ms/step - loss: 0.6192 - accuracy: 0.8500 - val_loss: 0.6039 - val_accuracy: 0.7447\n",
            "Epoch 1372/2000\n",
            "3/3 [==============================] - 0s 64ms/step - loss: 0.6337 - accuracy: 0.8400 - val_loss: 0.5866 - val_accuracy: 0.7660\n",
            "Epoch 1373/2000\n",
            "3/3 [==============================] - 0s 61ms/step - loss: 0.5093 - accuracy: 0.8000 - val_loss: 0.5848 - val_accuracy: 0.7447\n",
            "Epoch 1374/2000\n",
            "3/3 [==============================] - 0s 74ms/step - loss: 0.5815 - accuracy: 0.8000 - val_loss: 0.5783 - val_accuracy: 0.7660\n",
            "Epoch 1375/2000\n",
            "3/3 [==============================] - 0s 61ms/step - loss: 0.6300 - accuracy: 0.8200 - val_loss: 0.5799 - val_accuracy: 0.8298\n",
            "Epoch 1376/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 0.5397 - accuracy: 0.8400 - val_loss: 0.5803 - val_accuracy: 0.7872\n",
            "Epoch 1377/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 0.4794 - accuracy: 0.8600 - val_loss: 0.5739 - val_accuracy: 0.7660\n",
            "Epoch 1378/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 0.5998 - accuracy: 0.8000 - val_loss: 0.5668 - val_accuracy: 0.7660\n",
            "Epoch 1379/2000\n",
            "3/3 [==============================] - 0s 48ms/step - loss: 0.7631 - accuracy: 0.8200 - val_loss: 0.5680 - val_accuracy: 0.7872\n",
            "Epoch 1380/2000\n",
            "3/3 [==============================] - 0s 47ms/step - loss: 0.6798 - accuracy: 0.7300 - val_loss: 0.5735 - val_accuracy: 0.7872\n",
            "Epoch 1381/2000\n",
            "3/3 [==============================] - 0s 34ms/step - loss: 0.5455 - accuracy: 0.8400 - val_loss: 0.5780 - val_accuracy: 0.7660\n",
            "Epoch 1382/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 0.8451 - accuracy: 0.7800 - val_loss: 0.6286 - val_accuracy: 0.7021\n",
            "Epoch 1383/2000\n",
            "3/3 [==============================] - 0s 52ms/step - loss: 0.6592 - accuracy: 0.7700 - val_loss: 0.7042 - val_accuracy: 0.6809\n",
            "Epoch 1384/2000\n",
            "3/3 [==============================] - 0s 45ms/step - loss: 0.7222 - accuracy: 0.7600 - val_loss: 0.6490 - val_accuracy: 0.7021\n",
            "Epoch 1385/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 0.5321 - accuracy: 0.8100 - val_loss: 0.6007 - val_accuracy: 0.8085\n",
            "Epoch 1386/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 0.6270 - accuracy: 0.7600 - val_loss: 0.5742 - val_accuracy: 0.8298\n",
            "Epoch 1387/2000\n",
            "3/3 [==============================] - 0s 57ms/step - loss: 0.5520 - accuracy: 0.8300 - val_loss: 0.5627 - val_accuracy: 0.8085\n",
            "Epoch 1388/2000\n",
            "3/3 [==============================] - 0s 54ms/step - loss: 0.5224 - accuracy: 0.8300 - val_loss: 0.5682 - val_accuracy: 0.8298\n",
            "Epoch 1389/2000\n",
            "3/3 [==============================] - 0s 51ms/step - loss: 0.4878 - accuracy: 0.8300 - val_loss: 0.5951 - val_accuracy: 0.7872\n",
            "Epoch 1390/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 0.5370 - accuracy: 0.8000 - val_loss: 0.6041 - val_accuracy: 0.7872\n",
            "Epoch 1391/2000\n",
            "3/3 [==============================] - 0s 54ms/step - loss: 0.7131 - accuracy: 0.8200 - val_loss: 0.5886 - val_accuracy: 0.7660\n",
            "Epoch 1392/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 0.7100 - accuracy: 0.8000 - val_loss: 0.5668 - val_accuracy: 0.8298\n",
            "Epoch 1393/2000\n",
            "3/3 [==============================] - 0s 40ms/step - loss: 0.4038 - accuracy: 0.8600 - val_loss: 0.5475 - val_accuracy: 0.8298\n",
            "Epoch 1394/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 0.5989 - accuracy: 0.8100 - val_loss: 0.5630 - val_accuracy: 0.8085\n",
            "Epoch 1395/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 0.4991 - accuracy: 0.8100 - val_loss: 0.5870 - val_accuracy: 0.7872\n",
            "Epoch 1396/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 0.5538 - accuracy: 0.8100 - val_loss: 0.6038 - val_accuracy: 0.7872\n",
            "Epoch 1397/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 0.6360 - accuracy: 0.8100 - val_loss: 0.5964 - val_accuracy: 0.8085\n",
            "Epoch 1398/2000\n",
            "3/3 [==============================] - 0s 56ms/step - loss: 0.5792 - accuracy: 0.8200 - val_loss: 0.5788 - val_accuracy: 0.8298\n",
            "Epoch 1399/2000\n",
            "3/3 [==============================] - 0s 51ms/step - loss: 0.6903 - accuracy: 0.7800 - val_loss: 0.5637 - val_accuracy: 0.8085\n",
            "Epoch 1400/2000\n",
            "3/3 [==============================] - 0s 46ms/step - loss: 0.4179 - accuracy: 0.8900 - val_loss: 0.5731 - val_accuracy: 0.7660\n",
            "Epoch 1401/2000\n",
            "3/3 [==============================] - 0s 34ms/step - loss: 0.4647 - accuracy: 0.8600 - val_loss: 0.5787 - val_accuracy: 0.7660\n",
            "Epoch 1402/2000\n",
            "3/3 [==============================] - 0s 40ms/step - loss: 0.5950 - accuracy: 0.8400 - val_loss: 0.5922 - val_accuracy: 0.7872\n",
            "Epoch 1403/2000\n",
            "3/3 [==============================] - 0s 48ms/step - loss: 0.4783 - accuracy: 0.8400 - val_loss: 0.6345 - val_accuracy: 0.8085\n",
            "Epoch 1404/2000\n",
            "3/3 [==============================] - 0s 52ms/step - loss: 0.7224 - accuracy: 0.8200 - val_loss: 0.5947 - val_accuracy: 0.7872\n",
            "Epoch 1405/2000\n",
            "3/3 [==============================] - 0s 44ms/step - loss: 0.5542 - accuracy: 0.7700 - val_loss: 0.5449 - val_accuracy: 0.8085\n",
            "Epoch 1406/2000\n",
            "3/3 [==============================] - 0s 40ms/step - loss: 0.5298 - accuracy: 0.8100 - val_loss: 0.4979 - val_accuracy: 0.8298\n",
            "Epoch 1407/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 0.7712 - accuracy: 0.8000 - val_loss: 0.4916 - val_accuracy: 0.8298\n",
            "Epoch 1408/2000\n",
            "3/3 [==============================] - 0s 47ms/step - loss: 0.5947 - accuracy: 0.8300 - val_loss: 0.5016 - val_accuracy: 0.8511\n",
            "Epoch 1409/2000\n",
            "3/3 [==============================] - 0s 34ms/step - loss: 0.5383 - accuracy: 0.8100 - val_loss: 0.5150 - val_accuracy: 0.8511\n",
            "Epoch 1410/2000\n",
            "3/3 [==============================] - 0s 44ms/step - loss: 0.5639 - accuracy: 0.8100 - val_loss: 0.5396 - val_accuracy: 0.8298\n",
            "Epoch 1411/2000\n",
            "3/3 [==============================] - 0s 61ms/step - loss: 0.5919 - accuracy: 0.8300 - val_loss: 0.5769 - val_accuracy: 0.8298\n",
            "Epoch 1412/2000\n",
            "3/3 [==============================] - 0s 44ms/step - loss: 0.4742 - accuracy: 0.8100 - val_loss: 0.5817 - val_accuracy: 0.8085\n",
            "Epoch 1413/2000\n",
            "3/3 [==============================] - 0s 47ms/step - loss: 0.5150 - accuracy: 0.7900 - val_loss: 0.5782 - val_accuracy: 0.8085\n",
            "Epoch 1414/2000\n",
            "3/3 [==============================] - 0s 45ms/step - loss: 0.6622 - accuracy: 0.8500 - val_loss: 0.5658 - val_accuracy: 0.8085\n",
            "Epoch 1415/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 0.6022 - accuracy: 0.8100 - val_loss: 0.5385 - val_accuracy: 0.8085\n",
            "Epoch 1416/2000\n",
            "3/3 [==============================] - 0s 34ms/step - loss: 0.4994 - accuracy: 0.8500 - val_loss: 0.5170 - val_accuracy: 0.8511\n",
            "Epoch 1417/2000\n",
            "3/3 [==============================] - 0s 66ms/step - loss: 0.4065 - accuracy: 0.8300 - val_loss: 0.5113 - val_accuracy: 0.8298\n",
            "Epoch 1418/2000\n",
            "3/3 [==============================] - 0s 34ms/step - loss: 0.5847 - accuracy: 0.7900 - val_loss: 0.5407 - val_accuracy: 0.7872\n",
            "Epoch 1419/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 0.6863 - accuracy: 0.7300 - val_loss: 0.5613 - val_accuracy: 0.7872\n",
            "Epoch 1420/2000\n",
            "3/3 [==============================] - 0s 46ms/step - loss: 0.5274 - accuracy: 0.8200 - val_loss: 0.5116 - val_accuracy: 0.8085\n",
            "Epoch 1421/2000\n",
            "3/3 [==============================] - 0s 53ms/step - loss: 0.4985 - accuracy: 0.8300 - val_loss: 0.4695 - val_accuracy: 0.8085\n",
            "Epoch 1422/2000\n",
            "3/3 [==============================] - 0s 48ms/step - loss: 0.5908 - accuracy: 0.7900 - val_loss: 0.4734 - val_accuracy: 0.8085\n",
            "Epoch 1423/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 0.4530 - accuracy: 0.8300 - val_loss: 0.4933 - val_accuracy: 0.8085\n",
            "Epoch 1424/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 0.4378 - accuracy: 0.9000 - val_loss: 0.4681 - val_accuracy: 0.8298\n",
            "Epoch 1425/2000\n",
            "3/3 [==============================] - 0s 45ms/step - loss: 0.3659 - accuracy: 0.8600 - val_loss: 0.4401 - val_accuracy: 0.8085\n",
            "Epoch 1426/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 0.5294 - accuracy: 0.8600 - val_loss: 0.4277 - val_accuracy: 0.8298\n",
            "Epoch 1427/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 0.4440 - accuracy: 0.8600 - val_loss: 0.4246 - val_accuracy: 0.8298\n",
            "Epoch 1428/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 0.6945 - accuracy: 0.7900 - val_loss: 0.4260 - val_accuracy: 0.8511\n",
            "Epoch 1429/2000\n",
            "3/3 [==============================] - 0s 45ms/step - loss: 0.6024 - accuracy: 0.8400 - val_loss: 0.4467 - val_accuracy: 0.8298\n",
            "Epoch 1430/2000\n",
            "3/3 [==============================] - 0s 45ms/step - loss: 0.5857 - accuracy: 0.8200 - val_loss: 0.4667 - val_accuracy: 0.7872\n",
            "Epoch 1431/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 0.5784 - accuracy: 0.8800 - val_loss: 0.4924 - val_accuracy: 0.7872\n",
            "Epoch 1432/2000\n",
            "3/3 [==============================] - 0s 44ms/step - loss: 0.6720 - accuracy: 0.8000 - val_loss: 0.5140 - val_accuracy: 0.8085\n",
            "Epoch 1433/2000\n",
            "3/3 [==============================] - 0s 34ms/step - loss: 0.7110 - accuracy: 0.8300 - val_loss: 0.4974 - val_accuracy: 0.8511\n",
            "Epoch 1434/2000\n",
            "3/3 [==============================] - 0s 53ms/step - loss: 0.4969 - accuracy: 0.8400 - val_loss: 0.4886 - val_accuracy: 0.8298\n",
            "Epoch 1435/2000\n",
            "3/3 [==============================] - 0s 46ms/step - loss: 0.4921 - accuracy: 0.8500 - val_loss: 0.4810 - val_accuracy: 0.8511\n",
            "Epoch 1436/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 0.4501 - accuracy: 0.9000 - val_loss: 0.4682 - val_accuracy: 0.8085\n",
            "Epoch 1437/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 0.6234 - accuracy: 0.7800 - val_loss: 0.4639 - val_accuracy: 0.8723\n",
            "Epoch 1438/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 0.4656 - accuracy: 0.8300 - val_loss: 0.4856 - val_accuracy: 0.7872\n",
            "Epoch 1439/2000\n",
            "3/3 [==============================] - 0s 46ms/step - loss: 0.6734 - accuracy: 0.7900 - val_loss: 0.5237 - val_accuracy: 0.7660\n",
            "Epoch 1440/2000\n",
            "3/3 [==============================] - 0s 55ms/step - loss: 0.5203 - accuracy: 0.8300 - val_loss: 0.5451 - val_accuracy: 0.7872\n",
            "Epoch 1441/2000\n",
            "3/3 [==============================] - 0s 76ms/step - loss: 0.5384 - accuracy: 0.8200 - val_loss: 0.5418 - val_accuracy: 0.7872\n",
            "Epoch 1442/2000\n",
            "3/3 [==============================] - 0s 72ms/step - loss: 0.4318 - accuracy: 0.8900 - val_loss: 0.5327 - val_accuracy: 0.8298\n",
            "Epoch 1443/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 0.5985 - accuracy: 0.7900 - val_loss: 0.5234 - val_accuracy: 0.8085\n",
            "Epoch 1444/2000\n",
            "3/3 [==============================] - 0s 70ms/step - loss: 0.6314 - accuracy: 0.8000 - val_loss: 0.5143 - val_accuracy: 0.8085\n",
            "Epoch 1445/2000\n",
            "3/3 [==============================] - 0s 45ms/step - loss: 0.6026 - accuracy: 0.7500 - val_loss: 0.5009 - val_accuracy: 0.8511\n",
            "Epoch 1446/2000\n",
            "3/3 [==============================] - 0s 67ms/step - loss: 0.5687 - accuracy: 0.8700 - val_loss: 0.5132 - val_accuracy: 0.8511\n",
            "Epoch 1447/2000\n",
            "3/3 [==============================] - 0s 64ms/step - loss: 0.5276 - accuracy: 0.8700 - val_loss: 0.5219 - val_accuracy: 0.8298\n",
            "Epoch 1448/2000\n",
            "3/3 [==============================] - 0s 69ms/step - loss: 0.4777 - accuracy: 0.8200 - val_loss: 0.5324 - val_accuracy: 0.7872\n",
            "Epoch 1449/2000\n",
            "3/3 [==============================] - 0s 62ms/step - loss: 0.5822 - accuracy: 0.8100 - val_loss: 0.5498 - val_accuracy: 0.7234\n",
            "Epoch 1450/2000\n",
            "3/3 [==============================] - 0s 48ms/step - loss: 0.4786 - accuracy: 0.8100 - val_loss: 0.5464 - val_accuracy: 0.7234\n",
            "Epoch 1451/2000\n",
            "3/3 [==============================] - 0s 62ms/step - loss: 0.6321 - accuracy: 0.7900 - val_loss: 0.5475 - val_accuracy: 0.7021\n",
            "Epoch 1452/2000\n",
            "3/3 [==============================] - 0s 57ms/step - loss: 0.4038 - accuracy: 0.8200 - val_loss: 0.5592 - val_accuracy: 0.7447\n",
            "Epoch 1453/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 0.5071 - accuracy: 0.8100 - val_loss: 0.5513 - val_accuracy: 0.7447\n",
            "Epoch 1454/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 0.5928 - accuracy: 0.7500 - val_loss: 0.5382 - val_accuracy: 0.7872\n",
            "Epoch 1455/2000\n",
            "3/3 [==============================] - 0s 55ms/step - loss: 0.5047 - accuracy: 0.8200 - val_loss: 0.5583 - val_accuracy: 0.7660\n",
            "Epoch 1456/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 0.4884 - accuracy: 0.8400 - val_loss: 0.6114 - val_accuracy: 0.7234\n",
            "Epoch 1457/2000\n",
            "3/3 [==============================] - 0s 44ms/step - loss: 0.4577 - accuracy: 0.8500 - val_loss: 0.6516 - val_accuracy: 0.7872\n",
            "Epoch 1458/2000\n",
            "3/3 [==============================] - 0s 48ms/step - loss: 0.5775 - accuracy: 0.7400 - val_loss: 0.6591 - val_accuracy: 0.7872\n",
            "Epoch 1459/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 0.6569 - accuracy: 0.7900 - val_loss: 0.6353 - val_accuracy: 0.7234\n",
            "Epoch 1460/2000\n",
            "3/3 [==============================] - 0s 47ms/step - loss: 0.6949 - accuracy: 0.8100 - val_loss: 0.5957 - val_accuracy: 0.7660\n",
            "Epoch 1461/2000\n",
            "3/3 [==============================] - 0s 57ms/step - loss: 0.5955 - accuracy: 0.8500 - val_loss: 0.5551 - val_accuracy: 0.8085\n",
            "Epoch 1462/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 0.6121 - accuracy: 0.8000 - val_loss: 0.5191 - val_accuracy: 0.8085\n",
            "Epoch 1463/2000\n",
            "3/3 [==============================] - 0s 59ms/step - loss: 0.5076 - accuracy: 0.8500 - val_loss: 0.5155 - val_accuracy: 0.7872\n",
            "Epoch 1464/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 0.5815 - accuracy: 0.7500 - val_loss: 0.5388 - val_accuracy: 0.7660\n",
            "Epoch 1465/2000\n",
            "3/3 [==============================] - 0s 59ms/step - loss: 0.6675 - accuracy: 0.7700 - val_loss: 0.5486 - val_accuracy: 0.8085\n",
            "Epoch 1466/2000\n",
            "3/3 [==============================] - 0s 55ms/step - loss: 0.4878 - accuracy: 0.8600 - val_loss: 0.5663 - val_accuracy: 0.7660\n",
            "Epoch 1467/2000\n",
            "3/3 [==============================] - 0s 67ms/step - loss: 0.5988 - accuracy: 0.8100 - val_loss: 0.5779 - val_accuracy: 0.7660\n",
            "Epoch 1468/2000\n",
            "3/3 [==============================] - 0s 48ms/step - loss: 0.7209 - accuracy: 0.8100 - val_loss: 0.5814 - val_accuracy: 0.7660\n",
            "Epoch 1469/2000\n",
            "3/3 [==============================] - 0s 38ms/step - loss: 0.6706 - accuracy: 0.8200 - val_loss: 0.5765 - val_accuracy: 0.7660\n",
            "Epoch 1470/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 0.4908 - accuracy: 0.8100 - val_loss: 0.5884 - val_accuracy: 0.7660\n",
            "Epoch 1471/2000\n",
            "3/3 [==============================] - 0s 55ms/step - loss: 0.6672 - accuracy: 0.8100 - val_loss: 0.5865 - val_accuracy: 0.7447\n",
            "Epoch 1472/2000\n",
            "3/3 [==============================] - 0s 47ms/step - loss: 0.6505 - accuracy: 0.7900 - val_loss: 0.5943 - val_accuracy: 0.7660\n",
            "Epoch 1473/2000\n",
            "3/3 [==============================] - 0s 46ms/step - loss: 0.5418 - accuracy: 0.7800 - val_loss: 0.5833 - val_accuracy: 0.7660\n",
            "Epoch 1474/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 0.5400 - accuracy: 0.8100 - val_loss: 0.5834 - val_accuracy: 0.7660\n",
            "Epoch 1475/2000\n",
            "3/3 [==============================] - 0s 50ms/step - loss: 0.4643 - accuracy: 0.8300 - val_loss: 0.6258 - val_accuracy: 0.7660\n",
            "Epoch 1476/2000\n",
            "3/3 [==============================] - 0s 52ms/step - loss: 0.5212 - accuracy: 0.8400 - val_loss: 0.6543 - val_accuracy: 0.7660\n",
            "Epoch 1477/2000\n",
            "3/3 [==============================] - 0s 49ms/step - loss: 0.5006 - accuracy: 0.8400 - val_loss: 0.6571 - val_accuracy: 0.7447\n",
            "Epoch 1478/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 0.6213 - accuracy: 0.8000 - val_loss: 0.6550 - val_accuracy: 0.7660\n",
            "Epoch 1479/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 0.5117 - accuracy: 0.8700 - val_loss: 0.6294 - val_accuracy: 0.7872\n",
            "Epoch 1480/2000\n",
            "3/3 [==============================] - 0s 55ms/step - loss: 0.4560 - accuracy: 0.8500 - val_loss: 0.6176 - val_accuracy: 0.7872\n",
            "Epoch 1481/2000\n",
            "3/3 [==============================] - 0s 39ms/step - loss: 0.5434 - accuracy: 0.8000 - val_loss: 0.6025 - val_accuracy: 0.7872\n",
            "Epoch 1482/2000\n",
            "3/3 [==============================] - 0s 44ms/step - loss: 0.6844 - accuracy: 0.8000 - val_loss: 0.6032 - val_accuracy: 0.7660\n",
            "Epoch 1483/2000\n",
            "3/3 [==============================] - 0s 50ms/step - loss: 0.4665 - accuracy: 0.8300 - val_loss: 0.6061 - val_accuracy: 0.7872\n",
            "Epoch 1484/2000\n",
            "3/3 [==============================] - 0s 55ms/step - loss: 0.7384 - accuracy: 0.8200 - val_loss: 0.5790 - val_accuracy: 0.7872\n",
            "Epoch 1485/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 0.4851 - accuracy: 0.8500 - val_loss: 0.5728 - val_accuracy: 0.7660\n",
            "Epoch 1486/2000\n",
            "3/3 [==============================] - 0s 34ms/step - loss: 0.6529 - accuracy: 0.8200 - val_loss: 0.5742 - val_accuracy: 0.7660\n",
            "Epoch 1487/2000\n",
            "3/3 [==============================] - 0s 59ms/step - loss: 0.5916 - accuracy: 0.7800 - val_loss: 0.5629 - val_accuracy: 0.7872\n",
            "Epoch 1488/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 0.6317 - accuracy: 0.8100 - val_loss: 0.5539 - val_accuracy: 0.8085\n",
            "Epoch 1489/2000\n",
            "3/3 [==============================] - 0s 53ms/step - loss: 0.4388 - accuracy: 0.8800 - val_loss: 0.5466 - val_accuracy: 0.7872\n",
            "Epoch 1490/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 0.4474 - accuracy: 0.8700 - val_loss: 0.5494 - val_accuracy: 0.8085\n",
            "Epoch 1491/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 0.5402 - accuracy: 0.8000 - val_loss: 0.5400 - val_accuracy: 0.8085\n",
            "Epoch 1492/2000\n",
            "3/3 [==============================] - 0s 58ms/step - loss: 0.4383 - accuracy: 0.8300 - val_loss: 0.5292 - val_accuracy: 0.8085\n",
            "Epoch 1493/2000\n",
            "3/3 [==============================] - 0s 34ms/step - loss: 0.4742 - accuracy: 0.8400 - val_loss: 0.5342 - val_accuracy: 0.8085\n",
            "Epoch 1494/2000\n",
            "3/3 [==============================] - 0s 45ms/step - loss: 0.3902 - accuracy: 0.9000 - val_loss: 0.5441 - val_accuracy: 0.7872\n",
            "Epoch 1495/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 0.5663 - accuracy: 0.7600 - val_loss: 0.5403 - val_accuracy: 0.8085\n",
            "Epoch 1496/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 0.5382 - accuracy: 0.7800 - val_loss: 0.5480 - val_accuracy: 0.8085\n",
            "Epoch 1497/2000\n",
            "3/3 [==============================] - 0s 45ms/step - loss: 0.5555 - accuracy: 0.8000 - val_loss: 0.5296 - val_accuracy: 0.8298\n",
            "Epoch 1498/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 0.5132 - accuracy: 0.8400 - val_loss: 0.5196 - val_accuracy: 0.8298\n",
            "Epoch 1499/2000\n",
            "3/3 [==============================] - 0s 49ms/step - loss: 0.6024 - accuracy: 0.7800 - val_loss: 0.5128 - val_accuracy: 0.8298\n",
            "Epoch 1500/2000\n",
            "3/3 [==============================] - 0s 54ms/step - loss: 0.5505 - accuracy: 0.8000 - val_loss: 0.5101 - val_accuracy: 0.8298\n",
            "Epoch 1501/2000\n",
            "3/3 [==============================] - 0s 69ms/step - loss: 0.5257 - accuracy: 0.8200 - val_loss: 0.5023 - val_accuracy: 0.8298\n",
            "Epoch 1502/2000\n",
            "3/3 [==============================] - 0s 88ms/step - loss: 0.7607 - accuracy: 0.7900 - val_loss: 0.5097 - val_accuracy: 0.8298\n",
            "Epoch 1503/2000\n",
            "3/3 [==============================] - 0s 53ms/step - loss: 0.4558 - accuracy: 0.9200 - val_loss: 0.5311 - val_accuracy: 0.8298\n",
            "Epoch 1504/2000\n",
            "3/3 [==============================] - 0s 61ms/step - loss: 0.5496 - accuracy: 0.7900 - val_loss: 0.5396 - val_accuracy: 0.8511\n",
            "Epoch 1505/2000\n",
            "3/3 [==============================] - 0s 70ms/step - loss: 0.7183 - accuracy: 0.7800 - val_loss: 0.5384 - val_accuracy: 0.8298\n",
            "Epoch 1506/2000\n",
            "3/3 [==============================] - 0s 58ms/step - loss: 0.5367 - accuracy: 0.7900 - val_loss: 0.5349 - val_accuracy: 0.8085\n",
            "Epoch 1507/2000\n",
            "3/3 [==============================] - 0s 71ms/step - loss: 0.4918 - accuracy: 0.8700 - val_loss: 0.5059 - val_accuracy: 0.7872\n",
            "Epoch 1508/2000\n",
            "3/3 [==============================] - 0s 66ms/step - loss: 0.4331 - accuracy: 0.8700 - val_loss: 0.4865 - val_accuracy: 0.7660\n",
            "Epoch 1509/2000\n",
            "3/3 [==============================] - 0s 66ms/step - loss: 0.4496 - accuracy: 0.8600 - val_loss: 0.4838 - val_accuracy: 0.7872\n",
            "Epoch 1510/2000\n",
            "3/3 [==============================] - 0s 68ms/step - loss: 0.5006 - accuracy: 0.8200 - val_loss: 0.5164 - val_accuracy: 0.8085\n",
            "Epoch 1511/2000\n",
            "3/3 [==============================] - 0s 62ms/step - loss: 0.5438 - accuracy: 0.8300 - val_loss: 0.5669 - val_accuracy: 0.7021\n",
            "Epoch 1512/2000\n",
            "3/3 [==============================] - 0s 57ms/step - loss: 0.7019 - accuracy: 0.8200 - val_loss: 0.5581 - val_accuracy: 0.7021\n",
            "Epoch 1513/2000\n",
            "3/3 [==============================] - 0s 38ms/step - loss: 0.5661 - accuracy: 0.8200 - val_loss: 0.5004 - val_accuracy: 0.7660\n",
            "Epoch 1514/2000\n",
            "3/3 [==============================] - 0s 53ms/step - loss: 0.5101 - accuracy: 0.8400 - val_loss: 0.4544 - val_accuracy: 0.7660\n",
            "Epoch 1515/2000\n",
            "3/3 [==============================] - 0s 44ms/step - loss: 0.5569 - accuracy: 0.8200 - val_loss: 0.4476 - val_accuracy: 0.8085\n",
            "Epoch 1516/2000\n",
            "3/3 [==============================] - 0s 45ms/step - loss: 0.5458 - accuracy: 0.8500 - val_loss: 0.4467 - val_accuracy: 0.8298\n",
            "Epoch 1517/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 0.6034 - accuracy: 0.8100 - val_loss: 0.4590 - val_accuracy: 0.8298\n",
            "Epoch 1518/2000\n",
            "3/3 [==============================] - 0s 48ms/step - loss: 0.5690 - accuracy: 0.8100 - val_loss: 0.4902 - val_accuracy: 0.8298\n",
            "Epoch 1519/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 0.5398 - accuracy: 0.8100 - val_loss: 0.5282 - val_accuracy: 0.8298\n",
            "Epoch 1520/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 0.4623 - accuracy: 0.8500 - val_loss: 0.5417 - val_accuracy: 0.8085\n",
            "Epoch 1521/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 0.4710 - accuracy: 0.8200 - val_loss: 0.5368 - val_accuracy: 0.8085\n",
            "Epoch 1522/2000\n",
            "3/3 [==============================] - 0s 46ms/step - loss: 0.6422 - accuracy: 0.8100 - val_loss: 0.5477 - val_accuracy: 0.8298\n",
            "Epoch 1523/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 0.4712 - accuracy: 0.8700 - val_loss: 0.5572 - val_accuracy: 0.8298\n",
            "Epoch 1524/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 0.5272 - accuracy: 0.8200 - val_loss: 0.5388 - val_accuracy: 0.8085\n",
            "Epoch 1525/2000\n",
            "3/3 [==============================] - 0s 44ms/step - loss: 0.4375 - accuracy: 0.8500 - val_loss: 0.5079 - val_accuracy: 0.8085\n",
            "Epoch 1526/2000\n",
            "3/3 [==============================] - 0s 34ms/step - loss: 0.4338 - accuracy: 0.8400 - val_loss: 0.5122 - val_accuracy: 0.7872\n",
            "Epoch 1527/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 0.4585 - accuracy: 0.8500 - val_loss: 0.5327 - val_accuracy: 0.7872\n",
            "Epoch 1528/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 0.5459 - accuracy: 0.8400 - val_loss: 0.5785 - val_accuracy: 0.7447\n",
            "Epoch 1529/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 0.5746 - accuracy: 0.8400 - val_loss: 0.5764 - val_accuracy: 0.7447\n",
            "Epoch 1530/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 0.5317 - accuracy: 0.7800 - val_loss: 0.5686 - val_accuracy: 0.7660\n",
            "Epoch 1531/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 0.5376 - accuracy: 0.8200 - val_loss: 0.5540 - val_accuracy: 0.7872\n",
            "Epoch 1532/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 0.6152 - accuracy: 0.8200 - val_loss: 0.5504 - val_accuracy: 0.7660\n",
            "Epoch 1533/2000\n",
            "3/3 [==============================] - 0s 46ms/step - loss: 0.5348 - accuracy: 0.8500 - val_loss: 0.5405 - val_accuracy: 0.7660\n",
            "Epoch 1534/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 0.5556 - accuracy: 0.8000 - val_loss: 0.5514 - val_accuracy: 0.7660\n",
            "Epoch 1535/2000\n",
            "3/3 [==============================] - 0s 34ms/step - loss: 0.4623 - accuracy: 0.8500 - val_loss: 0.5782 - val_accuracy: 0.7660\n",
            "Epoch 1536/2000\n",
            "3/3 [==============================] - 0s 57ms/step - loss: 0.4761 - accuracy: 0.8500 - val_loss: 0.6111 - val_accuracy: 0.7447\n",
            "Epoch 1537/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 0.5310 - accuracy: 0.8000 - val_loss: 0.6258 - val_accuracy: 0.7660\n",
            "Epoch 1538/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 0.5448 - accuracy: 0.8300 - val_loss: 0.5779 - val_accuracy: 0.7660\n",
            "Epoch 1539/2000\n",
            "3/3 [==============================] - 0s 47ms/step - loss: 0.4090 - accuracy: 0.8400 - val_loss: 0.5289 - val_accuracy: 0.8085\n",
            "Epoch 1540/2000\n",
            "3/3 [==============================] - 0s 39ms/step - loss: 0.5269 - accuracy: 0.9000 - val_loss: 0.4991 - val_accuracy: 0.8085\n",
            "Epoch 1541/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 0.6088 - accuracy: 0.8100 - val_loss: 0.4810 - val_accuracy: 0.7872\n",
            "Epoch 1542/2000\n",
            "3/3 [==============================] - 0s 48ms/step - loss: 0.5073 - accuracy: 0.8600 - val_loss: 0.4801 - val_accuracy: 0.7872\n",
            "Epoch 1543/2000\n",
            "3/3 [==============================] - 0s 44ms/step - loss: 0.4820 - accuracy: 0.8100 - val_loss: 0.4829 - val_accuracy: 0.8085\n",
            "Epoch 1544/2000\n",
            "3/3 [==============================] - 0s 46ms/step - loss: 0.5923 - accuracy: 0.7800 - val_loss: 0.4970 - val_accuracy: 0.8085\n",
            "Epoch 1545/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 0.5748 - accuracy: 0.8300 - val_loss: 0.5126 - val_accuracy: 0.7872\n",
            "Epoch 1546/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 0.4903 - accuracy: 0.8400 - val_loss: 0.5133 - val_accuracy: 0.8298\n",
            "Epoch 1547/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 0.4908 - accuracy: 0.8400 - val_loss: 0.5179 - val_accuracy: 0.7872\n",
            "Epoch 1548/2000\n",
            "3/3 [==============================] - 0s 54ms/step - loss: 0.6515 - accuracy: 0.8300 - val_loss: 0.5005 - val_accuracy: 0.7872\n",
            "Epoch 1549/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 0.5818 - accuracy: 0.8100 - val_loss: 0.4916 - val_accuracy: 0.8085\n",
            "Epoch 1550/2000\n",
            "3/3 [==============================] - 0s 44ms/step - loss: 0.6404 - accuracy: 0.8800 - val_loss: 0.4852 - val_accuracy: 0.8085\n",
            "Epoch 1551/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 0.6676 - accuracy: 0.8500 - val_loss: 0.4831 - val_accuracy: 0.8298\n",
            "Epoch 1552/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 0.5138 - accuracy: 0.8400 - val_loss: 0.4746 - val_accuracy: 0.8085\n",
            "Epoch 1553/2000\n",
            "3/3 [==============================] - 0s 53ms/step - loss: 0.6416 - accuracy: 0.7900 - val_loss: 0.4717 - val_accuracy: 0.7872\n",
            "Epoch 1554/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 0.5149 - accuracy: 0.8400 - val_loss: 0.4963 - val_accuracy: 0.7872\n",
            "Epoch 1555/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 0.4940 - accuracy: 0.8300 - val_loss: 0.5419 - val_accuracy: 0.7660\n",
            "Epoch 1556/2000\n",
            "3/3 [==============================] - 0s 38ms/step - loss: 0.4545 - accuracy: 0.8600 - val_loss: 0.5635 - val_accuracy: 0.7660\n",
            "Epoch 1557/2000\n",
            "3/3 [==============================] - 0s 44ms/step - loss: 0.4591 - accuracy: 0.8400 - val_loss: 0.5624 - val_accuracy: 0.8298\n",
            "Epoch 1558/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 0.4949 - accuracy: 0.8100 - val_loss: 0.5549 - val_accuracy: 0.8298\n",
            "Epoch 1559/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 0.3834 - accuracy: 0.8900 - val_loss: 0.5335 - val_accuracy: 0.7872\n",
            "Epoch 1560/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 0.4923 - accuracy: 0.8500 - val_loss: 0.4897 - val_accuracy: 0.8298\n",
            "Epoch 1561/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 0.4723 - accuracy: 0.8600 - val_loss: 0.4694 - val_accuracy: 0.8085\n",
            "Epoch 1562/2000\n",
            "3/3 [==============================] - 0s 72ms/step - loss: 0.4948 - accuracy: 0.8200 - val_loss: 0.4677 - val_accuracy: 0.8298\n",
            "Epoch 1563/2000\n",
            "3/3 [==============================] - 0s 67ms/step - loss: 0.5497 - accuracy: 0.8300 - val_loss: 0.4539 - val_accuracy: 0.8085\n",
            "Epoch 1564/2000\n",
            "3/3 [==============================] - 0s 55ms/step - loss: 0.4878 - accuracy: 0.8700 - val_loss: 0.4630 - val_accuracy: 0.8298\n",
            "Epoch 1565/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 0.5878 - accuracy: 0.7800 - val_loss: 0.4865 - val_accuracy: 0.8298\n",
            "Epoch 1566/2000\n",
            "3/3 [==============================] - 0s 56ms/step - loss: 0.3611 - accuracy: 0.9100 - val_loss: 0.4900 - val_accuracy: 0.8085\n",
            "Epoch 1567/2000\n",
            "3/3 [==============================] - 0s 50ms/step - loss: 0.4895 - accuracy: 0.8000 - val_loss: 0.5048 - val_accuracy: 0.7872\n",
            "Epoch 1568/2000\n",
            "3/3 [==============================] - 0s 55ms/step - loss: 0.7259 - accuracy: 0.8000 - val_loss: 0.4927 - val_accuracy: 0.8085\n",
            "Epoch 1569/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 0.5473 - accuracy: 0.8000 - val_loss: 0.4798 - val_accuracy: 0.8085\n",
            "Epoch 1570/2000\n",
            "3/3 [==============================] - 0s 61ms/step - loss: 0.4146 - accuracy: 0.8400 - val_loss: 0.4694 - val_accuracy: 0.8085\n",
            "Epoch 1571/2000\n",
            "3/3 [==============================] - 0s 60ms/step - loss: 0.4963 - accuracy: 0.8300 - val_loss: 0.4556 - val_accuracy: 0.8085\n",
            "Epoch 1572/2000\n",
            "3/3 [==============================] - 0s 61ms/step - loss: 0.3983 - accuracy: 0.8600 - val_loss: 0.4513 - val_accuracy: 0.8085\n",
            "Epoch 1573/2000\n",
            "3/3 [==============================] - 0s 72ms/step - loss: 0.5619 - accuracy: 0.8300 - val_loss: 0.4477 - val_accuracy: 0.8085\n",
            "Epoch 1574/2000\n",
            "3/3 [==============================] - 0s 73ms/step - loss: 0.4621 - accuracy: 0.8400 - val_loss: 0.4693 - val_accuracy: 0.7872\n",
            "Epoch 1575/2000\n",
            "3/3 [==============================] - 0s 78ms/step - loss: 0.4453 - accuracy: 0.8500 - val_loss: 0.5014 - val_accuracy: 0.7660\n",
            "Epoch 1576/2000\n",
            "3/3 [==============================] - 0s 72ms/step - loss: 0.5413 - accuracy: 0.8300 - val_loss: 0.5200 - val_accuracy: 0.7447\n",
            "Epoch 1577/2000\n",
            "3/3 [==============================] - 0s 49ms/step - loss: 0.4281 - accuracy: 0.8600 - val_loss: 0.5385 - val_accuracy: 0.7447\n",
            "Epoch 1578/2000\n",
            "3/3 [==============================] - 0s 59ms/step - loss: 0.6969 - accuracy: 0.8100 - val_loss: 0.5321 - val_accuracy: 0.7447\n",
            "Epoch 1579/2000\n",
            "3/3 [==============================] - 0s 61ms/step - loss: 0.5725 - accuracy: 0.8400 - val_loss: 0.5197 - val_accuracy: 0.7872\n",
            "Epoch 1580/2000\n",
            "3/3 [==============================] - 0s 57ms/step - loss: 0.4981 - accuracy: 0.8400 - val_loss: 0.5014 - val_accuracy: 0.7872\n",
            "Epoch 1581/2000\n",
            "3/3 [==============================] - 0s 82ms/step - loss: 0.4217 - accuracy: 0.8400 - val_loss: 0.5037 - val_accuracy: 0.7872\n",
            "Epoch 1582/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 0.5975 - accuracy: 0.8100 - val_loss: 0.5000 - val_accuracy: 0.8085\n",
            "Epoch 1583/2000\n",
            "3/3 [==============================] - 0s 54ms/step - loss: 0.6669 - accuracy: 0.8200 - val_loss: 0.4930 - val_accuracy: 0.8085\n",
            "Epoch 1584/2000\n",
            "3/3 [==============================] - 0s 46ms/step - loss: 0.4524 - accuracy: 0.8400 - val_loss: 0.4849 - val_accuracy: 0.8298\n",
            "Epoch 1585/2000\n",
            "3/3 [==============================] - 0s 51ms/step - loss: 0.4592 - accuracy: 0.8800 - val_loss: 0.4567 - val_accuracy: 0.8511\n",
            "Epoch 1586/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 0.6379 - accuracy: 0.8000 - val_loss: 0.4636 - val_accuracy: 0.8511\n",
            "Epoch 1587/2000\n",
            "3/3 [==============================] - 0s 52ms/step - loss: 0.4903 - accuracy: 0.8600 - val_loss: 0.4823 - val_accuracy: 0.8511\n",
            "Epoch 1588/2000\n",
            "3/3 [==============================] - 0s 53ms/step - loss: 0.3700 - accuracy: 0.9100 - val_loss: 0.4785 - val_accuracy: 0.8298\n",
            "Epoch 1589/2000\n",
            "3/3 [==============================] - 0s 51ms/step - loss: 0.5308 - accuracy: 0.8400 - val_loss: 0.4718 - val_accuracy: 0.8085\n",
            "Epoch 1590/2000\n",
            "3/3 [==============================] - 0s 59ms/step - loss: 0.4677 - accuracy: 0.8300 - val_loss: 0.4743 - val_accuracy: 0.8085\n",
            "Epoch 1591/2000\n",
            "3/3 [==============================] - 0s 45ms/step - loss: 0.5765 - accuracy: 0.8400 - val_loss: 0.4798 - val_accuracy: 0.8298\n",
            "Epoch 1592/2000\n",
            "3/3 [==============================] - 0s 63ms/step - loss: 0.4555 - accuracy: 0.8900 - val_loss: 0.4862 - val_accuracy: 0.7872\n",
            "Epoch 1593/2000\n",
            "3/3 [==============================] - 0s 58ms/step - loss: 0.6910 - accuracy: 0.8400 - val_loss: 0.5026 - val_accuracy: 0.7660\n",
            "Epoch 1594/2000\n",
            "3/3 [==============================] - 0s 49ms/step - loss: 0.5389 - accuracy: 0.7800 - val_loss: 0.5199 - val_accuracy: 0.7872\n",
            "Epoch 1595/2000\n",
            "3/3 [==============================] - 0s 59ms/step - loss: 0.6064 - accuracy: 0.8300 - val_loss: 0.5283 - val_accuracy: 0.7660\n",
            "Epoch 1596/2000\n",
            "3/3 [==============================] - 0s 55ms/step - loss: 0.5083 - accuracy: 0.8500 - val_loss: 0.5485 - val_accuracy: 0.7447\n",
            "Epoch 1597/2000\n",
            "3/3 [==============================] - 0s 51ms/step - loss: 0.4422 - accuracy: 0.8200 - val_loss: 0.5721 - val_accuracy: 0.7234\n",
            "Epoch 1598/2000\n",
            "3/3 [==============================] - 0s 44ms/step - loss: 0.5249 - accuracy: 0.8200 - val_loss: 0.5655 - val_accuracy: 0.7447\n",
            "Epoch 1599/2000\n",
            "3/3 [==============================] - 0s 55ms/step - loss: 0.4840 - accuracy: 0.8600 - val_loss: 0.5708 - val_accuracy: 0.7872\n",
            "Epoch 1600/2000\n",
            "3/3 [==============================] - 0s 45ms/step - loss: 0.5659 - accuracy: 0.8200 - val_loss: 0.5567 - val_accuracy: 0.7447\n",
            "Epoch 1601/2000\n",
            "3/3 [==============================] - 0s 44ms/step - loss: 0.4540 - accuracy: 0.8500 - val_loss: 0.5229 - val_accuracy: 0.7447\n",
            "Epoch 1602/2000\n",
            "3/3 [==============================] - 0s 46ms/step - loss: 0.4553 - accuracy: 0.8500 - val_loss: 0.4847 - val_accuracy: 0.7872\n",
            "Epoch 1603/2000\n",
            "3/3 [==============================] - 0s 47ms/step - loss: 0.5075 - accuracy: 0.8300 - val_loss: 0.4479 - val_accuracy: 0.8298\n",
            "Epoch 1604/2000\n",
            "3/3 [==============================] - 0s 47ms/step - loss: 0.4166 - accuracy: 0.8600 - val_loss: 0.4186 - val_accuracy: 0.8723\n",
            "Epoch 1605/2000\n",
            "3/3 [==============================] - 0s 44ms/step - loss: 0.4643 - accuracy: 0.8400 - val_loss: 0.4085 - val_accuracy: 0.8511\n",
            "Epoch 1606/2000\n",
            "3/3 [==============================] - 0s 63ms/step - loss: 0.4482 - accuracy: 0.8500 - val_loss: 0.4027 - val_accuracy: 0.8723\n",
            "Epoch 1607/2000\n",
            "3/3 [==============================] - 0s 44ms/step - loss: 0.4817 - accuracy: 0.8300 - val_loss: 0.3996 - val_accuracy: 0.8723\n",
            "Epoch 1608/2000\n",
            "3/3 [==============================] - 0s 47ms/step - loss: 0.6603 - accuracy: 0.8400 - val_loss: 0.4091 - val_accuracy: 0.8511\n",
            "Epoch 1609/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 0.3597 - accuracy: 0.8800 - val_loss: 0.4541 - val_accuracy: 0.8511\n",
            "Epoch 1610/2000\n",
            "3/3 [==============================] - 0s 51ms/step - loss: 0.4501 - accuracy: 0.8300 - val_loss: 0.4894 - val_accuracy: 0.8085\n",
            "Epoch 1611/2000\n",
            "3/3 [==============================] - 0s 56ms/step - loss: 0.5594 - accuracy: 0.8500 - val_loss: 0.4605 - val_accuracy: 0.8085\n",
            "Epoch 1612/2000\n",
            "3/3 [==============================] - 0s 59ms/step - loss: 0.5857 - accuracy: 0.7900 - val_loss: 0.4516 - val_accuracy: 0.8085\n",
            "Epoch 1613/2000\n",
            "3/3 [==============================] - 0s 44ms/step - loss: 0.4835 - accuracy: 0.8400 - val_loss: 0.4335 - val_accuracy: 0.8085\n",
            "Epoch 1614/2000\n",
            "3/3 [==============================] - 0s 59ms/step - loss: 0.4037 - accuracy: 0.8900 - val_loss: 0.4164 - val_accuracy: 0.8085\n",
            "Epoch 1615/2000\n",
            "3/3 [==============================] - 0s 56ms/step - loss: 0.4009 - accuracy: 0.8500 - val_loss: 0.4053 - val_accuracy: 0.7872\n",
            "Epoch 1616/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 0.4953 - accuracy: 0.8700 - val_loss: 0.4008 - val_accuracy: 0.8298\n",
            "Epoch 1617/2000\n",
            "3/3 [==============================] - 0s 44ms/step - loss: 0.4783 - accuracy: 0.8500 - val_loss: 0.4031 - val_accuracy: 0.8298\n",
            "Epoch 1618/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 0.3914 - accuracy: 0.8700 - val_loss: 0.4033 - val_accuracy: 0.8511\n",
            "Epoch 1619/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 0.3499 - accuracy: 0.9000 - val_loss: 0.3968 - val_accuracy: 0.8511\n",
            "Epoch 1620/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 0.6366 - accuracy: 0.7900 - val_loss: 0.3926 - val_accuracy: 0.8511\n",
            "Epoch 1621/2000\n",
            "3/3 [==============================] - 0s 50ms/step - loss: 0.4769 - accuracy: 0.8700 - val_loss: 0.4033 - val_accuracy: 0.8511\n",
            "Epoch 1622/2000\n",
            "3/3 [==============================] - 0s 55ms/step - loss: 0.5718 - accuracy: 0.8200 - val_loss: 0.4116 - val_accuracy: 0.8723\n",
            "Epoch 1623/2000\n",
            "3/3 [==============================] - 0s 53ms/step - loss: 0.5539 - accuracy: 0.7700 - val_loss: 0.4268 - val_accuracy: 0.8085\n",
            "Epoch 1624/2000\n",
            "3/3 [==============================] - 0s 54ms/step - loss: 0.3900 - accuracy: 0.8300 - val_loss: 0.4446 - val_accuracy: 0.8085\n",
            "Epoch 1625/2000\n",
            "3/3 [==============================] - 0s 47ms/step - loss: 0.4064 - accuracy: 0.8400 - val_loss: 0.4646 - val_accuracy: 0.8085\n",
            "Epoch 1626/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 0.5744 - accuracy: 0.8100 - val_loss: 0.4783 - val_accuracy: 0.8298\n",
            "Epoch 1627/2000\n",
            "3/3 [==============================] - 0s 50ms/step - loss: 0.5431 - accuracy: 0.8500 - val_loss: 0.4864 - val_accuracy: 0.8085\n",
            "Epoch 1628/2000\n",
            "3/3 [==============================] - 0s 46ms/step - loss: 0.7461 - accuracy: 0.7800 - val_loss: 0.4877 - val_accuracy: 0.8085\n",
            "Epoch 1629/2000\n",
            "3/3 [==============================] - 0s 53ms/step - loss: 0.6665 - accuracy: 0.7800 - val_loss: 0.4889 - val_accuracy: 0.8085\n",
            "Epoch 1630/2000\n",
            "3/3 [==============================] - 0s 44ms/step - loss: 0.6316 - accuracy: 0.8600 - val_loss: 0.4914 - val_accuracy: 0.8298\n",
            "Epoch 1631/2000\n",
            "3/3 [==============================] - 0s 70ms/step - loss: 0.4834 - accuracy: 0.8200 - val_loss: 0.4927 - val_accuracy: 0.8511\n",
            "Epoch 1632/2000\n",
            "3/3 [==============================] - 0s 73ms/step - loss: 0.4415 - accuracy: 0.8800 - val_loss: 0.5057 - val_accuracy: 0.8298\n",
            "Epoch 1633/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 0.5323 - accuracy: 0.8300 - val_loss: 0.4890 - val_accuracy: 0.8298\n",
            "Epoch 1634/2000\n",
            "3/3 [==============================] - 0s 56ms/step - loss: 0.5344 - accuracy: 0.7900 - val_loss: 0.4662 - val_accuracy: 0.8723\n",
            "Epoch 1635/2000\n",
            "3/3 [==============================] - 0s 59ms/step - loss: 0.5987 - accuracy: 0.8000 - val_loss: 0.4708 - val_accuracy: 0.8723\n",
            "Epoch 1636/2000\n",
            "3/3 [==============================] - 0s 67ms/step - loss: 0.4352 - accuracy: 0.8300 - val_loss: 0.4712 - val_accuracy: 0.8298\n",
            "Epoch 1637/2000\n",
            "3/3 [==============================] - 0s 66ms/step - loss: 0.4741 - accuracy: 0.8300 - val_loss: 0.4574 - val_accuracy: 0.8298\n",
            "Epoch 1638/2000\n",
            "3/3 [==============================] - 0s 65ms/step - loss: 0.5410 - accuracy: 0.8100 - val_loss: 0.4469 - val_accuracy: 0.8298\n",
            "Epoch 1639/2000\n",
            "3/3 [==============================] - 0s 69ms/step - loss: 0.4865 - accuracy: 0.8100 - val_loss: 0.4115 - val_accuracy: 0.8298\n",
            "Epoch 1640/2000\n",
            "3/3 [==============================] - 0s 56ms/step - loss: 0.4908 - accuracy: 0.8000 - val_loss: 0.3985 - val_accuracy: 0.8511\n",
            "Epoch 1641/2000\n",
            "3/3 [==============================] - 0s 52ms/step - loss: 0.4771 - accuracy: 0.8500 - val_loss: 0.4059 - val_accuracy: 0.8511\n",
            "Epoch 1642/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 0.5325 - accuracy: 0.7900 - val_loss: 0.4154 - val_accuracy: 0.8511\n",
            "Epoch 1643/2000\n",
            "3/3 [==============================] - 0s 45ms/step - loss: 0.4198 - accuracy: 0.8200 - val_loss: 0.4320 - val_accuracy: 0.8298\n",
            "Epoch 1644/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 0.3805 - accuracy: 0.8600 - val_loss: 0.4374 - val_accuracy: 0.8298\n",
            "Epoch 1645/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 0.4909 - accuracy: 0.8700 - val_loss: 0.4481 - val_accuracy: 0.8085\n",
            "Epoch 1646/2000\n",
            "3/3 [==============================] - 0s 45ms/step - loss: 0.5376 - accuracy: 0.8600 - val_loss: 0.4460 - val_accuracy: 0.8085\n",
            "Epoch 1647/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 0.5092 - accuracy: 0.8600 - val_loss: 0.4462 - val_accuracy: 0.7872\n",
            "Epoch 1648/2000\n",
            "3/3 [==============================] - 0s 44ms/step - loss: 0.5840 - accuracy: 0.8200 - val_loss: 0.4642 - val_accuracy: 0.8085\n",
            "Epoch 1649/2000\n",
            "3/3 [==============================] - 0s 34ms/step - loss: 0.3909 - accuracy: 0.8500 - val_loss: 0.4856 - val_accuracy: 0.8085\n",
            "Epoch 1650/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 0.5697 - accuracy: 0.8200 - val_loss: 0.4862 - val_accuracy: 0.7872\n",
            "Epoch 1651/2000\n",
            "3/3 [==============================] - 0s 61ms/step - loss: 0.4705 - accuracy: 0.8300 - val_loss: 0.4961 - val_accuracy: 0.8085\n",
            "Epoch 1652/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 0.6114 - accuracy: 0.7900 - val_loss: 0.5175 - val_accuracy: 0.8085\n",
            "Epoch 1653/2000\n",
            "3/3 [==============================] - 0s 38ms/step - loss: 0.5133 - accuracy: 0.8100 - val_loss: 0.5106 - val_accuracy: 0.7872\n",
            "Epoch 1654/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 0.6968 - accuracy: 0.8400 - val_loss: 0.4609 - val_accuracy: 0.7872\n",
            "Epoch 1655/2000\n",
            "3/3 [==============================] - 0s 54ms/step - loss: 0.4770 - accuracy: 0.8300 - val_loss: 0.4467 - val_accuracy: 0.8298\n",
            "Epoch 1656/2000\n",
            "3/3 [==============================] - 0s 38ms/step - loss: 0.3837 - accuracy: 0.8700 - val_loss: 0.4513 - val_accuracy: 0.8298\n",
            "Epoch 1657/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 0.4889 - accuracy: 0.8300 - val_loss: 0.4632 - val_accuracy: 0.7872\n",
            "Epoch 1658/2000\n",
            "3/3 [==============================] - 0s 49ms/step - loss: 0.5561 - accuracy: 0.8800 - val_loss: 0.4748 - val_accuracy: 0.7872\n",
            "Epoch 1659/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 0.5344 - accuracy: 0.8500 - val_loss: 0.4831 - val_accuracy: 0.7872\n",
            "Epoch 1660/2000\n",
            "3/3 [==============================] - 0s 38ms/step - loss: 0.5071 - accuracy: 0.8200 - val_loss: 0.4759 - val_accuracy: 0.7660\n",
            "Epoch 1661/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 0.5764 - accuracy: 0.8800 - val_loss: 0.4774 - val_accuracy: 0.7872\n",
            "Epoch 1662/2000\n",
            "3/3 [==============================] - 0s 34ms/step - loss: 0.5234 - accuracy: 0.8500 - val_loss: 0.4528 - val_accuracy: 0.7872\n",
            "Epoch 1663/2000\n",
            "3/3 [==============================] - 0s 46ms/step - loss: 0.5182 - accuracy: 0.8100 - val_loss: 0.4415 - val_accuracy: 0.8085\n",
            "Epoch 1664/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 0.5184 - accuracy: 0.8300 - val_loss: 0.4483 - val_accuracy: 0.8085\n",
            "Epoch 1665/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 0.5512 - accuracy: 0.8100 - val_loss: 0.4734 - val_accuracy: 0.7872\n",
            "Epoch 1666/2000\n",
            "3/3 [==============================] - 0s 57ms/step - loss: 0.5461 - accuracy: 0.7700 - val_loss: 0.5031 - val_accuracy: 0.7660\n",
            "Epoch 1667/2000\n",
            "3/3 [==============================] - 0s 64ms/step - loss: 0.4070 - accuracy: 0.8900 - val_loss: 0.5083 - val_accuracy: 0.7872\n",
            "Epoch 1668/2000\n",
            "3/3 [==============================] - 0s 57ms/step - loss: 0.4356 - accuracy: 0.8700 - val_loss: 0.4717 - val_accuracy: 0.7872\n",
            "Epoch 1669/2000\n",
            "3/3 [==============================] - 0s 51ms/step - loss: 0.4036 - accuracy: 0.9200 - val_loss: 0.4591 - val_accuracy: 0.8511\n",
            "Epoch 1670/2000\n",
            "3/3 [==============================] - 0s 44ms/step - loss: 0.5415 - accuracy: 0.8100 - val_loss: 0.4528 - val_accuracy: 0.8085\n",
            "Epoch 1671/2000\n",
            "3/3 [==============================] - 0s 65ms/step - loss: 0.5207 - accuracy: 0.9000 - val_loss: 0.4560 - val_accuracy: 0.8085\n",
            "Epoch 1672/2000\n",
            "3/3 [==============================] - 0s 52ms/step - loss: 0.4877 - accuracy: 0.8400 - val_loss: 0.4589 - val_accuracy: 0.8085\n",
            "Epoch 1673/2000\n",
            "3/3 [==============================] - 0s 55ms/step - loss: 0.5990 - accuracy: 0.7800 - val_loss: 0.4483 - val_accuracy: 0.8298\n",
            "Epoch 1674/2000\n",
            "3/3 [==============================] - 0s 53ms/step - loss: 0.5447 - accuracy: 0.8300 - val_loss: 0.4166 - val_accuracy: 0.8298\n",
            "Epoch 1675/2000\n",
            "3/3 [==============================] - 0s 55ms/step - loss: 0.4484 - accuracy: 0.8400 - val_loss: 0.3997 - val_accuracy: 0.8511\n",
            "Epoch 1676/2000\n",
            "3/3 [==============================] - 0s 47ms/step - loss: 0.4129 - accuracy: 0.8700 - val_loss: 0.4028 - val_accuracy: 0.8298\n",
            "Epoch 1677/2000\n",
            "3/3 [==============================] - 0s 45ms/step - loss: 0.5540 - accuracy: 0.8300 - val_loss: 0.4116 - val_accuracy: 0.8511\n",
            "Epoch 1678/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 0.7362 - accuracy: 0.7900 - val_loss: 0.4262 - val_accuracy: 0.8298\n",
            "Epoch 1679/2000\n",
            "3/3 [==============================] - 0s 52ms/step - loss: 0.3934 - accuracy: 0.8500 - val_loss: 0.4379 - val_accuracy: 0.8511\n",
            "Epoch 1680/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 0.4576 - accuracy: 0.8600 - val_loss: 0.4343 - val_accuracy: 0.8298\n",
            "Epoch 1681/2000\n",
            "3/3 [==============================] - 0s 44ms/step - loss: 0.4257 - accuracy: 0.8300 - val_loss: 0.4267 - val_accuracy: 0.8298\n",
            "Epoch 1682/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 0.5123 - accuracy: 0.8000 - val_loss: 0.4224 - val_accuracy: 0.8298\n",
            "Epoch 1683/2000\n",
            "3/3 [==============================] - 0s 57ms/step - loss: 0.4181 - accuracy: 0.8400 - val_loss: 0.4310 - val_accuracy: 0.8085\n",
            "Epoch 1684/2000\n",
            "3/3 [==============================] - 0s 46ms/step - loss: 0.4414 - accuracy: 0.8300 - val_loss: 0.4385 - val_accuracy: 0.8298\n",
            "Epoch 1685/2000\n",
            "3/3 [==============================] - 0s 45ms/step - loss: 0.5024 - accuracy: 0.8300 - val_loss: 0.4345 - val_accuracy: 0.8085\n",
            "Epoch 1686/2000\n",
            "3/3 [==============================] - 0s 57ms/step - loss: 0.4373 - accuracy: 0.8600 - val_loss: 0.4459 - val_accuracy: 0.8085\n",
            "Epoch 1687/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 0.5217 - accuracy: 0.8300 - val_loss: 0.4649 - val_accuracy: 0.7872\n",
            "Epoch 1688/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 0.4435 - accuracy: 0.8200 - val_loss: 0.4812 - val_accuracy: 0.8085\n",
            "Epoch 1689/2000\n",
            "3/3 [==============================] - 0s 52ms/step - loss: 0.4804 - accuracy: 0.8100 - val_loss: 0.5002 - val_accuracy: 0.7872\n",
            "Epoch 1690/2000\n",
            "3/3 [==============================] - 0s 49ms/step - loss: 0.3655 - accuracy: 0.8900 - val_loss: 0.4915 - val_accuracy: 0.7660\n",
            "Epoch 1691/2000\n",
            "3/3 [==============================] - 0s 38ms/step - loss: 0.4157 - accuracy: 0.8700 - val_loss: 0.4620 - val_accuracy: 0.8085\n",
            "Epoch 1692/2000\n",
            "3/3 [==============================] - 0s 53ms/step - loss: 0.3959 - accuracy: 0.8700 - val_loss: 0.4467 - val_accuracy: 0.8298\n",
            "Epoch 1693/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 0.3498 - accuracy: 0.8900 - val_loss: 0.4467 - val_accuracy: 0.8298\n",
            "Epoch 1694/2000\n",
            "3/3 [==============================] - 0s 47ms/step - loss: 0.6712 - accuracy: 0.7800 - val_loss: 0.4395 - val_accuracy: 0.8511\n",
            "Epoch 1695/2000\n",
            "3/3 [==============================] - 0s 55ms/step - loss: 0.4652 - accuracy: 0.8500 - val_loss: 0.4372 - val_accuracy: 0.8298\n",
            "Epoch 1696/2000\n",
            "3/3 [==============================] - 0s 59ms/step - loss: 0.6916 - accuracy: 0.8100 - val_loss: 0.4437 - val_accuracy: 0.8298\n",
            "Epoch 1697/2000\n",
            "3/3 [==============================] - 0s 79ms/step - loss: 0.5388 - accuracy: 0.8600 - val_loss: 0.4313 - val_accuracy: 0.8511\n",
            "Epoch 1698/2000\n",
            "3/3 [==============================] - 0s 66ms/step - loss: 0.4499 - accuracy: 0.8400 - val_loss: 0.4274 - val_accuracy: 0.8723\n",
            "Epoch 1699/2000\n",
            "3/3 [==============================] - 0s 64ms/step - loss: 0.6027 - accuracy: 0.7800 - val_loss: 0.4382 - val_accuracy: 0.8298\n",
            "Epoch 1700/2000\n",
            "3/3 [==============================] - 0s 58ms/step - loss: 0.5228 - accuracy: 0.8400 - val_loss: 0.4462 - val_accuracy: 0.8511\n",
            "Epoch 1701/2000\n",
            "3/3 [==============================] - 0s 66ms/step - loss: 0.4733 - accuracy: 0.8300 - val_loss: 0.4702 - val_accuracy: 0.8298\n",
            "Epoch 1702/2000\n",
            "3/3 [==============================] - 0s 53ms/step - loss: 0.4259 - accuracy: 0.8500 - val_loss: 0.4754 - val_accuracy: 0.8298\n",
            "Epoch 1703/2000\n",
            "3/3 [==============================] - 0s 65ms/step - loss: 0.4120 - accuracy: 0.8400 - val_loss: 0.4643 - val_accuracy: 0.8511\n",
            "Epoch 1704/2000\n",
            "3/3 [==============================] - 0s 66ms/step - loss: 0.5024 - accuracy: 0.8600 - val_loss: 0.4433 - val_accuracy: 0.8723\n",
            "Epoch 1705/2000\n",
            "3/3 [==============================] - 0s 79ms/step - loss: 0.4120 - accuracy: 0.8200 - val_loss: 0.4527 - val_accuracy: 0.8723\n",
            "Epoch 1706/2000\n",
            "3/3 [==============================] - 0s 81ms/step - loss: 0.4399 - accuracy: 0.8600 - val_loss: 0.4837 - val_accuracy: 0.8085\n",
            "Epoch 1707/2000\n",
            "3/3 [==============================] - 0s 53ms/step - loss: 0.3467 - accuracy: 0.8800 - val_loss: 0.5137 - val_accuracy: 0.7872\n",
            "Epoch 1708/2000\n",
            "3/3 [==============================] - 0s 53ms/step - loss: 0.3736 - accuracy: 0.8600 - val_loss: 0.5399 - val_accuracy: 0.7872\n",
            "Epoch 1709/2000\n",
            "3/3 [==============================] - 0s 44ms/step - loss: 0.5045 - accuracy: 0.8000 - val_loss: 0.5337 - val_accuracy: 0.7872\n",
            "Epoch 1710/2000\n",
            "3/3 [==============================] - 0s 49ms/step - loss: 0.5005 - accuracy: 0.8200 - val_loss: 0.5430 - val_accuracy: 0.7447\n",
            "Epoch 1711/2000\n",
            "3/3 [==============================] - 0s 39ms/step - loss: 0.2748 - accuracy: 0.9300 - val_loss: 0.5393 - val_accuracy: 0.7872\n",
            "Epoch 1712/2000\n",
            "3/3 [==============================] - 0s 50ms/step - loss: 0.3885 - accuracy: 0.9000 - val_loss: 0.5265 - val_accuracy: 0.8085\n",
            "Epoch 1713/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 0.4228 - accuracy: 0.8700 - val_loss: 0.5001 - val_accuracy: 0.8085\n",
            "Epoch 1714/2000\n",
            "3/3 [==============================] - 0s 49ms/step - loss: 0.4102 - accuracy: 0.8800 - val_loss: 0.4978 - val_accuracy: 0.7660\n",
            "Epoch 1715/2000\n",
            "3/3 [==============================] - 0s 46ms/step - loss: 0.3761 - accuracy: 0.9000 - val_loss: 0.5106 - val_accuracy: 0.7660\n",
            "Epoch 1716/2000\n",
            "3/3 [==============================] - 0s 54ms/step - loss: 0.3772 - accuracy: 0.9000 - val_loss: 0.5172 - val_accuracy: 0.7660\n",
            "Epoch 1717/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 0.6599 - accuracy: 0.7900 - val_loss: 0.4900 - val_accuracy: 0.7872\n",
            "Epoch 1718/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 0.5605 - accuracy: 0.8100 - val_loss: 0.4783 - val_accuracy: 0.8085\n",
            "Epoch 1719/2000\n",
            "3/3 [==============================] - 0s 39ms/step - loss: 0.5195 - accuracy: 0.7800 - val_loss: 0.4952 - val_accuracy: 0.8085\n",
            "Epoch 1720/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 0.4469 - accuracy: 0.8400 - val_loss: 0.5105 - val_accuracy: 0.7872\n",
            "Epoch 1721/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 0.4985 - accuracy: 0.8400 - val_loss: 0.5030 - val_accuracy: 0.7660\n",
            "Epoch 1722/2000\n",
            "3/3 [==============================] - 0s 44ms/step - loss: 0.5839 - accuracy: 0.8700 - val_loss: 0.4962 - val_accuracy: 0.7234\n",
            "Epoch 1723/2000\n",
            "3/3 [==============================] - 0s 52ms/step - loss: 0.5398 - accuracy: 0.8300 - val_loss: 0.4956 - val_accuracy: 0.7447\n",
            "Epoch 1724/2000\n",
            "3/3 [==============================] - 0s 51ms/step - loss: 0.4916 - accuracy: 0.8700 - val_loss: 0.5228 - val_accuracy: 0.7447\n",
            "Epoch 1725/2000\n",
            "3/3 [==============================] - 0s 38ms/step - loss: 0.4669 - accuracy: 0.8600 - val_loss: 0.5655 - val_accuracy: 0.7234\n",
            "Epoch 1726/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 0.4167 - accuracy: 0.8400 - val_loss: 0.6006 - val_accuracy: 0.7234\n",
            "Epoch 1727/2000\n",
            "3/3 [==============================] - 0s 38ms/step - loss: 0.5270 - accuracy: 0.8100 - val_loss: 0.5671 - val_accuracy: 0.7447\n",
            "Epoch 1728/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 0.4655 - accuracy: 0.8100 - val_loss: 0.5540 - val_accuracy: 0.7660\n",
            "Epoch 1729/2000\n",
            "3/3 [==============================] - 0s 50ms/step - loss: 0.3703 - accuracy: 0.8800 - val_loss: 0.5821 - val_accuracy: 0.7447\n",
            "Epoch 1730/2000\n",
            "3/3 [==============================] - 0s 58ms/step - loss: 0.6230 - accuracy: 0.7800 - val_loss: 0.6196 - val_accuracy: 0.7660\n",
            "Epoch 1731/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 0.5753 - accuracy: 0.8100 - val_loss: 0.5807 - val_accuracy: 0.7447\n",
            "Epoch 1732/2000\n",
            "3/3 [==============================] - 0s 46ms/step - loss: 0.5711 - accuracy: 0.8200 - val_loss: 0.5370 - val_accuracy: 0.7447\n",
            "Epoch 1733/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 0.4736 - accuracy: 0.8600 - val_loss: 0.5104 - val_accuracy: 0.8085\n",
            "Epoch 1734/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 0.4300 - accuracy: 0.8700 - val_loss: 0.5105 - val_accuracy: 0.8298\n",
            "Epoch 1735/2000\n",
            "3/3 [==============================] - 0s 46ms/step - loss: 0.6130 - accuracy: 0.8500 - val_loss: 0.5291 - val_accuracy: 0.7872\n",
            "Epoch 1736/2000\n",
            "3/3 [==============================] - 0s 45ms/step - loss: 0.6344 - accuracy: 0.7600 - val_loss: 0.5382 - val_accuracy: 0.7872\n",
            "Epoch 1737/2000\n",
            "3/3 [==============================] - 0s 64ms/step - loss: 0.4571 - accuracy: 0.8300 - val_loss: 0.5371 - val_accuracy: 0.7447\n",
            "Epoch 1738/2000\n",
            "3/3 [==============================] - 0s 47ms/step - loss: 0.4222 - accuracy: 0.8400 - val_loss: 0.5236 - val_accuracy: 0.7660\n",
            "Epoch 1739/2000\n",
            "3/3 [==============================] - 0s 58ms/step - loss: 0.6019 - accuracy: 0.8300 - val_loss: 0.4849 - val_accuracy: 0.7872\n",
            "Epoch 1740/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 0.3920 - accuracy: 0.8700 - val_loss: 0.4725 - val_accuracy: 0.7872\n",
            "Epoch 1741/2000\n",
            "3/3 [==============================] - 0s 56ms/step - loss: 0.4168 - accuracy: 0.8700 - val_loss: 0.4777 - val_accuracy: 0.8085\n",
            "Epoch 1742/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 0.5263 - accuracy: 0.8200 - val_loss: 0.4601 - val_accuracy: 0.7872\n",
            "Epoch 1743/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 0.5013 - accuracy: 0.8200 - val_loss: 0.4427 - val_accuracy: 0.8085\n",
            "Epoch 1744/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 0.5037 - accuracy: 0.8600 - val_loss: 0.4218 - val_accuracy: 0.8298\n",
            "Epoch 1745/2000\n",
            "3/3 [==============================] - 0s 58ms/step - loss: 0.4950 - accuracy: 0.8400 - val_loss: 0.4075 - val_accuracy: 0.8511\n",
            "Epoch 1746/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 0.4444 - accuracy: 0.8800 - val_loss: 0.4057 - val_accuracy: 0.8511\n",
            "Epoch 1747/2000\n",
            "3/3 [==============================] - 0s 60ms/step - loss: 0.4640 - accuracy: 0.8600 - val_loss: 0.4332 - val_accuracy: 0.8298\n",
            "Epoch 1748/2000\n",
            "3/3 [==============================] - 0s 50ms/step - loss: 0.4778 - accuracy: 0.8100 - val_loss: 0.4666 - val_accuracy: 0.8298\n",
            "Epoch 1749/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 0.4875 - accuracy: 0.8600 - val_loss: 0.4803 - val_accuracy: 0.8298\n",
            "Epoch 1750/2000\n",
            "3/3 [==============================] - 0s 45ms/step - loss: 0.5081 - accuracy: 0.8100 - val_loss: 0.4811 - val_accuracy: 0.8298\n",
            "Epoch 1751/2000\n",
            "3/3 [==============================] - 0s 40ms/step - loss: 0.5627 - accuracy: 0.8100 - val_loss: 0.4701 - val_accuracy: 0.8085\n",
            "Epoch 1752/2000\n",
            "3/3 [==============================] - 0s 30ms/step - loss: 0.5524 - accuracy: 0.8300 - val_loss: 0.4623 - val_accuracy: 0.7872\n",
            "Epoch 1753/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 0.3905 - accuracy: 0.9000 - val_loss: 0.4743 - val_accuracy: 0.7872\n",
            "Epoch 1754/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 0.4984 - accuracy: 0.8300 - val_loss: 0.4960 - val_accuracy: 0.7660\n",
            "Epoch 1755/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 0.4361 - accuracy: 0.8700 - val_loss: 0.5127 - val_accuracy: 0.7872\n",
            "Epoch 1756/2000\n",
            "3/3 [==============================] - 0s 61ms/step - loss: 0.4840 - accuracy: 0.8000 - val_loss: 0.5164 - val_accuracy: 0.8085\n",
            "Epoch 1757/2000\n",
            "3/3 [==============================] - 0s 53ms/step - loss: 0.5091 - accuracy: 0.8500 - val_loss: 0.4957 - val_accuracy: 0.8298\n",
            "Epoch 1758/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 0.4124 - accuracy: 0.8500 - val_loss: 0.4821 - val_accuracy: 0.8085\n",
            "Epoch 1759/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 0.4526 - accuracy: 0.8400 - val_loss: 0.4816 - val_accuracy: 0.7872\n",
            "Epoch 1760/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 0.3506 - accuracy: 0.8700 - val_loss: 0.4856 - val_accuracy: 0.7872\n",
            "Epoch 1761/2000\n",
            "3/3 [==============================] - 0s 48ms/step - loss: 0.4928 - accuracy: 0.8300 - val_loss: 0.4835 - val_accuracy: 0.7872\n",
            "Epoch 1762/2000\n",
            "3/3 [==============================] - 0s 54ms/step - loss: 0.4513 - accuracy: 0.8700 - val_loss: 0.4918 - val_accuracy: 0.7872\n",
            "Epoch 1763/2000\n",
            "3/3 [==============================] - 0s 44ms/step - loss: 0.3785 - accuracy: 0.8700 - val_loss: 0.4846 - val_accuracy: 0.8298\n",
            "Epoch 1764/2000\n",
            "3/3 [==============================] - 0s 44ms/step - loss: 0.5721 - accuracy: 0.8400 - val_loss: 0.4724 - val_accuracy: 0.8723\n",
            "Epoch 1765/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 0.4894 - accuracy: 0.8800 - val_loss: 0.4700 - val_accuracy: 0.8723\n",
            "Epoch 1766/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 0.6546 - accuracy: 0.8400 - val_loss: 0.4802 - val_accuracy: 0.8085\n",
            "Epoch 1767/2000\n",
            "3/3 [==============================] - 0s 49ms/step - loss: 0.4358 - accuracy: 0.8600 - val_loss: 0.4744 - val_accuracy: 0.8298\n",
            "Epoch 1768/2000\n",
            "3/3 [==============================] - 0s 69ms/step - loss: 0.4576 - accuracy: 0.8400 - val_loss: 0.4633 - val_accuracy: 0.8298\n",
            "Epoch 1769/2000\n",
            "3/3 [==============================] - 0s 65ms/step - loss: 0.6274 - accuracy: 0.8300 - val_loss: 0.4746 - val_accuracy: 0.7872\n",
            "Epoch 1770/2000\n",
            "3/3 [==============================] - 0s 61ms/step - loss: 0.3781 - accuracy: 0.8500 - val_loss: 0.4784 - val_accuracy: 0.7872\n",
            "Epoch 1771/2000\n",
            "3/3 [==============================] - 0s 66ms/step - loss: 0.5405 - accuracy: 0.8300 - val_loss: 0.4752 - val_accuracy: 0.8085\n",
            "Epoch 1772/2000\n",
            "3/3 [==============================] - 0s 47ms/step - loss: 0.5009 - accuracy: 0.8500 - val_loss: 0.4890 - val_accuracy: 0.7872\n",
            "Epoch 1773/2000\n",
            "3/3 [==============================] - 0s 53ms/step - loss: 0.6006 - accuracy: 0.8000 - val_loss: 0.5034 - val_accuracy: 0.7660\n",
            "Epoch 1774/2000\n",
            "3/3 [==============================] - 0s 68ms/step - loss: 0.6781 - accuracy: 0.7900 - val_loss: 0.4991 - val_accuracy: 0.7872\n",
            "Epoch 1775/2000\n",
            "3/3 [==============================] - 0s 76ms/step - loss: 0.4431 - accuracy: 0.8600 - val_loss: 0.4961 - val_accuracy: 0.7872\n",
            "Epoch 1776/2000\n",
            "3/3 [==============================] - 0s 50ms/step - loss: 0.4048 - accuracy: 0.8900 - val_loss: 0.5052 - val_accuracy: 0.7872\n",
            "Epoch 1777/2000\n",
            "3/3 [==============================] - 0s 80ms/step - loss: 0.4655 - accuracy: 0.8300 - val_loss: 0.5492 - val_accuracy: 0.8298\n",
            "Epoch 1778/2000\n",
            "3/3 [==============================] - 0s 89ms/step - loss: 0.3916 - accuracy: 0.8200 - val_loss: 0.5819 - val_accuracy: 0.8298\n",
            "Epoch 1779/2000\n",
            "3/3 [==============================] - 0s 47ms/step - loss: 0.5688 - accuracy: 0.8100 - val_loss: 0.5974 - val_accuracy: 0.8298\n",
            "Epoch 1780/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 0.4558 - accuracy: 0.8100 - val_loss: 0.5868 - val_accuracy: 0.7660\n",
            "Epoch 1781/2000\n",
            "3/3 [==============================] - 0s 55ms/step - loss: 0.5705 - accuracy: 0.8300 - val_loss: 0.5620 - val_accuracy: 0.7872\n",
            "Epoch 1782/2000\n",
            "3/3 [==============================] - 0s 58ms/step - loss: 0.5390 - accuracy: 0.8400 - val_loss: 0.5226 - val_accuracy: 0.7872\n",
            "Epoch 1783/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 0.5303 - accuracy: 0.8700 - val_loss: 0.4954 - val_accuracy: 0.7872\n",
            "Epoch 1784/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 0.4623 - accuracy: 0.8300 - val_loss: 0.4977 - val_accuracy: 0.8298\n",
            "Epoch 1785/2000\n",
            "3/3 [==============================] - 0s 46ms/step - loss: 0.5151 - accuracy: 0.8000 - val_loss: 0.4881 - val_accuracy: 0.8085\n",
            "Epoch 1786/2000\n",
            "3/3 [==============================] - 0s 44ms/step - loss: 0.5700 - accuracy: 0.7800 - val_loss: 0.4918 - val_accuracy: 0.8085\n",
            "Epoch 1787/2000\n",
            "3/3 [==============================] - 0s 46ms/step - loss: 0.5245 - accuracy: 0.8200 - val_loss: 0.4603 - val_accuracy: 0.8298\n",
            "Epoch 1788/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 0.5701 - accuracy: 0.8000 - val_loss: 0.4461 - val_accuracy: 0.8511\n",
            "Epoch 1789/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 0.4469 - accuracy: 0.8200 - val_loss: 0.4503 - val_accuracy: 0.8511\n",
            "Epoch 1790/2000\n",
            "3/3 [==============================] - 0s 51ms/step - loss: 0.3941 - accuracy: 0.9000 - val_loss: 0.4751 - val_accuracy: 0.8723\n",
            "Epoch 1791/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 0.5539 - accuracy: 0.8500 - val_loss: 0.5041 - val_accuracy: 0.8085\n",
            "Epoch 1792/2000\n",
            "3/3 [==============================] - 0s 53ms/step - loss: 0.3342 - accuracy: 0.8700 - val_loss: 0.5260 - val_accuracy: 0.7872\n",
            "Epoch 1793/2000\n",
            "3/3 [==============================] - 0s 60ms/step - loss: 0.5450 - accuracy: 0.8500 - val_loss: 0.5464 - val_accuracy: 0.7660\n",
            "Epoch 1794/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 0.4162 - accuracy: 0.8900 - val_loss: 0.5694 - val_accuracy: 0.7660\n",
            "Epoch 1795/2000\n",
            "3/3 [==============================] - 0s 46ms/step - loss: 0.7603 - accuracy: 0.7600 - val_loss: 0.6054 - val_accuracy: 0.7660\n",
            "Epoch 1796/2000\n",
            "3/3 [==============================] - 0s 44ms/step - loss: 0.5673 - accuracy: 0.8400 - val_loss: 0.6127 - val_accuracy: 0.7872\n",
            "Epoch 1797/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 0.5139 - accuracy: 0.8300 - val_loss: 0.5864 - val_accuracy: 0.7660\n",
            "Epoch 1798/2000\n",
            "3/3 [==============================] - 0s 46ms/step - loss: 0.4756 - accuracy: 0.8700 - val_loss: 0.5491 - val_accuracy: 0.7660\n",
            "Epoch 1799/2000\n",
            "3/3 [==============================] - 0s 66ms/step - loss: 0.3646 - accuracy: 0.8700 - val_loss: 0.5518 - val_accuracy: 0.7660\n",
            "Epoch 1800/2000\n",
            "3/3 [==============================] - 0s 55ms/step - loss: 0.4438 - accuracy: 0.8600 - val_loss: 0.5681 - val_accuracy: 0.7660\n",
            "Epoch 1801/2000\n",
            "3/3 [==============================] - 0s 51ms/step - loss: 0.3999 - accuracy: 0.8900 - val_loss: 0.5647 - val_accuracy: 0.7872\n",
            "Epoch 1802/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 0.4144 - accuracy: 0.9100 - val_loss: 0.5309 - val_accuracy: 0.8085\n",
            "Epoch 1803/2000\n",
            "3/3 [==============================] - 0s 56ms/step - loss: 0.4702 - accuracy: 0.8500 - val_loss: 0.5001 - val_accuracy: 0.8298\n",
            "Epoch 1804/2000\n",
            "3/3 [==============================] - 0s 47ms/step - loss: 0.5478 - accuracy: 0.8100 - val_loss: 0.4745 - val_accuracy: 0.8298\n",
            "Epoch 1805/2000\n",
            "3/3 [==============================] - 0s 59ms/step - loss: 0.4272 - accuracy: 0.8500 - val_loss: 0.4629 - val_accuracy: 0.8298\n",
            "Epoch 1806/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 0.4333 - accuracy: 0.8300 - val_loss: 0.4670 - val_accuracy: 0.8511\n",
            "Epoch 1807/2000\n",
            "3/3 [==============================] - 0s 46ms/step - loss: 0.4942 - accuracy: 0.8100 - val_loss: 0.4986 - val_accuracy: 0.8298\n",
            "Epoch 1808/2000\n",
            "3/3 [==============================] - 0s 52ms/step - loss: 0.5860 - accuracy: 0.8100 - val_loss: 0.5142 - val_accuracy: 0.8511\n",
            "Epoch 1809/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 0.3960 - accuracy: 0.8500 - val_loss: 0.5086 - val_accuracy: 0.8298\n",
            "Epoch 1810/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 0.4472 - accuracy: 0.7900 - val_loss: 0.4528 - val_accuracy: 0.8298\n",
            "Epoch 1811/2000\n",
            "3/3 [==============================] - 0s 44ms/step - loss: 0.3379 - accuracy: 0.9000 - val_loss: 0.4129 - val_accuracy: 0.8723\n",
            "Epoch 1812/2000\n",
            "3/3 [==============================] - 0s 55ms/step - loss: 0.6372 - accuracy: 0.8100 - val_loss: 0.3940 - val_accuracy: 0.8936\n",
            "Epoch 1813/2000\n",
            "3/3 [==============================] - 0s 53ms/step - loss: 0.4753 - accuracy: 0.8200 - val_loss: 0.3908 - val_accuracy: 0.8936\n",
            "Epoch 1814/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 0.5636 - accuracy: 0.8100 - val_loss: 0.4025 - val_accuracy: 0.8723\n",
            "Epoch 1815/2000\n",
            "3/3 [==============================] - 0s 52ms/step - loss: 0.3080 - accuracy: 0.9000 - val_loss: 0.4178 - val_accuracy: 0.8511\n",
            "Epoch 1816/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 0.4485 - accuracy: 0.8500 - val_loss: 0.4248 - val_accuracy: 0.8511\n",
            "Epoch 1817/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 0.4025 - accuracy: 0.9000 - val_loss: 0.4312 - val_accuracy: 0.8511\n",
            "Epoch 1818/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 0.4693 - accuracy: 0.8300 - val_loss: 0.4385 - val_accuracy: 0.8298\n",
            "Epoch 1819/2000\n",
            "3/3 [==============================] - 0s 58ms/step - loss: 0.3688 - accuracy: 0.8500 - val_loss: 0.4248 - val_accuracy: 0.8511\n",
            "Epoch 1820/2000\n",
            "3/3 [==============================] - 0s 53ms/step - loss: 0.4679 - accuracy: 0.8400 - val_loss: 0.4179 - val_accuracy: 0.8298\n",
            "Epoch 1821/2000\n",
            "3/3 [==============================] - 0s 53ms/step - loss: 0.6975 - accuracy: 0.8000 - val_loss: 0.4114 - val_accuracy: 0.8298\n",
            "Epoch 1822/2000\n",
            "3/3 [==============================] - 0s 53ms/step - loss: 0.4185 - accuracy: 0.8900 - val_loss: 0.4180 - val_accuracy: 0.8298\n",
            "Epoch 1823/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 0.4347 - accuracy: 0.8700 - val_loss: 0.4435 - val_accuracy: 0.8298\n",
            "Epoch 1824/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 0.4258 - accuracy: 0.8700 - val_loss: 0.4681 - val_accuracy: 0.8085\n",
            "Epoch 1825/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 0.4831 - accuracy: 0.8400 - val_loss: 0.4963 - val_accuracy: 0.8085\n",
            "Epoch 1826/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 0.4829 - accuracy: 0.8100 - val_loss: 0.4934 - val_accuracy: 0.8085\n",
            "Epoch 1827/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 0.4715 - accuracy: 0.8900 - val_loss: 0.4972 - val_accuracy: 0.8298\n",
            "Epoch 1828/2000\n",
            "3/3 [==============================] - 0s 65ms/step - loss: 0.3443 - accuracy: 0.9000 - val_loss: 0.5108 - val_accuracy: 0.8085\n",
            "Epoch 1829/2000\n",
            "3/3 [==============================] - 0s 62ms/step - loss: 0.4033 - accuracy: 0.8800 - val_loss: 0.4968 - val_accuracy: 0.8298\n",
            "Epoch 1830/2000\n",
            "3/3 [==============================] - 0s 63ms/step - loss: 0.3440 - accuracy: 0.8800 - val_loss: 0.4800 - val_accuracy: 0.8298\n",
            "Epoch 1831/2000\n",
            "3/3 [==============================] - 0s 71ms/step - loss: 0.3657 - accuracy: 0.8700 - val_loss: 0.4712 - val_accuracy: 0.8085\n",
            "Epoch 1832/2000\n",
            "3/3 [==============================] - 0s 74ms/step - loss: 0.4786 - accuracy: 0.8300 - val_loss: 0.4670 - val_accuracy: 0.8085\n",
            "Epoch 1833/2000\n",
            "3/3 [==============================] - 0s 67ms/step - loss: 0.3254 - accuracy: 0.9000 - val_loss: 0.4684 - val_accuracy: 0.8085\n",
            "Epoch 1834/2000\n",
            "3/3 [==============================] - 0s 60ms/step - loss: 0.5356 - accuracy: 0.8300 - val_loss: 0.4541 - val_accuracy: 0.8085\n",
            "Epoch 1835/2000\n",
            "3/3 [==============================] - 0s 46ms/step - loss: 0.6463 - accuracy: 0.8300 - val_loss: 0.4379 - val_accuracy: 0.8085\n",
            "Epoch 1836/2000\n",
            "3/3 [==============================] - 0s 73ms/step - loss: 0.4363 - accuracy: 0.8500 - val_loss: 0.4324 - val_accuracy: 0.8085\n",
            "Epoch 1837/2000\n",
            "3/3 [==============================] - 0s 60ms/step - loss: 0.4287 - accuracy: 0.8500 - val_loss: 0.4269 - val_accuracy: 0.8298\n",
            "Epoch 1838/2000\n",
            "3/3 [==============================] - 0s 67ms/step - loss: 0.4258 - accuracy: 0.8600 - val_loss: 0.4475 - val_accuracy: 0.8298\n",
            "Epoch 1839/2000\n",
            "3/3 [==============================] - 0s 61ms/step - loss: 0.3357 - accuracy: 0.9000 - val_loss: 0.4643 - val_accuracy: 0.7872\n",
            "Epoch 1840/2000\n",
            "3/3 [==============================] - 0s 67ms/step - loss: 0.4419 - accuracy: 0.8300 - val_loss: 0.4872 - val_accuracy: 0.7872\n",
            "Epoch 1841/2000\n",
            "3/3 [==============================] - 0s 76ms/step - loss: 0.4975 - accuracy: 0.8600 - val_loss: 0.5389 - val_accuracy: 0.7447\n",
            "Epoch 1842/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 0.4976 - accuracy: 0.8400 - val_loss: 0.5708 - val_accuracy: 0.7447\n",
            "Epoch 1843/2000\n",
            "3/3 [==============================] - 0s 45ms/step - loss: 0.6105 - accuracy: 0.8000 - val_loss: 0.5808 - val_accuracy: 0.7447\n",
            "Epoch 1844/2000\n",
            "3/3 [==============================] - 0s 33ms/step - loss: 0.4524 - accuracy: 0.8800 - val_loss: 0.5401 - val_accuracy: 0.7660\n",
            "Epoch 1845/2000\n",
            "3/3 [==============================] - 0s 45ms/step - loss: 0.5043 - accuracy: 0.8200 - val_loss: 0.4967 - val_accuracy: 0.7660\n",
            "Epoch 1846/2000\n",
            "3/3 [==============================] - 0s 44ms/step - loss: 0.4090 - accuracy: 0.8800 - val_loss: 0.4590 - val_accuracy: 0.7872\n",
            "Epoch 1847/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 0.4728 - accuracy: 0.8400 - val_loss: 0.4614 - val_accuracy: 0.8298\n",
            "Epoch 1848/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 0.6278 - accuracy: 0.7600 - val_loss: 0.4858 - val_accuracy: 0.8298\n",
            "Epoch 1849/2000\n",
            "3/3 [==============================] - 0s 44ms/step - loss: 0.5015 - accuracy: 0.8600 - val_loss: 0.5077 - val_accuracy: 0.8085\n",
            "Epoch 1850/2000\n",
            "3/3 [==============================] - 0s 48ms/step - loss: 0.4707 - accuracy: 0.8400 - val_loss: 0.5108 - val_accuracy: 0.8298\n",
            "Epoch 1851/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 0.6038 - accuracy: 0.7800 - val_loss: 0.4921 - val_accuracy: 0.8298\n",
            "Epoch 1852/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 0.3992 - accuracy: 0.8700 - val_loss: 0.4734 - val_accuracy: 0.8511\n",
            "Epoch 1853/2000\n",
            "3/3 [==============================] - 0s 57ms/step - loss: 0.4877 - accuracy: 0.7900 - val_loss: 0.4516 - val_accuracy: 0.8511\n",
            "Epoch 1854/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 0.5248 - accuracy: 0.7500 - val_loss: 0.4485 - val_accuracy: 0.8511\n",
            "Epoch 1855/2000\n",
            "3/3 [==============================] - 0s 49ms/step - loss: 0.6213 - accuracy: 0.8300 - val_loss: 0.4568 - val_accuracy: 0.8511\n",
            "Epoch 1856/2000\n",
            "3/3 [==============================] - 0s 45ms/step - loss: 0.3981 - accuracy: 0.8900 - val_loss: 0.4647 - val_accuracy: 0.8085\n",
            "Epoch 1857/2000\n",
            "3/3 [==============================] - 0s 44ms/step - loss: 0.4724 - accuracy: 0.8200 - val_loss: 0.4671 - val_accuracy: 0.8085\n",
            "Epoch 1858/2000\n",
            "3/3 [==============================] - 0s 61ms/step - loss: 0.4173 - accuracy: 0.8900 - val_loss: 0.4736 - val_accuracy: 0.7660\n",
            "Epoch 1859/2000\n",
            "3/3 [==============================] - 0s 47ms/step - loss: 0.5309 - accuracy: 0.8000 - val_loss: 0.4783 - val_accuracy: 0.8085\n",
            "Epoch 1860/2000\n",
            "3/3 [==============================] - 0s 48ms/step - loss: 0.4403 - accuracy: 0.8600 - val_loss: 0.4820 - val_accuracy: 0.8085\n",
            "Epoch 1861/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 0.4986 - accuracy: 0.8000 - val_loss: 0.4763 - val_accuracy: 0.8085\n",
            "Epoch 1862/2000\n",
            "3/3 [==============================] - 0s 51ms/step - loss: 0.5302 - accuracy: 0.8400 - val_loss: 0.5306 - val_accuracy: 0.8085\n",
            "Epoch 1863/2000\n",
            "3/3 [==============================] - 0s 50ms/step - loss: 0.5368 - accuracy: 0.8300 - val_loss: 0.5426 - val_accuracy: 0.8511\n",
            "Epoch 1864/2000\n",
            "3/3 [==============================] - 0s 56ms/step - loss: 0.4779 - accuracy: 0.8300 - val_loss: 0.5550 - val_accuracy: 0.8298\n",
            "Epoch 1865/2000\n",
            "3/3 [==============================] - 0s 49ms/step - loss: 0.4567 - accuracy: 0.8100 - val_loss: 0.5456 - val_accuracy: 0.8085\n",
            "Epoch 1866/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 0.4546 - accuracy: 0.8800 - val_loss: 0.5255 - val_accuracy: 0.7872\n",
            "Epoch 1867/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 0.4260 - accuracy: 0.8500 - val_loss: 0.5134 - val_accuracy: 0.7872\n",
            "Epoch 1868/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 0.5321 - accuracy: 0.8200 - val_loss: 0.5239 - val_accuracy: 0.7660\n",
            "Epoch 1869/2000\n",
            "3/3 [==============================] - 0s 38ms/step - loss: 0.3729 - accuracy: 0.8900 - val_loss: 0.5186 - val_accuracy: 0.7660\n",
            "Epoch 1870/2000\n",
            "3/3 [==============================] - 0s 51ms/step - loss: 0.4180 - accuracy: 0.8600 - val_loss: 0.5301 - val_accuracy: 0.7872\n",
            "Epoch 1871/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 0.3480 - accuracy: 0.8400 - val_loss: 0.5413 - val_accuracy: 0.7872\n",
            "Epoch 1872/2000\n",
            "3/3 [==============================] - 0s 56ms/step - loss: 0.4221 - accuracy: 0.8500 - val_loss: 0.5634 - val_accuracy: 0.8085\n",
            "Epoch 1873/2000\n",
            "3/3 [==============================] - 0s 49ms/step - loss: 0.5040 - accuracy: 0.8200 - val_loss: 0.5569 - val_accuracy: 0.8085\n",
            "Epoch 1874/2000\n",
            "3/3 [==============================] - 0s 57ms/step - loss: 0.4413 - accuracy: 0.8700 - val_loss: 0.5450 - val_accuracy: 0.8298\n",
            "Epoch 1875/2000\n",
            "3/3 [==============================] - 0s 59ms/step - loss: 0.4270 - accuracy: 0.8500 - val_loss: 0.5281 - val_accuracy: 0.8298\n",
            "Epoch 1876/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 0.4747 - accuracy: 0.8400 - val_loss: 0.5072 - val_accuracy: 0.8298\n",
            "Epoch 1877/2000\n",
            "3/3 [==============================] - 0s 40ms/step - loss: 0.3335 - accuracy: 0.9100 - val_loss: 0.4882 - val_accuracy: 0.8298\n",
            "Epoch 1878/2000\n",
            "3/3 [==============================] - 0s 47ms/step - loss: 0.5392 - accuracy: 0.8500 - val_loss: 0.4688 - val_accuracy: 0.8298\n",
            "Epoch 1879/2000\n",
            "3/3 [==============================] - 0s 44ms/step - loss: 0.5043 - accuracy: 0.8500 - val_loss: 0.4686 - val_accuracy: 0.8298\n",
            "Epoch 1880/2000\n",
            "3/3 [==============================] - 0s 56ms/step - loss: 0.3975 - accuracy: 0.8700 - val_loss: 0.4835 - val_accuracy: 0.8298\n",
            "Epoch 1881/2000\n",
            "3/3 [==============================] - 0s 49ms/step - loss: 0.4726 - accuracy: 0.8500 - val_loss: 0.4902 - val_accuracy: 0.8298\n",
            "Epoch 1882/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 0.4520 - accuracy: 0.8100 - val_loss: 0.4930 - val_accuracy: 0.8298\n",
            "Epoch 1883/2000\n",
            "3/3 [==============================] - 0s 54ms/step - loss: 0.4394 - accuracy: 0.9100 - val_loss: 0.4939 - val_accuracy: 0.8298\n",
            "Epoch 1884/2000\n",
            "3/3 [==============================] - 0s 55ms/step - loss: 0.4720 - accuracy: 0.8500 - val_loss: 0.4794 - val_accuracy: 0.8298\n",
            "Epoch 1885/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 0.4402 - accuracy: 0.8600 - val_loss: 0.4664 - val_accuracy: 0.8298\n",
            "Epoch 1886/2000\n",
            "3/3 [==============================] - 0s 42ms/step - loss: 0.5132 - accuracy: 0.8800 - val_loss: 0.4682 - val_accuracy: 0.8298\n",
            "Epoch 1887/2000\n",
            "3/3 [==============================] - 0s 59ms/step - loss: 0.4832 - accuracy: 0.9100 - val_loss: 0.4831 - val_accuracy: 0.7872\n",
            "Epoch 1888/2000\n",
            "3/3 [==============================] - 0s 49ms/step - loss: 0.4870 - accuracy: 0.8400 - val_loss: 0.4920 - val_accuracy: 0.8085\n",
            "Epoch 1889/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 0.3594 - accuracy: 0.8700 - val_loss: 0.5081 - val_accuracy: 0.8085\n",
            "Epoch 1890/2000\n",
            "3/3 [==============================] - 0s 38ms/step - loss: 0.5173 - accuracy: 0.8600 - val_loss: 0.5286 - val_accuracy: 0.8298\n",
            "Epoch 1891/2000\n",
            "3/3 [==============================] - 0s 47ms/step - loss: 0.5386 - accuracy: 0.8200 - val_loss: 0.5346 - val_accuracy: 0.8085\n",
            "Epoch 1892/2000\n",
            "3/3 [==============================] - 0s 46ms/step - loss: 0.5123 - accuracy: 0.8000 - val_loss: 0.5225 - val_accuracy: 0.8085\n",
            "Epoch 1893/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 0.4341 - accuracy: 0.8900 - val_loss: 0.5095 - val_accuracy: 0.8085\n",
            "Epoch 1894/2000\n",
            "3/3 [==============================] - 0s 34ms/step - loss: 0.4873 - accuracy: 0.8300 - val_loss: 0.5030 - val_accuracy: 0.8085\n",
            "Epoch 1895/2000\n",
            "3/3 [==============================] - 0s 74ms/step - loss: 0.4245 - accuracy: 0.8600 - val_loss: 0.5160 - val_accuracy: 0.8085\n",
            "Epoch 1896/2000\n",
            "3/3 [==============================] - 0s 78ms/step - loss: 0.3633 - accuracy: 0.8700 - val_loss: 0.5324 - val_accuracy: 0.7872\n",
            "Epoch 1897/2000\n",
            "3/3 [==============================] - 0s 73ms/step - loss: 0.4932 - accuracy: 0.8200 - val_loss: 0.5380 - val_accuracy: 0.8085\n",
            "Epoch 1898/2000\n",
            "3/3 [==============================] - 0s 61ms/step - loss: 0.6416 - accuracy: 0.8100 - val_loss: 0.5145 - val_accuracy: 0.7872\n",
            "Epoch 1899/2000\n",
            "3/3 [==============================] - 0s 66ms/step - loss: 0.3958 - accuracy: 0.8500 - val_loss: 0.5074 - val_accuracy: 0.8085\n",
            "Epoch 1900/2000\n",
            "3/3 [==============================] - 0s 80ms/step - loss: 0.5761 - accuracy: 0.8500 - val_loss: 0.5485 - val_accuracy: 0.7872\n",
            "Epoch 1901/2000\n",
            "3/3 [==============================] - 0s 61ms/step - loss: 0.3146 - accuracy: 0.9000 - val_loss: 0.5878 - val_accuracy: 0.8085\n",
            "Epoch 1902/2000\n",
            "3/3 [==============================] - 0s 48ms/step - loss: 0.4541 - accuracy: 0.8300 - val_loss: 0.5676 - val_accuracy: 0.8085\n",
            "Epoch 1903/2000\n",
            "3/3 [==============================] - 0s 59ms/step - loss: 0.4734 - accuracy: 0.8500 - val_loss: 0.5121 - val_accuracy: 0.8085\n",
            "Epoch 1904/2000\n",
            "3/3 [==============================] - 0s 75ms/step - loss: 0.6807 - accuracy: 0.8200 - val_loss: 0.4582 - val_accuracy: 0.8298\n",
            "Epoch 1905/2000\n",
            "3/3 [==============================] - 0s 63ms/step - loss: 0.4236 - accuracy: 0.8300 - val_loss: 0.4104 - val_accuracy: 0.8085\n",
            "Epoch 1906/2000\n",
            "3/3 [==============================] - 0s 89ms/step - loss: 0.4479 - accuracy: 0.8800 - val_loss: 0.3763 - val_accuracy: 0.7872\n",
            "Epoch 1907/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 0.4037 - accuracy: 0.8400 - val_loss: 0.3758 - val_accuracy: 0.8298\n",
            "Epoch 1908/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 0.5799 - accuracy: 0.8100 - val_loss: 0.3975 - val_accuracy: 0.8085\n",
            "Epoch 1909/2000\n",
            "3/3 [==============================] - 0s 56ms/step - loss: 0.3903 - accuracy: 0.8800 - val_loss: 0.4641 - val_accuracy: 0.8085\n",
            "Epoch 1910/2000\n",
            "3/3 [==============================] - 0s 44ms/step - loss: 0.5471 - accuracy: 0.8400 - val_loss: 0.5324 - val_accuracy: 0.7660\n",
            "Epoch 1911/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 0.6466 - accuracy: 0.7900 - val_loss: 0.5386 - val_accuracy: 0.8085\n",
            "Epoch 1912/2000\n",
            "3/3 [==============================] - 0s 52ms/step - loss: 0.5674 - accuracy: 0.8300 - val_loss: 0.5181 - val_accuracy: 0.8085\n",
            "Epoch 1913/2000\n",
            "3/3 [==============================] - 0s 57ms/step - loss: 0.4347 - accuracy: 0.8200 - val_loss: 0.4741 - val_accuracy: 0.8511\n",
            "Epoch 1914/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 0.4011 - accuracy: 0.9000 - val_loss: 0.4157 - val_accuracy: 0.8298\n",
            "Epoch 1915/2000\n",
            "3/3 [==============================] - 0s 53ms/step - loss: 0.3482 - accuracy: 0.8800 - val_loss: 0.3868 - val_accuracy: 0.8298\n",
            "Epoch 1916/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 0.4595 - accuracy: 0.8400 - val_loss: 0.3764 - val_accuracy: 0.8511\n",
            "Epoch 1917/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 0.4447 - accuracy: 0.8300 - val_loss: 0.3714 - val_accuracy: 0.8511\n",
            "Epoch 1918/2000\n",
            "3/3 [==============================] - 0s 61ms/step - loss: 0.4124 - accuracy: 0.8900 - val_loss: 0.4097 - val_accuracy: 0.8298\n",
            "Epoch 1919/2000\n",
            "3/3 [==============================] - 0s 52ms/step - loss: 0.5546 - accuracy: 0.8100 - val_loss: 0.4346 - val_accuracy: 0.8085\n",
            "Epoch 1920/2000\n",
            "3/3 [==============================] - 0s 57ms/step - loss: 0.5483 - accuracy: 0.8300 - val_loss: 0.4108 - val_accuracy: 0.8085\n",
            "Epoch 1921/2000\n",
            "3/3 [==============================] - 0s 47ms/step - loss: 0.4809 - accuracy: 0.8200 - val_loss: 0.4175 - val_accuracy: 0.8723\n",
            "Epoch 1922/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 0.4464 - accuracy: 0.8700 - val_loss: 0.4502 - val_accuracy: 0.8511\n",
            "Epoch 1923/2000\n",
            "3/3 [==============================] - 0s 52ms/step - loss: 0.6718 - accuracy: 0.8100 - val_loss: 0.4634 - val_accuracy: 0.8298\n",
            "Epoch 1924/2000\n",
            "3/3 [==============================] - 0s 56ms/step - loss: 0.4429 - accuracy: 0.8100 - val_loss: 0.4737 - val_accuracy: 0.8085\n",
            "Epoch 1925/2000\n",
            "3/3 [==============================] - 0s 52ms/step - loss: 0.4706 - accuracy: 0.8600 - val_loss: 0.4806 - val_accuracy: 0.8085\n",
            "Epoch 1926/2000\n",
            "3/3 [==============================] - 0s 50ms/step - loss: 0.4323 - accuracy: 0.8300 - val_loss: 0.4771 - val_accuracy: 0.7872\n",
            "Epoch 1927/2000\n",
            "3/3 [==============================] - 0s 44ms/step - loss: 0.4220 - accuracy: 0.8400 - val_loss: 0.4739 - val_accuracy: 0.8085\n",
            "Epoch 1928/2000\n",
            "3/3 [==============================] - 0s 39ms/step - loss: 0.4166 - accuracy: 0.9000 - val_loss: 0.4694 - val_accuracy: 0.8298\n",
            "Epoch 1929/2000\n",
            "3/3 [==============================] - 0s 56ms/step - loss: 0.4873 - accuracy: 0.8300 - val_loss: 0.4484 - val_accuracy: 0.8298\n",
            "Epoch 1930/2000\n",
            "3/3 [==============================] - 0s 45ms/step - loss: 0.4806 - accuracy: 0.8300 - val_loss: 0.4162 - val_accuracy: 0.8511\n",
            "Epoch 1931/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 0.6031 - accuracy: 0.8000 - val_loss: 0.4053 - val_accuracy: 0.8511\n",
            "Epoch 1932/2000\n",
            "3/3 [==============================] - 0s 57ms/step - loss: 0.4378 - accuracy: 0.8500 - val_loss: 0.4248 - val_accuracy: 0.8085\n",
            "Epoch 1933/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 0.3734 - accuracy: 0.8500 - val_loss: 0.4277 - val_accuracy: 0.8298\n",
            "Epoch 1934/2000\n",
            "3/3 [==============================] - 0s 57ms/step - loss: 0.5039 - accuracy: 0.8100 - val_loss: 0.4354 - val_accuracy: 0.8085\n",
            "Epoch 1935/2000\n",
            "3/3 [==============================] - 0s 54ms/step - loss: 0.4210 - accuracy: 0.8400 - val_loss: 0.4320 - val_accuracy: 0.8085\n",
            "Epoch 1936/2000\n",
            "3/3 [==============================] - 0s 54ms/step - loss: 0.5583 - accuracy: 0.7800 - val_loss: 0.4231 - val_accuracy: 0.8298\n",
            "Epoch 1937/2000\n",
            "3/3 [==============================] - 0s 38ms/step - loss: 0.3855 - accuracy: 0.9000 - val_loss: 0.4111 - val_accuracy: 0.8085\n",
            "Epoch 1938/2000\n",
            "3/3 [==============================] - 0s 46ms/step - loss: 0.4937 - accuracy: 0.8300 - val_loss: 0.4205 - val_accuracy: 0.8085\n",
            "Epoch 1939/2000\n",
            "3/3 [==============================] - 0s 45ms/step - loss: 0.4773 - accuracy: 0.8500 - val_loss: 0.4309 - val_accuracy: 0.8085\n",
            "Epoch 1940/2000\n",
            "3/3 [==============================] - 0s 53ms/step - loss: 0.4832 - accuracy: 0.8300 - val_loss: 0.4210 - val_accuracy: 0.8298\n",
            "Epoch 1941/2000\n",
            "3/3 [==============================] - 0s 65ms/step - loss: 0.4087 - accuracy: 0.8700 - val_loss: 0.4202 - val_accuracy: 0.8511\n",
            "Epoch 1942/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 0.4080 - accuracy: 0.8500 - val_loss: 0.4196 - val_accuracy: 0.8511\n",
            "Epoch 1943/2000\n",
            "3/3 [==============================] - 0s 45ms/step - loss: 0.5428 - accuracy: 0.7900 - val_loss: 0.4108 - val_accuracy: 0.8723\n",
            "Epoch 1944/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 0.6526 - accuracy: 0.8000 - val_loss: 0.4081 - val_accuracy: 0.8723\n",
            "Epoch 1945/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 0.3245 - accuracy: 0.9000 - val_loss: 0.4107 - val_accuracy: 0.8298\n",
            "Epoch 1946/2000\n",
            "3/3 [==============================] - 0s 45ms/step - loss: 0.4872 - accuracy: 0.7800 - val_loss: 0.4148 - val_accuracy: 0.8723\n",
            "Epoch 1947/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 0.3139 - accuracy: 0.9000 - val_loss: 0.4199 - val_accuracy: 0.8511\n",
            "Epoch 1948/2000\n",
            "3/3 [==============================] - 0s 47ms/step - loss: 0.6008 - accuracy: 0.8300 - val_loss: 0.4278 - val_accuracy: 0.8085\n",
            "Epoch 1949/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 0.4182 - accuracy: 0.8800 - val_loss: 0.4217 - val_accuracy: 0.8298\n",
            "Epoch 1950/2000\n",
            "3/3 [==============================] - 0s 62ms/step - loss: 0.5460 - accuracy: 0.8800 - val_loss: 0.4146 - val_accuracy: 0.8298\n",
            "Epoch 1951/2000\n",
            "3/3 [==============================] - 0s 46ms/step - loss: 0.3817 - accuracy: 0.8900 - val_loss: 0.4093 - val_accuracy: 0.8298\n",
            "Epoch 1952/2000\n",
            "3/3 [==============================] - 0s 45ms/step - loss: 0.5500 - accuracy: 0.8100 - val_loss: 0.4061 - val_accuracy: 0.8298\n",
            "Epoch 1953/2000\n",
            "3/3 [==============================] - 0s 53ms/step - loss: 0.4772 - accuracy: 0.8300 - val_loss: 0.4045 - val_accuracy: 0.8298\n",
            "Epoch 1954/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 0.4209 - accuracy: 0.8700 - val_loss: 0.3961 - val_accuracy: 0.8511\n",
            "Epoch 1955/2000\n",
            "3/3 [==============================] - 0s 60ms/step - loss: 0.5658 - accuracy: 0.8000 - val_loss: 0.4007 - val_accuracy: 0.8298\n",
            "Epoch 1956/2000\n",
            "3/3 [==============================] - 0s 59ms/step - loss: 0.4392 - accuracy: 0.8700 - val_loss: 0.4156 - val_accuracy: 0.8511\n",
            "Epoch 1957/2000\n",
            "3/3 [==============================] - 0s 46ms/step - loss: 0.5878 - accuracy: 0.8000 - val_loss: 0.4221 - val_accuracy: 0.8298\n",
            "Epoch 1958/2000\n",
            "3/3 [==============================] - 0s 62ms/step - loss: 0.5231 - accuracy: 0.8200 - val_loss: 0.4146 - val_accuracy: 0.8511\n",
            "Epoch 1959/2000\n",
            "3/3 [==============================] - 0s 64ms/step - loss: 0.4515 - accuracy: 0.8800 - val_loss: 0.4114 - val_accuracy: 0.8723\n",
            "Epoch 1960/2000\n",
            "3/3 [==============================] - 0s 69ms/step - loss: 0.5113 - accuracy: 0.8600 - val_loss: 0.4196 - val_accuracy: 0.8723\n",
            "Epoch 1961/2000\n",
            "3/3 [==============================] - 0s 63ms/step - loss: 0.4379 - accuracy: 0.8600 - val_loss: 0.4156 - val_accuracy: 0.8723\n",
            "Epoch 1962/2000\n",
            "3/3 [==============================] - 0s 57ms/step - loss: 0.4379 - accuracy: 0.8300 - val_loss: 0.4186 - val_accuracy: 0.8511\n",
            "Epoch 1963/2000\n",
            "3/3 [==============================] - 0s 49ms/step - loss: 0.5147 - accuracy: 0.8400 - val_loss: 0.4309 - val_accuracy: 0.8298\n",
            "Epoch 1964/2000\n",
            "3/3 [==============================] - 0s 66ms/step - loss: 0.5254 - accuracy: 0.8300 - val_loss: 0.4446 - val_accuracy: 0.8511\n",
            "Epoch 1965/2000\n",
            "3/3 [==============================] - 0s 70ms/step - loss: 0.5293 - accuracy: 0.8500 - val_loss: 0.4437 - val_accuracy: 0.8298\n",
            "Epoch 1966/2000\n",
            "3/3 [==============================] - 0s 74ms/step - loss: 0.3633 - accuracy: 0.9300 - val_loss: 0.4495 - val_accuracy: 0.7872\n",
            "Epoch 1967/2000\n",
            "3/3 [==============================] - 0s 85ms/step - loss: 0.4563 - accuracy: 0.8700 - val_loss: 0.4513 - val_accuracy: 0.8085\n",
            "Epoch 1968/2000\n",
            "3/3 [==============================] - 0s 73ms/step - loss: 0.3478 - accuracy: 0.9000 - val_loss: 0.4457 - val_accuracy: 0.8085\n",
            "Epoch 1969/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 0.5243 - accuracy: 0.8400 - val_loss: 0.4444 - val_accuracy: 0.7872\n",
            "Epoch 1970/2000\n",
            "3/3 [==============================] - 0s 53ms/step - loss: 0.3433 - accuracy: 0.8800 - val_loss: 0.4280 - val_accuracy: 0.8298\n",
            "Epoch 1971/2000\n",
            "3/3 [==============================] - 0s 49ms/step - loss: 0.4326 - accuracy: 0.8700 - val_loss: 0.4263 - val_accuracy: 0.8298\n",
            "Epoch 1972/2000\n",
            "3/3 [==============================] - 0s 46ms/step - loss: 0.4609 - accuracy: 0.8300 - val_loss: 0.4416 - val_accuracy: 0.7872\n",
            "Epoch 1973/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 0.3584 - accuracy: 0.8900 - val_loss: 0.4477 - val_accuracy: 0.8085\n",
            "Epoch 1974/2000\n",
            "3/3 [==============================] - 0s 44ms/step - loss: 0.4128 - accuracy: 0.8300 - val_loss: 0.4267 - val_accuracy: 0.8085\n",
            "Epoch 1975/2000\n",
            "3/3 [==============================] - 0s 63ms/step - loss: 0.4802 - accuracy: 0.8400 - val_loss: 0.4294 - val_accuracy: 0.8511\n",
            "Epoch 1976/2000\n",
            "3/3 [==============================] - 0s 38ms/step - loss: 0.4103 - accuracy: 0.8500 - val_loss: 0.4424 - val_accuracy: 0.8511\n",
            "Epoch 1977/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 0.4230 - accuracy: 0.9000 - val_loss: 0.4533 - val_accuracy: 0.8298\n",
            "Epoch 1978/2000\n",
            "3/3 [==============================] - 0s 58ms/step - loss: 0.4369 - accuracy: 0.8500 - val_loss: 0.4662 - val_accuracy: 0.8298\n",
            "Epoch 1979/2000\n",
            "3/3 [==============================] - 0s 51ms/step - loss: 0.3379 - accuracy: 0.8900 - val_loss: 0.4761 - val_accuracy: 0.8511\n",
            "Epoch 1980/2000\n",
            "3/3 [==============================] - 0s 36ms/step - loss: 0.4265 - accuracy: 0.8300 - val_loss: 0.4596 - val_accuracy: 0.8511\n",
            "Epoch 1981/2000\n",
            "3/3 [==============================] - 0s 45ms/step - loss: 0.5459 - accuracy: 0.8200 - val_loss: 0.4261 - val_accuracy: 0.8298\n",
            "Epoch 1982/2000\n",
            "3/3 [==============================] - 0s 44ms/step - loss: 0.3351 - accuracy: 0.9400 - val_loss: 0.4440 - val_accuracy: 0.7872\n",
            "Epoch 1983/2000\n",
            "3/3 [==============================] - 0s 44ms/step - loss: 0.3900 - accuracy: 0.8900 - val_loss: 0.4926 - val_accuracy: 0.7872\n",
            "Epoch 1984/2000\n",
            "3/3 [==============================] - 0s 46ms/step - loss: 0.7086 - accuracy: 0.8000 - val_loss: 0.5125 - val_accuracy: 0.7660\n",
            "Epoch 1985/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 0.5735 - accuracy: 0.8100 - val_loss: 0.4928 - val_accuracy: 0.7872\n",
            "Epoch 1986/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 0.6596 - accuracy: 0.7900 - val_loss: 0.4652 - val_accuracy: 0.8298\n",
            "Epoch 1987/2000\n",
            "3/3 [==============================] - 0s 44ms/step - loss: 0.3947 - accuracy: 0.8900 - val_loss: 0.4440 - val_accuracy: 0.8298\n",
            "Epoch 1988/2000\n",
            "3/3 [==============================] - 0s 44ms/step - loss: 0.4035 - accuracy: 0.8600 - val_loss: 0.4310 - val_accuracy: 0.8511\n",
            "Epoch 1989/2000\n",
            "3/3 [==============================] - 0s 35ms/step - loss: 0.6939 - accuracy: 0.7700 - val_loss: 0.4164 - val_accuracy: 0.8723\n",
            "Epoch 1990/2000\n",
            "3/3 [==============================] - 0s 43ms/step - loss: 0.5722 - accuracy: 0.8600 - val_loss: 0.4569 - val_accuracy: 0.8723\n",
            "Epoch 1991/2000\n",
            "3/3 [==============================] - 0s 55ms/step - loss: 0.4862 - accuracy: 0.8300 - val_loss: 0.4869 - val_accuracy: 0.8936\n",
            "Epoch 1992/2000\n",
            "3/3 [==============================] - 0s 41ms/step - loss: 0.4838 - accuracy: 0.8600 - val_loss: 0.4624 - val_accuracy: 0.9149\n",
            "Epoch 1993/2000\n",
            "3/3 [==============================] - 0s 33ms/step - loss: 0.5114 - accuracy: 0.8100 - val_loss: 0.4244 - val_accuracy: 0.8723\n",
            "Epoch 1994/2000\n",
            "3/3 [==============================] - 0s 49ms/step - loss: 0.6362 - accuracy: 0.7500 - val_loss: 0.3982 - val_accuracy: 0.8511\n",
            "Epoch 1995/2000\n",
            "3/3 [==============================] - 0s 46ms/step - loss: 0.5122 - accuracy: 0.7700 - val_loss: 0.4046 - val_accuracy: 0.8298\n",
            "Epoch 1996/2000\n",
            "3/3 [==============================] - 0s 55ms/step - loss: 0.5153 - accuracy: 0.8700 - val_loss: 0.4011 - val_accuracy: 0.8298\n",
            "Epoch 1997/2000\n",
            "3/3 [==============================] - 0s 49ms/step - loss: 0.5005 - accuracy: 0.8100 - val_loss: 0.3844 - val_accuracy: 0.8085\n",
            "Epoch 1998/2000\n",
            "3/3 [==============================] - 0s 49ms/step - loss: 0.3897 - accuracy: 0.8900 - val_loss: 0.3721 - val_accuracy: 0.8511\n",
            "Epoch 1999/2000\n",
            "3/3 [==============================] - 0s 37ms/step - loss: 0.4910 - accuracy: 0.8100 - val_loss: 0.3789 - val_accuracy: 0.8511\n",
            "Epoch 2000/2000\n",
            "3/3 [==============================] - 0s 47ms/step - loss: 0.4693 - accuracy: 0.8200 - val_loss: 0.3912 - val_accuracy: 0.8511\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "drawAndGuess(model)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 378
        },
        "id": "MSXhRfBAML3-",
        "outputId": "80be7fdc-be50-4a5a-af8a-0cf98219e44e"
      },
      "execution_count": 89,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Draw a score and then click on finish to try to recognize it.\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              "\n",
              "<canvas width=280 height=280 style=\"border:1px solid #000000;\"></canvas><br/>\n",
              "<button>Finish</button>\n",
              "<script>\n",
              "var canvas = document.querySelector('canvas')\n",
              "var ctx = canvas.getContext('2d')\n",
              "ctx.lineWidth = 20\n",
              "ctx.lineCap = 'round';\n",
              "var button = document.querySelector('button')\n",
              "var mouse = {x: 0, y: 0}\n",
              "canvas.addEventListener('mousemove', function(e) {\n",
              "  mouse.x = e.pageX - this.offsetLeft\n",
              "  mouse.y = e.pageY - this.offsetTop\n",
              "})\n",
              "canvas.onmousedown = ()=>{\n",
              "  ctx.beginPath()\n",
              "  ctx.moveTo(mouse.x, mouse.y)\n",
              "  canvas.addEventListener('mousemove', onPaint)\n",
              "}\n",
              "canvas.onmouseup = ()=>{\n",
              "  canvas.removeEventListener('mousemove', onPaint)\n",
              "}\n",
              "var onPaint = ()=>{\n",
              "  ctx.lineTo(mouse.x, mouse.y)\n",
              "  ctx.stroke()\n",
              "}\n",
              "var data = new Promise(resolve=>{\n",
              "  button.onclick = ()=>{\n",
              "    resolve(canvas.toDataURL('image/png'))\n",
              "  }\n",
              "})\n",
              "</script>\n"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "1/1 [==============================] - 0s 22ms/step\n",
            "This is a D- with confidence 64.0%.\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "print(indexes_bad_prediction(X_test, Y_test))"
      ],
      "metadata": {
        "id": "OFOLCf7uxKyd"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "# **Augmentation 2 (Changed batch Size and used Adam) New Best //Improved on base**"
      ],
      "metadata": {
        "id": "4SLUBBAvce6B"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Take only the first 1740 samples of the training dataset\n",
        "X_train_small = X_train[0:1740, :, :].reshape((-1, 28, 28, 1))\n",
        "Y_train_small = Y_train[0:1740, :]\n",
        "\n",
        "from tensorflow.keras.optimizers import Adam\n",
        "\n",
        "# Using a smaller batch size for potentially more robust gradient estimates\n",
        "new_batch_size = 16\n",
        "\n",
        "# Adjusting the ImageDataGenerator parameters slightly\n",
        "generator = ImageDataGenerator(\n",
        "    rotation_range=15,  # Reduced rotation range\n",
        "    width_shift_range=0.1,  # Reduced shift range\n",
        "    height_shift_range=0.1,  # Reduced shift range\n",
        "    zoom_range=0.1,  # Keeping zoom range same\n",
        "    shear_range=5  # Reduced shear range\n",
        ")\n",
        "\n",
        "# Define the model with a slightly deeper architecture\n",
        "model1 = Sequential()\n",
        "model1.add(Flatten(input_shape=(28, 28)))\n",
        "model1.add(Dense(1024, activation='relu'))  # Increased from 32 to 64\n",
        "model1.add(Dense(512, activation='relu'))  # Increased from 16 to 32\n",
        "model1.add(Dense(256, activation='relu'))  # Additional layer\n",
        "model1.add(Dense(128, activation='relu'))  # Additional layer\n",
        "model1.add(Dense(64, activation='relu'))  # Additional layer\n",
        "model1.add(Dense(15, activation='softmax'))\n",
        "\n",
        "# Adjusting the optimizer's learning rate\n",
        "adam_optimizer = Adam(learning_rate=0.001)  # Adjusted learning rate\n",
        "\n",
        "model1.compile(loss='categorical_crossentropy',\n",
        "              optimizer=adam_optimizer,\n",
        "              metrics=['accuracy'])\n",
        "\n",
        "# Training the model with the adjusted batch size\n",
        "history = model1.fit(\n",
        "    generator.flow(X_train_small, Y_train_small, batch_size=new_batch_size),\n",
        "    steps_per_epoch=len(X_train_small) / new_batch_size,  # Adjusted to match the new batch size\n",
        "    epochs=100,  # Reduced epochs for quicker evaluation\n",
        "    validation_data=(X_test, Y_test)\n",
        ")\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "VuH7Rgo3ceYa",
        "outputId": "3e63797c-ead5-4404-e8e2-4233ce5f896a"
      },
      "execution_count": 136,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/100\n",
            "93/93 [==============================] - 6s 50ms/step - loss: 2.4144 - accuracy: 0.1873 - val_loss: 1.7086 - val_accuracy: 0.4424\n",
            "Epoch 2/100\n",
            "93/93 [==============================] - 3s 31ms/step - loss: 1.8247 - accuracy: 0.3833 - val_loss: 1.3073 - val_accuracy: 0.5511\n",
            "Epoch 3/100\n",
            "93/93 [==============================] - 2s 25ms/step - loss: 1.4346 - accuracy: 0.5127 - val_loss: 1.1847 - val_accuracy: 0.6197\n",
            "Epoch 4/100\n",
            "93/93 [==============================] - 3s 28ms/step - loss: 1.1692 - accuracy: 0.6160 - val_loss: 0.9096 - val_accuracy: 0.7102\n",
            "Epoch 5/100\n",
            "93/93 [==============================] - 3s 35ms/step - loss: 1.0737 - accuracy: 0.6340 - val_loss: 0.8961 - val_accuracy: 0.6960\n",
            "Epoch 6/100\n",
            "93/93 [==============================] - 4s 40ms/step - loss: 0.9572 - accuracy: 0.6927 - val_loss: 0.9566 - val_accuracy: 0.7076\n",
            "Epoch 7/100\n",
            "93/93 [==============================] - 4s 40ms/step - loss: 0.8748 - accuracy: 0.7033 - val_loss: 0.8468 - val_accuracy: 0.7115\n",
            "Epoch 8/100\n",
            "93/93 [==============================] - 2s 25ms/step - loss: 0.8682 - accuracy: 0.7107 - val_loss: 0.6179 - val_accuracy: 0.7853\n",
            "Epoch 9/100\n",
            "93/93 [==============================] - 3s 28ms/step - loss: 0.7822 - accuracy: 0.7360 - val_loss: 0.6174 - val_accuracy: 0.8060\n",
            "Epoch 10/100\n",
            "93/93 [==============================] - 3s 36ms/step - loss: 0.6913 - accuracy: 0.7720 - val_loss: 0.7266 - val_accuracy: 0.7865\n",
            "Epoch 11/100\n",
            "93/93 [==============================] - 2s 26ms/step - loss: 0.6672 - accuracy: 0.7867 - val_loss: 0.6452 - val_accuracy: 0.7982\n",
            "Epoch 12/100\n",
            "93/93 [==============================] - 2s 25ms/step - loss: 0.6590 - accuracy: 0.7780 - val_loss: 0.4850 - val_accuracy: 0.8215\n",
            "Epoch 13/100\n",
            "93/93 [==============================] - 2s 26ms/step - loss: 0.6077 - accuracy: 0.8007 - val_loss: 0.4552 - val_accuracy: 0.8629\n",
            "Epoch 14/100\n",
            "93/93 [==============================] - 3s 36ms/step - loss: 0.5333 - accuracy: 0.8220 - val_loss: 0.4565 - val_accuracy: 0.8564\n",
            "Epoch 15/100\n",
            "93/93 [==============================] - 3s 26ms/step - loss: 0.6022 - accuracy: 0.8133 - val_loss: 0.5215 - val_accuracy: 0.8370\n",
            "Epoch 16/100\n",
            "93/93 [==============================] - 2s 25ms/step - loss: 0.5113 - accuracy: 0.8367 - val_loss: 0.5290 - val_accuracy: 0.8409\n",
            "Epoch 17/100\n",
            "93/93 [==============================] - 2s 24ms/step - loss: 0.5154 - accuracy: 0.8373 - val_loss: 0.6224 - val_accuracy: 0.8228\n",
            "Epoch 18/100\n",
            "93/93 [==============================] - 3s 28ms/step - loss: 0.4639 - accuracy: 0.8527 - val_loss: 0.7512 - val_accuracy: 0.7930\n",
            "Epoch 19/100\n",
            "93/93 [==============================] - 4s 38ms/step - loss: 0.4717 - accuracy: 0.8420 - val_loss: 0.4597 - val_accuracy: 0.8784\n",
            "Epoch 20/100\n",
            "93/93 [==============================] - 2s 24ms/step - loss: 0.4709 - accuracy: 0.8467 - val_loss: 0.7401 - val_accuracy: 0.8047\n",
            "Epoch 21/100\n",
            "93/93 [==============================] - 2s 25ms/step - loss: 0.4522 - accuracy: 0.8513 - val_loss: 0.3925 - val_accuracy: 0.8823\n",
            "Epoch 22/100\n",
            "93/93 [==============================] - 2s 25ms/step - loss: 0.4006 - accuracy: 0.8687 - val_loss: 0.4202 - val_accuracy: 0.8693\n",
            "Epoch 23/100\n",
            "93/93 [==============================] - 3s 30ms/step - loss: 0.4668 - accuracy: 0.8480 - val_loss: 0.5413 - val_accuracy: 0.8512\n",
            "Epoch 24/100\n",
            "93/93 [==============================] - 2s 24ms/step - loss: 0.4088 - accuracy: 0.8713 - val_loss: 0.4359 - val_accuracy: 0.8771\n",
            "Epoch 25/100\n",
            "93/93 [==============================] - 3s 28ms/step - loss: 0.4068 - accuracy: 0.8687 - val_loss: 0.3646 - val_accuracy: 0.8952\n",
            "Epoch 26/100\n",
            "93/93 [==============================] - 3s 30ms/step - loss: 0.3799 - accuracy: 0.8753 - val_loss: 0.3707 - val_accuracy: 0.8862\n",
            "Epoch 27/100\n",
            "93/93 [==============================] - 3s 32ms/step - loss: 0.3851 - accuracy: 0.8713 - val_loss: 0.4560 - val_accuracy: 0.8706\n",
            "Epoch 28/100\n",
            "93/93 [==============================] - 2s 25ms/step - loss: 0.3638 - accuracy: 0.8780 - val_loss: 0.5061 - val_accuracy: 0.8771\n",
            "Epoch 29/100\n",
            "93/93 [==============================] - 2s 25ms/step - loss: 0.3285 - accuracy: 0.8920 - val_loss: 0.6428 - val_accuracy: 0.8370\n",
            "Epoch 30/100\n",
            "93/93 [==============================] - 2s 25ms/step - loss: 0.3048 - accuracy: 0.8980 - val_loss: 0.4073 - val_accuracy: 0.8991\n",
            "Epoch 31/100\n",
            "93/93 [==============================] - 3s 36ms/step - loss: 0.3514 - accuracy: 0.8793 - val_loss: 0.4307 - val_accuracy: 0.8797\n",
            "Epoch 32/100\n",
            "93/93 [==============================] - 4s 41ms/step - loss: 0.3824 - accuracy: 0.8760 - val_loss: 0.3774 - val_accuracy: 0.8926\n",
            "Epoch 33/100\n",
            "93/93 [==============================] - 2s 25ms/step - loss: 0.3483 - accuracy: 0.8953 - val_loss: 0.3880 - val_accuracy: 0.8862\n",
            "Epoch 34/100\n",
            "93/93 [==============================] - 2s 24ms/step - loss: 0.2901 - accuracy: 0.9053 - val_loss: 0.3481 - val_accuracy: 0.9094\n",
            "Epoch 35/100\n",
            "93/93 [==============================] - 3s 30ms/step - loss: 0.3075 - accuracy: 0.9013 - val_loss: 0.4706 - val_accuracy: 0.8810\n",
            "Epoch 36/100\n",
            "93/93 [==============================] - 3s 28ms/step - loss: 0.3631 - accuracy: 0.8900 - val_loss: 0.3749 - val_accuracy: 0.9017\n",
            "Epoch 37/100\n",
            "93/93 [==============================] - 3s 30ms/step - loss: 0.3321 - accuracy: 0.8880 - val_loss: 0.3104 - val_accuracy: 0.9198\n",
            "Epoch 38/100\n",
            "93/93 [==============================] - 2s 25ms/step - loss: 0.2714 - accuracy: 0.9133 - val_loss: 0.3058 - val_accuracy: 0.9185\n",
            "Epoch 39/100\n",
            "93/93 [==============================] - 2s 26ms/step - loss: 0.3022 - accuracy: 0.9100 - val_loss: 0.2684 - val_accuracy: 0.9185\n",
            "Epoch 40/100\n",
            "93/93 [==============================] - 3s 30ms/step - loss: 0.2663 - accuracy: 0.9140 - val_loss: 0.3832 - val_accuracy: 0.9004\n",
            "Epoch 41/100\n",
            "93/93 [==============================] - 3s 36ms/step - loss: 0.3012 - accuracy: 0.9013 - val_loss: 0.5258 - val_accuracy: 0.8849\n",
            "Epoch 42/100\n",
            "93/93 [==============================] - 3s 31ms/step - loss: 0.2818 - accuracy: 0.9160 - val_loss: 0.4613 - val_accuracy: 0.8758\n",
            "Epoch 43/100\n",
            "93/93 [==============================] - 3s 29ms/step - loss: 0.2785 - accuracy: 0.9127 - val_loss: 0.3353 - val_accuracy: 0.9107\n",
            "Epoch 44/100\n",
            "93/93 [==============================] - 3s 35ms/step - loss: 0.2547 - accuracy: 0.9140 - val_loss: 0.5020 - val_accuracy: 0.8668\n",
            "Epoch 45/100\n",
            "93/93 [==============================] - 3s 28ms/step - loss: 0.2572 - accuracy: 0.9120 - val_loss: 0.3668 - val_accuracy: 0.9043\n",
            "Epoch 46/100\n",
            "93/93 [==============================] - 3s 36ms/step - loss: 0.2744 - accuracy: 0.9140 - val_loss: 0.4241 - val_accuracy: 0.8939\n",
            "Epoch 47/100\n",
            "93/93 [==============================] - 2s 25ms/step - loss: 0.2481 - accuracy: 0.9160 - val_loss: 0.2887 - val_accuracy: 0.9198\n",
            "Epoch 48/100\n",
            "93/93 [==============================] - 2s 24ms/step - loss: 0.3076 - accuracy: 0.9020 - val_loss: 0.3491 - val_accuracy: 0.8900\n",
            "Epoch 49/100\n",
            "93/93 [==============================] - 2s 24ms/step - loss: 0.2441 - accuracy: 0.9233 - val_loss: 0.3366 - val_accuracy: 0.9198\n",
            "Epoch 50/100\n",
            "93/93 [==============================] - 3s 32ms/step - loss: 0.2880 - accuracy: 0.9060 - val_loss: 0.3518 - val_accuracy: 0.9030\n",
            "Epoch 51/100\n",
            "93/93 [==============================] - 3s 27ms/step - loss: 0.2599 - accuracy: 0.9207 - val_loss: 0.2857 - val_accuracy: 0.9159\n",
            "Epoch 52/100\n",
            "93/93 [==============================] - 2s 25ms/step - loss: 0.2657 - accuracy: 0.9040 - val_loss: 0.3998 - val_accuracy: 0.8926\n",
            "Epoch 53/100\n",
            "93/93 [==============================] - 2s 25ms/step - loss: 0.2225 - accuracy: 0.9280 - val_loss: 0.3724 - val_accuracy: 0.8965\n",
            "Epoch 54/100\n",
            "93/93 [==============================] - 3s 34ms/step - loss: 0.2339 - accuracy: 0.9287 - val_loss: 0.3100 - val_accuracy: 0.9237\n",
            "Epoch 55/100\n",
            "93/93 [==============================] - 2s 25ms/step - loss: 0.2450 - accuracy: 0.9247 - val_loss: 0.3738 - val_accuracy: 0.9107\n",
            "Epoch 56/100\n",
            "93/93 [==============================] - 2s 25ms/step - loss: 0.2521 - accuracy: 0.9220 - val_loss: 0.2534 - val_accuracy: 0.9340\n",
            "Epoch 57/100\n",
            "93/93 [==============================] - 2s 25ms/step - loss: 0.2393 - accuracy: 0.9267 - val_loss: 0.2918 - val_accuracy: 0.9198\n",
            "Epoch 58/100\n",
            "93/93 [==============================] - 3s 33ms/step - loss: 0.2185 - accuracy: 0.9260 - val_loss: 0.3790 - val_accuracy: 0.9198\n",
            "Epoch 59/100\n",
            "93/93 [==============================] - 3s 27ms/step - loss: 0.2057 - accuracy: 0.9313 - val_loss: 0.3825 - val_accuracy: 0.9082\n",
            "Epoch 60/100\n",
            "93/93 [==============================] - 2s 25ms/step - loss: 0.2031 - accuracy: 0.9367 - val_loss: 0.3543 - val_accuracy: 0.9172\n",
            "Epoch 61/100\n",
            "93/93 [==============================] - 3s 34ms/step - loss: 0.1873 - accuracy: 0.9440 - val_loss: 0.3460 - val_accuracy: 0.9198\n",
            "Epoch 62/100\n",
            "93/93 [==============================] - 2s 25ms/step - loss: 0.2216 - accuracy: 0.9267 - val_loss: 0.4441 - val_accuracy: 0.9120\n",
            "Epoch 63/100\n",
            "93/93 [==============================] - 2s 26ms/step - loss: 0.2201 - accuracy: 0.9347 - val_loss: 0.3349 - val_accuracy: 0.9224\n",
            "Epoch 64/100\n",
            "93/93 [==============================] - 2s 25ms/step - loss: 0.2500 - accuracy: 0.9233 - val_loss: 0.2972 - val_accuracy: 0.9340\n",
            "Epoch 65/100\n",
            "93/93 [==============================] - 3s 36ms/step - loss: 0.1976 - accuracy: 0.9353 - val_loss: 0.3040 - val_accuracy: 0.9250\n",
            "Epoch 66/100\n",
            "93/93 [==============================] - 2s 25ms/step - loss: 0.1841 - accuracy: 0.9420 - val_loss: 0.3528 - val_accuracy: 0.9133\n",
            "Epoch 67/100\n",
            "93/93 [==============================] - 2s 26ms/step - loss: 0.2143 - accuracy: 0.9313 - val_loss: 0.3133 - val_accuracy: 0.9224\n",
            "Epoch 68/100\n",
            "93/93 [==============================] - 3s 35ms/step - loss: 0.2019 - accuracy: 0.9393 - val_loss: 0.4129 - val_accuracy: 0.9146\n",
            "Epoch 69/100\n",
            "93/93 [==============================] - 3s 29ms/step - loss: 0.2146 - accuracy: 0.9293 - val_loss: 0.3146 - val_accuracy: 0.9250\n",
            "Epoch 70/100\n",
            "93/93 [==============================] - 2s 25ms/step - loss: 0.2270 - accuracy: 0.9320 - val_loss: 0.2908 - val_accuracy: 0.9185\n",
            "Epoch 71/100\n",
            "93/93 [==============================] - 3s 28ms/step - loss: 0.1846 - accuracy: 0.9460 - val_loss: 0.3323 - val_accuracy: 0.9185\n",
            "Epoch 72/100\n",
            "93/93 [==============================] - 4s 38ms/step - loss: 0.2189 - accuracy: 0.9287 - val_loss: 0.2999 - val_accuracy: 0.9314\n",
            "Epoch 73/100\n",
            "93/93 [==============================] - 3s 28ms/step - loss: 0.1969 - accuracy: 0.9333 - val_loss: 0.4048 - val_accuracy: 0.9056\n",
            "Epoch 74/100\n",
            "93/93 [==============================] - 3s 28ms/step - loss: 0.2306 - accuracy: 0.9253 - val_loss: 0.2987 - val_accuracy: 0.9250\n",
            "Epoch 75/100\n",
            "93/93 [==============================] - 4s 48ms/step - loss: 0.2134 - accuracy: 0.9380 - val_loss: 0.2873 - val_accuracy: 0.9107\n",
            "Epoch 76/100\n",
            "93/93 [==============================] - 3s 35ms/step - loss: 0.2017 - accuracy: 0.9353 - val_loss: 0.3555 - val_accuracy: 0.9172\n",
            "Epoch 77/100\n",
            "93/93 [==============================] - 3s 27ms/step - loss: 0.1934 - accuracy: 0.9387 - val_loss: 0.4123 - val_accuracy: 0.9069\n",
            "Epoch 78/100\n",
            "93/93 [==============================] - 3s 36ms/step - loss: 0.1960 - accuracy: 0.9367 - val_loss: 0.2682 - val_accuracy: 0.9237\n",
            "Epoch 79/100\n",
            "93/93 [==============================] - 3s 27ms/step - loss: 0.2131 - accuracy: 0.9367 - val_loss: 0.3197 - val_accuracy: 0.9366\n",
            "Epoch 80/100\n",
            "93/93 [==============================] - 2s 25ms/step - loss: 0.1697 - accuracy: 0.9467 - val_loss: 0.3710 - val_accuracy: 0.9069\n",
            "Epoch 81/100\n",
            "93/93 [==============================] - 2s 26ms/step - loss: 0.2086 - accuracy: 0.9233 - val_loss: 0.2551 - val_accuracy: 0.9392\n",
            "Epoch 82/100\n",
            "93/93 [==============================] - 2s 25ms/step - loss: 0.1542 - accuracy: 0.9467 - val_loss: 0.3045 - val_accuracy: 0.9185\n",
            "Epoch 83/100\n",
            "93/93 [==============================] - 4s 38ms/step - loss: 0.1627 - accuracy: 0.9480 - val_loss: 0.2914 - val_accuracy: 0.9418\n",
            "Epoch 84/100\n",
            "93/93 [==============================] - 2s 25ms/step - loss: 0.2327 - accuracy: 0.9253 - val_loss: 0.3433 - val_accuracy: 0.9172\n",
            "Epoch 85/100\n",
            "93/93 [==============================] - 3s 27ms/step - loss: 0.2125 - accuracy: 0.9327 - val_loss: 0.3301 - val_accuracy: 0.9301\n",
            "Epoch 86/100\n",
            "93/93 [==============================] - 3s 28ms/step - loss: 0.1730 - accuracy: 0.9487 - val_loss: 0.2580 - val_accuracy: 0.9444\n",
            "Epoch 87/100\n",
            "93/93 [==============================] - 3s 37ms/step - loss: 0.1817 - accuracy: 0.9373 - val_loss: 0.2928 - val_accuracy: 0.9353\n",
            "Epoch 88/100\n",
            "93/93 [==============================] - 3s 27ms/step - loss: 0.1909 - accuracy: 0.9427 - val_loss: 0.2906 - val_accuracy: 0.9288\n",
            "Epoch 89/100\n",
            "93/93 [==============================] - 3s 27ms/step - loss: 0.1556 - accuracy: 0.9467 - val_loss: 0.3310 - val_accuracy: 0.9366\n",
            "Epoch 90/100\n",
            "93/93 [==============================] - 2s 25ms/step - loss: 0.1446 - accuracy: 0.9540 - val_loss: 0.3399 - val_accuracy: 0.9237\n",
            "Epoch 91/100\n",
            "93/93 [==============================] - 2s 24ms/step - loss: 0.1617 - accuracy: 0.9473 - val_loss: 0.4259 - val_accuracy: 0.9094\n",
            "Epoch 92/100\n",
            "93/93 [==============================] - 2s 25ms/step - loss: 0.2018 - accuracy: 0.9313 - val_loss: 0.3275 - val_accuracy: 0.9301\n",
            "Epoch 93/100\n",
            "93/93 [==============================] - 3s 30ms/step - loss: 0.1559 - accuracy: 0.9493 - val_loss: 0.3186 - val_accuracy: 0.9327\n",
            "Epoch 94/100\n",
            "93/93 [==============================] - 3s 28ms/step - loss: 0.1723 - accuracy: 0.9447 - val_loss: 0.3189 - val_accuracy: 0.9276\n",
            "Epoch 95/100\n",
            "93/93 [==============================] - 4s 39ms/step - loss: 0.1349 - accuracy: 0.9580 - val_loss: 0.3191 - val_accuracy: 0.9327\n",
            "Epoch 96/100\n",
            "93/93 [==============================] - 4s 38ms/step - loss: 0.2013 - accuracy: 0.9373 - val_loss: 0.3173 - val_accuracy: 0.9288\n",
            "Epoch 97/100\n",
            "93/93 [==============================] - 2s 25ms/step - loss: 0.1569 - accuracy: 0.9500 - val_loss: 0.2859 - val_accuracy: 0.9340\n",
            "Epoch 98/100\n",
            "93/93 [==============================] - 4s 41ms/step - loss: 0.1479 - accuracy: 0.9507 - val_loss: 0.2752 - val_accuracy: 0.9340\n",
            "Epoch 99/100\n",
            "93/93 [==============================] - 3s 35ms/step - loss: 0.1378 - accuracy: 0.9593 - val_loss: 0.3297 - val_accuracy: 0.9276\n",
            "Epoch 100/100\n",
            "93/93 [==============================] - 2s 25ms/step - loss: 0.1554 - accuracy: 0.9500 - val_loss: 0.4137 - val_accuracy: 0.9069\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "drawAndGuess(model)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 378
        },
        "id": "VQzymwmnenes",
        "outputId": "249089c1-f62e-4326-e013-895d4e3d17d3"
      },
      "execution_count": 110,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Draw a score and then click on finish to try to recognize it.\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              "\n",
              "<canvas width=280 height=280 style=\"border:1px solid #000000;\"></canvas><br/>\n",
              "<button>Finish</button>\n",
              "<script>\n",
              "var canvas = document.querySelector('canvas')\n",
              "var ctx = canvas.getContext('2d')\n",
              "ctx.lineWidth = 20\n",
              "ctx.lineCap = 'round';\n",
              "var button = document.querySelector('button')\n",
              "var mouse = {x: 0, y: 0}\n",
              "canvas.addEventListener('mousemove', function(e) {\n",
              "  mouse.x = e.pageX - this.offsetLeft\n",
              "  mouse.y = e.pageY - this.offsetTop\n",
              "})\n",
              "canvas.onmousedown = ()=>{\n",
              "  ctx.beginPath()\n",
              "  ctx.moveTo(mouse.x, mouse.y)\n",
              "  canvas.addEventListener('mousemove', onPaint)\n",
              "}\n",
              "canvas.onmouseup = ()=>{\n",
              "  canvas.removeEventListener('mousemove', onPaint)\n",
              "}\n",
              "var onPaint = ()=>{\n",
              "  ctx.lineTo(mouse.x, mouse.y)\n",
              "  ctx.stroke()\n",
              "}\n",
              "var data = new Promise(resolve=>{\n",
              "  button.onclick = ()=>{\n",
              "    resolve(canvas.toDataURL('image/png'))\n",
              "  }\n",
              "})\n",
              "</script>\n"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "1/1 [==============================] - 0s 23ms/step\n",
            "This is a C with confidence 100.0%.\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "print(indexes_bad_prediction(X_test, Y_test))"
      ],
      "metadata": {
        "id": "H3ymsGBtxPAT"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "# **CNN**"
      ],
      "metadata": {
        "id": "jHJRfstihRak"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "from tensorflow.keras.models import Sequential\n",
        "from tensorflow.keras.layers import Conv2D, MaxPooling2D, Flatten, Dense, Dropout\n",
        "from tensorflow.keras.optimizers import Adam\n",
        "from tensorflow.keras.preprocessing.image import ImageDataGenerator\n",
        "import tensorflow.keras as keras\n",
        "\n",
        "# Adjusting the ImageDataGenerator parameters\n",
        "generator = ImageDataGenerator(\n",
        "    rotation_range=15,\n",
        "    width_shift_range=0.1,\n",
        "    height_shift_range=0.1,\n",
        "    zoom_range=0.2,\n",
        "    shear_range=10,\n",
        "    horizontal_flip=False,\n",
        "    fill_mode='nearest'\n",
        ")\n",
        "\n",
        "# Define the CNN model\n",
        "model = Sequential()\n",
        "\n",
        "# Convolutional layer block 1\n",
        "model.add(Conv2D(32, kernel_size=(3, 3), activation='relu', input_shape=(28, 28, 1)))\n",
        "model.add(MaxPooling2D(pool_size=(2, 2)))\n",
        "model.add(Dropout(0.25))\n",
        "\n",
        "# Convolutional layer block 2\n",
        "model.add(Conv2D(64, (3, 3), activation='relu'))\n",
        "#model.add(Conv2D(64, (3, 3), activation='relu'))\n",
        "model.add(MaxPooling2D(pool_size=(2, 2)))\n",
        "model.add(Dropout(0.25))\n",
        "\n",
        "model.add(Conv2D(128, (3, 3), activation='relu'))\n",
        "model.add(MaxPooling2D(pool_size=(2, 2)))\n",
        "model.add(Dropout(0.25))\n",
        "\n",
        "model.add(Flatten())  # Flatten the output and feed it into the dense layers\n",
        "model.add(Dense(128, activation='relu'))\n",
        "model.add(Dropout(0.5))\n",
        "model.add(Dense(15, activation='softmax'))  # Final layer with softmax activation for 15 classes\n",
        "\n",
        "# Compiling the model\n",
        "model.compile(loss='categorical_crossentropy',\n",
        "              optimizer=Adam(learning_rate=0.001),  # Adjusted learning rate\n",
        "              metrics=['accuracy'])\n",
        "\n",
        "# Training the CNN model with the adjusted batch size\n",
        "history = model.fit(\n",
        "    generator.flow(X_train.reshape(-1,28,28,1), Y_train, batch_size=16),  # Using a smaller batch size\n",
        "    steps_per_epoch=len(X_train) / 16,  # Adjusted to match the new batch size\n",
        "    epochs=300,  # Training for more epochs\n",
        "    validation_data=(X_test, Y_test)\n",
        ")\n",
        "\n",
        "# Visualize the training evolution\n",
        "visualize_training(history)\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "72EGBCehhQxK",
        "outputId": "fc059461-f31c-44fc-f9dd-9b8bdb9c9e90"
      },
      "execution_count": 135,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/300\n",
            "93/93 [==============================] - 5s 37ms/step - loss: 2.7082 - accuracy: 0.0693 - val_loss: 2.6740 - val_accuracy: 0.1656\n",
            "Epoch 2/300\n",
            "93/93 [==============================] - 3s 27ms/step - loss: 2.5158 - accuracy: 0.1487 - val_loss: 2.0882 - val_accuracy: 0.2135\n",
            "Epoch 3/300\n",
            "93/93 [==============================] - 3s 27ms/step - loss: 2.2312 - accuracy: 0.1973 - val_loss: 1.8067 - val_accuracy: 0.3700\n",
            "Epoch 4/300\n",
            "93/93 [==============================] - 3s 35ms/step - loss: 2.0038 - accuracy: 0.2760 - val_loss: 1.5641 - val_accuracy: 0.4502\n",
            "Epoch 5/300\n",
            "93/93 [==============================] - 2s 26ms/step - loss: 1.8152 - accuracy: 0.3253 - val_loss: 1.2059 - val_accuracy: 0.6688\n",
            "Epoch 6/300\n",
            "93/93 [==============================] - 2s 26ms/step - loss: 1.6079 - accuracy: 0.4000 - val_loss: 1.1796 - val_accuracy: 0.6145\n",
            "Epoch 7/300\n",
            "93/93 [==============================] - 2s 26ms/step - loss: 1.5021 - accuracy: 0.4467 - val_loss: 0.9850 - val_accuracy: 0.6740\n",
            "Epoch 8/300\n",
            "93/93 [==============================] - 4s 38ms/step - loss: 1.3902 - accuracy: 0.5033 - val_loss: 0.7778 - val_accuracy: 0.7335\n",
            "Epoch 9/300\n",
            "93/93 [==============================] - 3s 27ms/step - loss: 1.3150 - accuracy: 0.5260 - val_loss: 0.7415 - val_accuracy: 0.7555\n",
            "Epoch 10/300\n",
            "93/93 [==============================] - 2s 26ms/step - loss: 1.2428 - accuracy: 0.5487 - val_loss: 0.6555 - val_accuracy: 0.7710\n",
            "Epoch 11/300\n",
            "93/93 [==============================] - 2s 25ms/step - loss: 1.1988 - accuracy: 0.5593 - val_loss: 0.6351 - val_accuracy: 0.7865\n",
            "Epoch 12/300\n",
            "93/93 [==============================] - 2s 26ms/step - loss: 1.1228 - accuracy: 0.5993 - val_loss: 0.5658 - val_accuracy: 0.7969\n",
            "Epoch 13/300\n",
            "93/93 [==============================] - 4s 39ms/step - loss: 1.0887 - accuracy: 0.5873 - val_loss: 0.5443 - val_accuracy: 0.8202\n",
            "Epoch 14/300\n",
            "93/93 [==============================] - 3s 27ms/step - loss: 1.0235 - accuracy: 0.6260 - val_loss: 0.5776 - val_accuracy: 0.8008\n",
            "Epoch 15/300\n",
            "93/93 [==============================] - 2s 26ms/step - loss: 0.9934 - accuracy: 0.6633 - val_loss: 0.4634 - val_accuracy: 0.8461\n",
            "Epoch 16/300\n",
            "93/93 [==============================] - 2s 26ms/step - loss: 0.9496 - accuracy: 0.6573 - val_loss: 0.5118 - val_accuracy: 0.8331\n",
            "Epoch 17/300\n",
            "93/93 [==============================] - 4s 40ms/step - loss: 0.9130 - accuracy: 0.6760 - val_loss: 0.4349 - val_accuracy: 0.8422\n",
            "Epoch 18/300\n",
            "93/93 [==============================] - 3s 27ms/step - loss: 0.9326 - accuracy: 0.6740 - val_loss: 0.4325 - val_accuracy: 0.8499\n",
            "Epoch 19/300\n",
            "93/93 [==============================] - 3s 27ms/step - loss: 0.8872 - accuracy: 0.6707 - val_loss: 0.4139 - val_accuracy: 0.8590\n",
            "Epoch 20/300\n",
            "93/93 [==============================] - 4s 39ms/step - loss: 0.8681 - accuracy: 0.6987 - val_loss: 0.4226 - val_accuracy: 0.8577\n",
            "Epoch 21/300\n",
            "93/93 [==============================] - 3s 28ms/step - loss: 0.8707 - accuracy: 0.6800 - val_loss: 0.4473 - val_accuracy: 0.8642\n",
            "Epoch 22/300\n",
            "93/93 [==============================] - 3s 27ms/step - loss: 0.8632 - accuracy: 0.6980 - val_loss: 0.3905 - val_accuracy: 0.8616\n",
            "Epoch 23/300\n",
            "93/93 [==============================] - 4s 39ms/step - loss: 0.8237 - accuracy: 0.7073 - val_loss: 0.3663 - val_accuracy: 0.8680\n",
            "Epoch 24/300\n",
            "93/93 [==============================] - 2s 26ms/step - loss: 0.8033 - accuracy: 0.7093 - val_loss: 0.3644 - val_accuracy: 0.8810\n",
            "Epoch 25/300\n",
            "93/93 [==============================] - 2s 25ms/step - loss: 0.8113 - accuracy: 0.7273 - val_loss: 0.3550 - val_accuracy: 0.8875\n",
            "Epoch 26/300\n",
            "93/93 [==============================] - 2s 25ms/step - loss: 0.7526 - accuracy: 0.7200 - val_loss: 0.3578 - val_accuracy: 0.8875\n",
            "Epoch 27/300\n",
            "93/93 [==============================] - 3s 27ms/step - loss: 0.7742 - accuracy: 0.7287 - val_loss: 0.3391 - val_accuracy: 0.8758\n",
            "Epoch 28/300\n",
            "93/93 [==============================] - 4s 37ms/step - loss: 0.7637 - accuracy: 0.7193 - val_loss: 0.3566 - val_accuracy: 0.8836\n",
            "Epoch 29/300\n",
            "93/93 [==============================] - 2s 26ms/step - loss: 0.7501 - accuracy: 0.7293 - val_loss: 0.3599 - val_accuracy: 0.8758\n",
            "Epoch 30/300\n",
            "93/93 [==============================] - 2s 26ms/step - loss: 0.7303 - accuracy: 0.7367 - val_loss: 0.3568 - val_accuracy: 0.8784\n",
            "Epoch 31/300\n",
            "93/93 [==============================] - 2s 26ms/step - loss: 0.7369 - accuracy: 0.7340 - val_loss: 0.3511 - val_accuracy: 0.8797\n",
            "Epoch 32/300\n",
            "93/93 [==============================] - 4s 40ms/step - loss: 0.6875 - accuracy: 0.7487 - val_loss: 0.3359 - val_accuracy: 0.8823\n",
            "Epoch 33/300\n",
            "93/93 [==============================] - 2s 26ms/step - loss: 0.7335 - accuracy: 0.7540 - val_loss: 0.3285 - val_accuracy: 0.8784\n",
            "Epoch 34/300\n",
            "93/93 [==============================] - 2s 26ms/step - loss: 0.7082 - accuracy: 0.7560 - val_loss: 0.3318 - val_accuracy: 0.8887\n",
            "Epoch 35/300\n",
            "93/93 [==============================] - 2s 26ms/step - loss: 0.6865 - accuracy: 0.7527 - val_loss: 0.3176 - val_accuracy: 0.8862\n",
            "Epoch 36/300\n",
            "93/93 [==============================] - 3s 30ms/step - loss: 0.6916 - accuracy: 0.7447 - val_loss: 0.3458 - val_accuracy: 0.8745\n",
            "Epoch 37/300\n",
            "93/93 [==============================] - 3s 35ms/step - loss: 0.6946 - accuracy: 0.7493 - val_loss: 0.3121 - val_accuracy: 0.8978\n",
            "Epoch 38/300\n",
            "93/93 [==============================] - 2s 26ms/step - loss: 0.6782 - accuracy: 0.7547 - val_loss: 0.3333 - val_accuracy: 0.8797\n",
            "Epoch 39/300\n",
            "93/93 [==============================] - 3s 27ms/step - loss: 0.6570 - accuracy: 0.7627 - val_loss: 0.3370 - val_accuracy: 0.8862\n",
            "Epoch 40/300\n",
            "93/93 [==============================] - 3s 29ms/step - loss: 0.6388 - accuracy: 0.7720 - val_loss: 0.3107 - val_accuracy: 0.9069\n",
            "Epoch 41/300\n",
            "93/93 [==============================] - 3s 27ms/step - loss: 0.7093 - accuracy: 0.7473 - val_loss: 0.3113 - val_accuracy: 0.9017\n",
            "Epoch 42/300\n",
            "93/93 [==============================] - 3s 27ms/step - loss: 0.6912 - accuracy: 0.7520 - val_loss: 0.3030 - val_accuracy: 0.9043\n",
            "Epoch 43/300\n",
            "93/93 [==============================] - 4s 37ms/step - loss: 0.6373 - accuracy: 0.7753 - val_loss: 0.3005 - val_accuracy: 0.9056\n",
            "Epoch 44/300\n",
            "93/93 [==============================] - 3s 27ms/step - loss: 0.6512 - accuracy: 0.7567 - val_loss: 0.3004 - val_accuracy: 0.8965\n",
            "Epoch 45/300\n",
            "93/93 [==============================] - 2s 26ms/step - loss: 0.6442 - accuracy: 0.7700 - val_loss: 0.3092 - val_accuracy: 0.8900\n",
            "Epoch 46/300\n",
            "93/93 [==============================] - 4s 38ms/step - loss: 0.6414 - accuracy: 0.7720 - val_loss: 0.2924 - val_accuracy: 0.9082\n",
            "Epoch 47/300\n",
            "93/93 [==============================] - 2s 25ms/step - loss: 0.6316 - accuracy: 0.7753 - val_loss: 0.3119 - val_accuracy: 0.8836\n",
            "Epoch 48/300\n",
            "93/93 [==============================] - 2s 25ms/step - loss: 0.5869 - accuracy: 0.7900 - val_loss: 0.2937 - val_accuracy: 0.9094\n",
            "Epoch 49/300\n",
            "93/93 [==============================] - 3s 27ms/step - loss: 0.5698 - accuracy: 0.7987 - val_loss: 0.3192 - val_accuracy: 0.8952\n",
            "Epoch 50/300\n",
            "93/93 [==============================] - 2s 26ms/step - loss: 0.6000 - accuracy: 0.7820 - val_loss: 0.2894 - val_accuracy: 0.9017\n",
            "Epoch 51/300\n",
            "93/93 [==============================] - 4s 40ms/step - loss: 0.6387 - accuracy: 0.7787 - val_loss: 0.3022 - val_accuracy: 0.8952\n",
            "Epoch 52/300\n",
            "93/93 [==============================] - 3s 27ms/step - loss: 0.6251 - accuracy: 0.7787 - val_loss: 0.2879 - val_accuracy: 0.9146\n",
            "Epoch 53/300\n",
            "93/93 [==============================] - 3s 27ms/step - loss: 0.5886 - accuracy: 0.7940 - val_loss: 0.2935 - val_accuracy: 0.9017\n",
            "Epoch 54/300\n",
            "93/93 [==============================] - 4s 40ms/step - loss: 0.5704 - accuracy: 0.7927 - val_loss: 0.2976 - val_accuracy: 0.8797\n",
            "Epoch 55/300\n",
            "93/93 [==============================] - 2s 25ms/step - loss: 0.5872 - accuracy: 0.7873 - val_loss: 0.2928 - val_accuracy: 0.9017\n",
            "Epoch 56/300\n",
            "93/93 [==============================] - 2s 26ms/step - loss: 0.5868 - accuracy: 0.8000 - val_loss: 0.2844 - val_accuracy: 0.9056\n",
            "Epoch 57/300\n",
            "93/93 [==============================] - 2s 26ms/step - loss: 0.5866 - accuracy: 0.7893 - val_loss: 0.2986 - val_accuracy: 0.9056\n",
            "Epoch 58/300\n",
            "93/93 [==============================] - 3s 35ms/step - loss: 0.6114 - accuracy: 0.7833 - val_loss: 0.3171 - val_accuracy: 0.8758\n",
            "Epoch 59/300\n",
            "93/93 [==============================] - 3s 27ms/step - loss: 0.5608 - accuracy: 0.7987 - val_loss: 0.2944 - val_accuracy: 0.9017\n",
            "Epoch 60/300\n",
            "93/93 [==============================] - 2s 25ms/step - loss: 0.5440 - accuracy: 0.8013 - val_loss: 0.2904 - val_accuracy: 0.9069\n",
            "Epoch 61/300\n",
            "93/93 [==============================] - 3s 35ms/step - loss: 0.5421 - accuracy: 0.8113 - val_loss: 0.3006 - val_accuracy: 0.8939\n",
            "Epoch 62/300\n",
            "93/93 [==============================] - 3s 28ms/step - loss: 0.5445 - accuracy: 0.8067 - val_loss: 0.2889 - val_accuracy: 0.8991\n",
            "Epoch 63/300\n",
            "93/93 [==============================] - 3s 27ms/step - loss: 0.5416 - accuracy: 0.8113 - val_loss: 0.2933 - val_accuracy: 0.9082\n",
            "Epoch 64/300\n",
            "93/93 [==============================] - 4s 39ms/step - loss: 0.5622 - accuracy: 0.8040 - val_loss: 0.2930 - val_accuracy: 0.8926\n",
            "Epoch 65/300\n",
            "93/93 [==============================] - 3s 26ms/step - loss: 0.5753 - accuracy: 0.7867 - val_loss: 0.2917 - val_accuracy: 0.9004\n",
            "Epoch 66/300\n",
            "93/93 [==============================] - 3s 32ms/step - loss: 0.5814 - accuracy: 0.7953 - val_loss: 0.2838 - val_accuracy: 0.9094\n",
            "Epoch 67/300\n",
            "93/93 [==============================] - 3s 33ms/step - loss: 0.5437 - accuracy: 0.7987 - val_loss: 0.2903 - val_accuracy: 0.9056\n",
            "Epoch 68/300\n",
            "93/93 [==============================] - 2s 26ms/step - loss: 0.5430 - accuracy: 0.8053 - val_loss: 0.2923 - val_accuracy: 0.9069\n",
            "Epoch 69/300\n",
            "93/93 [==============================] - 2s 25ms/step - loss: 0.5382 - accuracy: 0.8060 - val_loss: 0.2920 - val_accuracy: 0.9082\n",
            "Epoch 70/300\n",
            "93/93 [==============================] - 2s 26ms/step - loss: 0.5340 - accuracy: 0.8147 - val_loss: 0.2923 - val_accuracy: 0.9030\n",
            "Epoch 71/300\n",
            "93/93 [==============================] - 3s 34ms/step - loss: 0.5557 - accuracy: 0.8067 - val_loss: 0.2835 - val_accuracy: 0.9146\n",
            "Epoch 72/300\n",
            "93/93 [==============================] - 3s 30ms/step - loss: 0.5488 - accuracy: 0.8113 - val_loss: 0.2708 - val_accuracy: 0.9120\n",
            "Epoch 73/300\n",
            "93/93 [==============================] - 2s 26ms/step - loss: 0.4957 - accuracy: 0.8180 - val_loss: 0.2622 - val_accuracy: 0.9107\n",
            "Epoch 74/300\n",
            "93/93 [==============================] - 2s 26ms/step - loss: 0.5383 - accuracy: 0.8053 - val_loss: 0.2668 - val_accuracy: 0.9082\n",
            "Epoch 75/300\n",
            "93/93 [==============================] - 4s 40ms/step - loss: 0.5331 - accuracy: 0.8027 - val_loss: 0.2869 - val_accuracy: 0.8939\n",
            "Epoch 76/300\n",
            "93/93 [==============================] - 3s 27ms/step - loss: 0.5593 - accuracy: 0.8000 - val_loss: 0.2855 - val_accuracy: 0.9069\n",
            "Epoch 77/300\n",
            "93/93 [==============================] - 2s 26ms/step - loss: 0.5211 - accuracy: 0.8193 - val_loss: 0.2820 - val_accuracy: 0.9043\n",
            "Epoch 78/300\n",
            "93/93 [==============================] - 2s 26ms/step - loss: 0.5382 - accuracy: 0.8107 - val_loss: 0.2780 - val_accuracy: 0.9094\n",
            "Epoch 79/300\n",
            "93/93 [==============================] - 3s 36ms/step - loss: 0.5128 - accuracy: 0.8220 - val_loss: 0.2614 - val_accuracy: 0.9172\n",
            "Epoch 80/300\n",
            "93/93 [==============================] - 2s 26ms/step - loss: 0.5497 - accuracy: 0.8067 - val_loss: 0.2728 - val_accuracy: 0.9017\n",
            "Epoch 81/300\n",
            "93/93 [==============================] - 2s 25ms/step - loss: 0.5050 - accuracy: 0.8180 - val_loss: 0.2827 - val_accuracy: 0.9043\n",
            "Epoch 82/300\n",
            "93/93 [==============================] - 2s 26ms/step - loss: 0.5230 - accuracy: 0.8007 - val_loss: 0.2672 - val_accuracy: 0.9146\n",
            "Epoch 83/300\n",
            "93/93 [==============================] - 3s 36ms/step - loss: 0.5108 - accuracy: 0.8120 - val_loss: 0.2654 - val_accuracy: 0.9159\n",
            "Epoch 84/300\n",
            "93/93 [==============================] - 2s 26ms/step - loss: 0.4822 - accuracy: 0.8347 - val_loss: 0.2663 - val_accuracy: 0.9133\n",
            "Epoch 85/300\n",
            "93/93 [==============================] - 2s 26ms/step - loss: 0.5136 - accuracy: 0.8213 - val_loss: 0.2673 - val_accuracy: 0.9146\n",
            "Epoch 86/300\n",
            "93/93 [==============================] - 3s 27ms/step - loss: 0.4906 - accuracy: 0.8160 - val_loss: 0.2582 - val_accuracy: 0.9211\n",
            "Epoch 87/300\n",
            "93/93 [==============================] - 4s 39ms/step - loss: 0.4961 - accuracy: 0.8160 - val_loss: 0.2694 - val_accuracy: 0.9224\n",
            "Epoch 88/300\n",
            "93/93 [==============================] - 2s 26ms/step - loss: 0.5372 - accuracy: 0.8047 - val_loss: 0.2726 - val_accuracy: 0.9069\n",
            "Epoch 89/300\n",
            "93/93 [==============================] - 2s 26ms/step - loss: 0.4801 - accuracy: 0.8300 - val_loss: 0.2632 - val_accuracy: 0.9146\n",
            "Epoch 90/300\n",
            "93/93 [==============================] - 2s 26ms/step - loss: 0.4970 - accuracy: 0.8240 - val_loss: 0.2551 - val_accuracy: 0.9211\n",
            "Epoch 91/300\n",
            "93/93 [==============================] - 4s 38ms/step - loss: 0.4945 - accuracy: 0.8173 - val_loss: 0.2735 - val_accuracy: 0.9133\n",
            "Epoch 92/300\n",
            "93/93 [==============================] - 3s 27ms/step - loss: 0.4959 - accuracy: 0.8253 - val_loss: 0.2744 - val_accuracy: 0.9159\n",
            "Epoch 93/300\n",
            "93/93 [==============================] - 2s 26ms/step - loss: 0.4873 - accuracy: 0.8300 - val_loss: 0.2761 - val_accuracy: 0.9224\n",
            "Epoch 94/300\n",
            "93/93 [==============================] - 2s 26ms/step - loss: 0.4460 - accuracy: 0.8347 - val_loss: 0.2543 - val_accuracy: 0.9237\n",
            "Epoch 95/300\n",
            "93/93 [==============================] - 4s 40ms/step - loss: 0.5213 - accuracy: 0.8160 - val_loss: 0.2620 - val_accuracy: 0.9250\n",
            "Epoch 96/300\n",
            "93/93 [==============================] - 3s 27ms/step - loss: 0.4450 - accuracy: 0.8387 - val_loss: 0.2602 - val_accuracy: 0.9185\n",
            "Epoch 97/300\n",
            "93/93 [==============================] - 2s 25ms/step - loss: 0.4727 - accuracy: 0.8287 - val_loss: 0.2717 - val_accuracy: 0.9172\n",
            "Epoch 98/300\n",
            "93/93 [==============================] - 2s 25ms/step - loss: 0.4681 - accuracy: 0.8213 - val_loss: 0.2586 - val_accuracy: 0.9314\n",
            "Epoch 99/300\n",
            "93/93 [==============================] - 4s 41ms/step - loss: 0.4657 - accuracy: 0.8347 - val_loss: 0.2543 - val_accuracy: 0.9276\n",
            "Epoch 100/300\n",
            "93/93 [==============================] - 2s 26ms/step - loss: 0.4983 - accuracy: 0.8227 - val_loss: 0.2530 - val_accuracy: 0.9276\n",
            "Epoch 101/300\n",
            "93/93 [==============================] - 2s 26ms/step - loss: 0.4860 - accuracy: 0.8153 - val_loss: 0.2550 - val_accuracy: 0.9185\n",
            "Epoch 102/300\n",
            "93/93 [==============================] - 3s 27ms/step - loss: 0.4640 - accuracy: 0.8287 - val_loss: 0.2659 - val_accuracy: 0.9133\n",
            "Epoch 103/300\n",
            "93/93 [==============================] - 4s 39ms/step - loss: 0.4582 - accuracy: 0.8340 - val_loss: 0.2641 - val_accuracy: 0.9120\n",
            "Epoch 104/300\n",
            "93/93 [==============================] - 3s 27ms/step - loss: 0.4832 - accuracy: 0.8233 - val_loss: 0.2706 - val_accuracy: 0.9185\n",
            "Epoch 105/300\n",
            "93/93 [==============================] - 2s 26ms/step - loss: 0.4915 - accuracy: 0.8307 - val_loss: 0.2717 - val_accuracy: 0.9133\n",
            "Epoch 106/300\n",
            "93/93 [==============================] - 3s 27ms/step - loss: 0.4545 - accuracy: 0.8373 - val_loss: 0.2552 - val_accuracy: 0.9211\n",
            "Epoch 107/300\n",
            "93/93 [==============================] - 4s 39ms/step - loss: 0.4637 - accuracy: 0.8427 - val_loss: 0.2628 - val_accuracy: 0.9288\n",
            "Epoch 108/300\n",
            "93/93 [==============================] - 2s 25ms/step - loss: 0.4653 - accuracy: 0.8240 - val_loss: 0.2585 - val_accuracy: 0.9237\n",
            "Epoch 109/300\n",
            "93/93 [==============================] - 2s 26ms/step - loss: 0.4958 - accuracy: 0.8280 - val_loss: 0.2563 - val_accuracy: 0.9250\n",
            "Epoch 110/300\n",
            "93/93 [==============================] - 3s 27ms/step - loss: 0.4601 - accuracy: 0.8327 - val_loss: 0.2631 - val_accuracy: 0.9159\n",
            "Epoch 111/300\n",
            "93/93 [==============================] - 2s 25ms/step - loss: 0.4438 - accuracy: 0.8460 - val_loss: 0.2494 - val_accuracy: 0.9172\n",
            "Epoch 112/300\n",
            "93/93 [==============================] - 4s 38ms/step - loss: 0.4470 - accuracy: 0.8333 - val_loss: 0.2645 - val_accuracy: 0.9250\n",
            "Epoch 113/300\n",
            "93/93 [==============================] - 2s 25ms/step - loss: 0.5142 - accuracy: 0.8180 - val_loss: 0.2476 - val_accuracy: 0.9276\n",
            "Epoch 114/300\n",
            "93/93 [==============================] - 3s 26ms/step - loss: 0.4658 - accuracy: 0.8393 - val_loss: 0.2642 - val_accuracy: 0.9237\n",
            "Epoch 115/300\n",
            "93/93 [==============================] - 2s 26ms/step - loss: 0.4661 - accuracy: 0.8267 - val_loss: 0.2537 - val_accuracy: 0.9301\n",
            "Epoch 116/300\n",
            "93/93 [==============================] - 4s 40ms/step - loss: 0.4746 - accuracy: 0.8287 - val_loss: 0.2576 - val_accuracy: 0.9224\n",
            "Epoch 117/300\n",
            "93/93 [==============================] - 2s 25ms/step - loss: 0.4385 - accuracy: 0.8380 - val_loss: 0.2831 - val_accuracy: 0.9250\n",
            "Epoch 118/300\n",
            "93/93 [==============================] - 2s 25ms/step - loss: 0.4801 - accuracy: 0.8307 - val_loss: 0.2617 - val_accuracy: 0.9211\n",
            "Epoch 119/300\n",
            "93/93 [==============================] - 2s 25ms/step - loss: 0.4329 - accuracy: 0.8440 - val_loss: 0.2621 - val_accuracy: 0.9237\n",
            "Epoch 120/300\n",
            "93/93 [==============================] - 3s 35ms/step - loss: 0.4729 - accuracy: 0.8293 - val_loss: 0.2780 - val_accuracy: 0.8887\n",
            "Epoch 121/300\n",
            "93/93 [==============================] - 3s 28ms/step - loss: 0.4289 - accuracy: 0.8433 - val_loss: 0.2664 - val_accuracy: 0.9172\n",
            "Epoch 122/300\n",
            "93/93 [==============================] - 2s 26ms/step - loss: 0.4821 - accuracy: 0.8200 - val_loss: 0.2713 - val_accuracy: 0.9056\n",
            "Epoch 123/300\n",
            "93/93 [==============================] - 4s 39ms/step - loss: 0.4665 - accuracy: 0.8187 - val_loss: 0.2617 - val_accuracy: 0.9263\n",
            "Epoch 124/300\n",
            "93/93 [==============================] - 3s 27ms/step - loss: 0.4240 - accuracy: 0.8480 - val_loss: 0.2676 - val_accuracy: 0.9340\n",
            "Epoch 125/300\n",
            "93/93 [==============================] - 2s 26ms/step - loss: 0.4229 - accuracy: 0.8520 - val_loss: 0.2726 - val_accuracy: 0.9172\n",
            "Epoch 126/300\n",
            "93/93 [==============================] - 2s 26ms/step - loss: 0.4708 - accuracy: 0.8293 - val_loss: 0.2636 - val_accuracy: 0.9263\n",
            "Epoch 127/300\n",
            "93/93 [==============================] - 3s 36ms/step - loss: 0.4890 - accuracy: 0.8307 - val_loss: 0.2585 - val_accuracy: 0.9301\n",
            "Epoch 128/300\n",
            "93/93 [==============================] - 3s 26ms/step - loss: 0.4554 - accuracy: 0.8253 - val_loss: 0.2606 - val_accuracy: 0.9185\n",
            "Epoch 129/300\n",
            "93/93 [==============================] - 2s 26ms/step - loss: 0.4197 - accuracy: 0.8453 - val_loss: 0.2659 - val_accuracy: 0.9237\n",
            "Epoch 130/300\n",
            "93/93 [==============================] - 2s 25ms/step - loss: 0.4392 - accuracy: 0.8420 - val_loss: 0.2495 - val_accuracy: 0.9224\n",
            "Epoch 131/300\n",
            "93/93 [==============================] - 4s 38ms/step - loss: 0.4834 - accuracy: 0.8293 - val_loss: 0.2546 - val_accuracy: 0.9185\n",
            "Epoch 132/300\n",
            "93/93 [==============================] - 3s 27ms/step - loss: 0.4485 - accuracy: 0.8373 - val_loss: 0.2470 - val_accuracy: 0.9224\n",
            "Epoch 133/300\n",
            "93/93 [==============================] - 2s 26ms/step - loss: 0.4168 - accuracy: 0.8433 - val_loss: 0.2371 - val_accuracy: 0.9276\n",
            "Epoch 134/300\n",
            "93/93 [==============================] - 2s 26ms/step - loss: 0.4643 - accuracy: 0.8240 - val_loss: 0.2497 - val_accuracy: 0.9120\n",
            "Epoch 135/300\n",
            "93/93 [==============================] - 4s 38ms/step - loss: 0.4313 - accuracy: 0.8447 - val_loss: 0.2622 - val_accuracy: 0.9237\n",
            "Epoch 136/300\n",
            "93/93 [==============================] - 3s 28ms/step - loss: 0.4066 - accuracy: 0.8493 - val_loss: 0.2795 - val_accuracy: 0.9237\n",
            "Epoch 137/300\n",
            "93/93 [==============================] - 3s 27ms/step - loss: 0.4532 - accuracy: 0.8327 - val_loss: 0.2779 - val_accuracy: 0.9133\n",
            "Epoch 138/300\n",
            "93/93 [==============================] - 4s 42ms/step - loss: 0.4531 - accuracy: 0.8287 - val_loss: 0.2519 - val_accuracy: 0.9250\n",
            "Epoch 139/300\n",
            "93/93 [==============================] - 3s 28ms/step - loss: 0.4632 - accuracy: 0.8280 - val_loss: 0.2512 - val_accuracy: 0.9224\n",
            "Epoch 140/300\n",
            "93/93 [==============================] - 3s 27ms/step - loss: 0.4602 - accuracy: 0.8307 - val_loss: 0.2481 - val_accuracy: 0.9263\n",
            "Epoch 141/300\n",
            "93/93 [==============================] - 3s 32ms/step - loss: 0.4641 - accuracy: 0.8320 - val_loss: 0.2697 - val_accuracy: 0.9107\n",
            "Epoch 142/300\n",
            "93/93 [==============================] - 2s 26ms/step - loss: 0.4360 - accuracy: 0.8300 - val_loss: 0.2628 - val_accuracy: 0.9211\n",
            "Epoch 143/300\n",
            "93/93 [==============================] - 2s 26ms/step - loss: 0.4556 - accuracy: 0.8467 - val_loss: 0.2724 - val_accuracy: 0.9017\n",
            "Epoch 144/300\n",
            "93/93 [==============================] - 2s 26ms/step - loss: 0.4641 - accuracy: 0.8327 - val_loss: 0.2570 - val_accuracy: 0.9185\n",
            "Epoch 145/300\n",
            "93/93 [==============================] - 4s 38ms/step - loss: 0.4406 - accuracy: 0.8347 - val_loss: 0.2657 - val_accuracy: 0.9185\n",
            "Epoch 146/300\n",
            "93/93 [==============================] - 3s 27ms/step - loss: 0.3989 - accuracy: 0.8453 - val_loss: 0.2406 - val_accuracy: 0.9237\n",
            "Epoch 147/300\n",
            "93/93 [==============================] - 3s 27ms/step - loss: 0.4332 - accuracy: 0.8413 - val_loss: 0.2524 - val_accuracy: 0.9288\n",
            "Epoch 148/300\n",
            "93/93 [==============================] - 2s 26ms/step - loss: 0.4385 - accuracy: 0.8420 - val_loss: 0.2437 - val_accuracy: 0.9276\n",
            "Epoch 149/300\n",
            "93/93 [==============================] - 2s 26ms/step - loss: 0.4086 - accuracy: 0.8553 - val_loss: 0.2542 - val_accuracy: 0.9314\n",
            "Epoch 150/300\n",
            "93/93 [==============================] - 4s 40ms/step - loss: 0.4131 - accuracy: 0.8513 - val_loss: 0.2548 - val_accuracy: 0.9198\n",
            "Epoch 151/300\n",
            "93/93 [==============================] - 3s 27ms/step - loss: 0.4134 - accuracy: 0.8580 - val_loss: 0.2530 - val_accuracy: 0.9146\n",
            "Epoch 152/300\n",
            "93/93 [==============================] - 2s 26ms/step - loss: 0.4554 - accuracy: 0.8287 - val_loss: 0.2650 - val_accuracy: 0.9120\n",
            "Epoch 153/300\n",
            "93/93 [==============================] - 2s 26ms/step - loss: 0.4215 - accuracy: 0.8527 - val_loss: 0.2610 - val_accuracy: 0.9030\n",
            "Epoch 154/300\n",
            "93/93 [==============================] - 4s 42ms/step - loss: 0.4275 - accuracy: 0.8367 - val_loss: 0.2453 - val_accuracy: 0.9159\n",
            "Epoch 155/300\n",
            "93/93 [==============================] - 3s 27ms/step - loss: 0.4238 - accuracy: 0.8413 - val_loss: 0.2412 - val_accuracy: 0.9185\n",
            "Epoch 156/300\n",
            "93/93 [==============================] - 3s 27ms/step - loss: 0.4476 - accuracy: 0.8433 - val_loss: 0.2599 - val_accuracy: 0.9094\n",
            "Epoch 157/300\n",
            "93/93 [==============================] - 3s 27ms/step - loss: 0.3894 - accuracy: 0.8507 - val_loss: 0.2673 - val_accuracy: 0.9237\n",
            "Epoch 158/300\n",
            "93/93 [==============================] - 3s 26ms/step - loss: 0.4120 - accuracy: 0.8513 - val_loss: 0.2603 - val_accuracy: 0.9263\n",
            "Epoch 159/300\n",
            "93/93 [==============================] - 2s 25ms/step - loss: 0.4580 - accuracy: 0.8440 - val_loss: 0.2540 - val_accuracy: 0.9224\n",
            "Epoch 160/300\n",
            "93/93 [==============================] - 3s 30ms/step - loss: 0.4102 - accuracy: 0.8433 - val_loss: 0.2509 - val_accuracy: 0.9185\n",
            "Epoch 161/300\n",
            "93/93 [==============================] - 3s 29ms/step - loss: 0.4174 - accuracy: 0.8580 - val_loss: 0.2419 - val_accuracy: 0.9211\n",
            "Epoch 162/300\n",
            "93/93 [==============================] - 3s 27ms/step - loss: 0.4556 - accuracy: 0.8407 - val_loss: 0.2403 - val_accuracy: 0.9250\n",
            "Epoch 163/300\n",
            "93/93 [==============================] - 3s 27ms/step - loss: 0.3977 - accuracy: 0.8567 - val_loss: 0.2467 - val_accuracy: 0.9263\n",
            "Epoch 164/300\n",
            "93/93 [==============================] - 2s 26ms/step - loss: 0.4071 - accuracy: 0.8473 - val_loss: 0.2472 - val_accuracy: 0.9224\n",
            "Epoch 165/300\n",
            "93/93 [==============================] - 3s 27ms/step - loss: 0.4413 - accuracy: 0.8340 - val_loss: 0.2360 - val_accuracy: 0.9301\n",
            "Epoch 166/300\n",
            "93/93 [==============================] - 4s 40ms/step - loss: 0.3991 - accuracy: 0.8560 - val_loss: 0.2492 - val_accuracy: 0.9288\n",
            "Epoch 167/300\n",
            "93/93 [==============================] - 2s 26ms/step - loss: 0.4097 - accuracy: 0.8460 - val_loss: 0.2609 - val_accuracy: 0.9288\n",
            "Epoch 168/300\n",
            "93/93 [==============================] - 3s 27ms/step - loss: 0.4476 - accuracy: 0.8307 - val_loss: 0.2578 - val_accuracy: 0.9276\n",
            "Epoch 169/300\n",
            "93/93 [==============================] - 4s 41ms/step - loss: 0.4273 - accuracy: 0.8573 - val_loss: 0.2685 - val_accuracy: 0.9172\n",
            "Epoch 170/300\n",
            "93/93 [==============================] - 3s 27ms/step - loss: 0.4514 - accuracy: 0.8453 - val_loss: 0.2703 - val_accuracy: 0.9107\n",
            "Epoch 171/300\n",
            "93/93 [==============================] - 3s 27ms/step - loss: 0.4107 - accuracy: 0.8447 - val_loss: 0.2615 - val_accuracy: 0.9301\n",
            "Epoch 172/300\n",
            "93/93 [==============================] - 3s 27ms/step - loss: 0.4393 - accuracy: 0.8313 - val_loss: 0.2472 - val_accuracy: 0.9314\n",
            "Epoch 173/300\n",
            "93/93 [==============================] - 4s 38ms/step - loss: 0.4251 - accuracy: 0.8420 - val_loss: 0.2453 - val_accuracy: 0.9237\n",
            "Epoch 174/300\n",
            "93/93 [==============================] - 3s 27ms/step - loss: 0.4353 - accuracy: 0.8447 - val_loss: 0.2331 - val_accuracy: 0.9327\n",
            "Epoch 175/300\n",
            "93/93 [==============================] - 2s 26ms/step - loss: 0.3920 - accuracy: 0.8540 - val_loss: 0.2439 - val_accuracy: 0.9263\n",
            "Epoch 176/300\n",
            "93/93 [==============================] - 2s 26ms/step - loss: 0.4366 - accuracy: 0.8413 - val_loss: 0.2416 - val_accuracy: 0.9276\n",
            "Epoch 177/300\n",
            "93/93 [==============================] - 3s 35ms/step - loss: 0.4200 - accuracy: 0.8487 - val_loss: 0.2571 - val_accuracy: 0.9159\n",
            "Epoch 178/300\n",
            "93/93 [==============================] - 2s 26ms/step - loss: 0.4368 - accuracy: 0.8440 - val_loss: 0.2479 - val_accuracy: 0.9263\n",
            "Epoch 179/300\n",
            "93/93 [==============================] - 2s 26ms/step - loss: 0.4163 - accuracy: 0.8493 - val_loss: 0.2645 - val_accuracy: 0.9224\n",
            "Epoch 180/300\n",
            "93/93 [==============================] - 3s 27ms/step - loss: 0.3990 - accuracy: 0.8573 - val_loss: 0.2550 - val_accuracy: 0.9314\n",
            "Epoch 181/300\n",
            "93/93 [==============================] - 3s 35ms/step - loss: 0.4402 - accuracy: 0.8367 - val_loss: 0.2656 - val_accuracy: 0.9288\n",
            "Epoch 182/300\n",
            "93/93 [==============================] - 3s 31ms/step - loss: 0.4034 - accuracy: 0.8487 - val_loss: 0.2724 - val_accuracy: 0.9276\n",
            "Epoch 183/300\n",
            "93/93 [==============================] - 3s 27ms/step - loss: 0.4223 - accuracy: 0.8427 - val_loss: 0.2749 - val_accuracy: 0.9030\n",
            "Epoch 184/300\n",
            "93/93 [==============================] - 3s 27ms/step - loss: 0.4242 - accuracy: 0.8427 - val_loss: 0.2605 - val_accuracy: 0.9276\n",
            "Epoch 185/300\n",
            "93/93 [==============================] - 4s 40ms/step - loss: 0.4199 - accuracy: 0.8467 - val_loss: 0.2574 - val_accuracy: 0.9237\n",
            "Epoch 186/300\n",
            "93/93 [==============================] - 2s 26ms/step - loss: 0.3828 - accuracy: 0.8627 - val_loss: 0.2543 - val_accuracy: 0.9301\n",
            "Epoch 187/300\n",
            "93/93 [==============================] - 2s 26ms/step - loss: 0.3873 - accuracy: 0.8553 - val_loss: 0.2570 - val_accuracy: 0.9250\n",
            "Epoch 188/300\n",
            "93/93 [==============================] - 2s 26ms/step - loss: 0.4292 - accuracy: 0.8420 - val_loss: 0.2445 - val_accuracy: 0.9263\n",
            "Epoch 189/300\n",
            "93/93 [==============================] - 4s 38ms/step - loss: 0.4070 - accuracy: 0.8580 - val_loss: 0.2468 - val_accuracy: 0.9288\n",
            "Epoch 190/300\n",
            "93/93 [==============================] - 3s 27ms/step - loss: 0.4289 - accuracy: 0.8500 - val_loss: 0.2576 - val_accuracy: 0.9185\n",
            "Epoch 191/300\n",
            "93/93 [==============================] - 2s 26ms/step - loss: 0.4111 - accuracy: 0.8507 - val_loss: 0.2268 - val_accuracy: 0.9288\n",
            "Epoch 192/300\n",
            "93/93 [==============================] - 3s 35ms/step - loss: 0.4310 - accuracy: 0.8407 - val_loss: 0.2401 - val_accuracy: 0.9276\n",
            "Epoch 193/300\n",
            "93/93 [==============================] - 2s 25ms/step - loss: 0.4359 - accuracy: 0.8413 - val_loss: 0.2362 - val_accuracy: 0.9353\n",
            "Epoch 194/300\n",
            "93/93 [==============================] - 3s 27ms/step - loss: 0.4015 - accuracy: 0.8680 - val_loss: 0.2432 - val_accuracy: 0.9405\n",
            "Epoch 195/300\n",
            "93/93 [==============================] - 2s 25ms/step - loss: 0.3904 - accuracy: 0.8587 - val_loss: 0.2548 - val_accuracy: 0.9185\n",
            "Epoch 196/300\n",
            "93/93 [==============================] - 4s 38ms/step - loss: 0.4244 - accuracy: 0.8453 - val_loss: 0.2383 - val_accuracy: 0.9314\n",
            "Epoch 197/300\n",
            "93/93 [==============================] - 3s 26ms/step - loss: 0.4761 - accuracy: 0.8340 - val_loss: 0.2389 - val_accuracy: 0.9340\n",
            "Epoch 198/300\n",
            "93/93 [==============================] - 2s 25ms/step - loss: 0.3923 - accuracy: 0.8627 - val_loss: 0.2758 - val_accuracy: 0.9004\n",
            "Epoch 199/300\n",
            "93/93 [==============================] - 2s 26ms/step - loss: 0.3708 - accuracy: 0.8633 - val_loss: 0.2615 - val_accuracy: 0.9379\n",
            "Epoch 200/300\n",
            "93/93 [==============================] - 3s 27ms/step - loss: 0.4578 - accuracy: 0.8467 - val_loss: 0.2462 - val_accuracy: 0.9198\n",
            "Epoch 201/300\n",
            "93/93 [==============================] - 4s 40ms/step - loss: 0.4132 - accuracy: 0.8587 - val_loss: 0.2438 - val_accuracy: 0.9250\n",
            "Epoch 202/300\n",
            "93/93 [==============================] - 3s 27ms/step - loss: 0.3964 - accuracy: 0.8627 - val_loss: 0.2593 - val_accuracy: 0.9276\n",
            "Epoch 203/300\n",
            "93/93 [==============================] - 2s 26ms/step - loss: 0.3951 - accuracy: 0.8533 - val_loss: 0.2595 - val_accuracy: 0.9301\n",
            "Epoch 204/300\n",
            "93/93 [==============================] - 2s 26ms/step - loss: 0.4376 - accuracy: 0.8513 - val_loss: 0.2466 - val_accuracy: 0.9327\n",
            "Epoch 205/300\n",
            "93/93 [==============================] - 4s 38ms/step - loss: 0.4209 - accuracy: 0.8487 - val_loss: 0.2365 - val_accuracy: 0.9379\n",
            "Epoch 206/300\n",
            "93/93 [==============================] - 3s 27ms/step - loss: 0.4086 - accuracy: 0.8593 - val_loss: 0.2384 - val_accuracy: 0.9146\n",
            "Epoch 207/300\n",
            "93/93 [==============================] - 2s 25ms/step - loss: 0.3955 - accuracy: 0.8533 - val_loss: 0.2364 - val_accuracy: 0.9353\n",
            "Epoch 208/300\n",
            "93/93 [==============================] - 2s 26ms/step - loss: 0.3909 - accuracy: 0.8633 - val_loss: 0.2530 - val_accuracy: 0.9301\n",
            "Epoch 209/300\n",
            "93/93 [==============================] - 4s 39ms/step - loss: 0.4383 - accuracy: 0.8460 - val_loss: 0.2467 - val_accuracy: 0.9237\n",
            "Epoch 210/300\n",
            "93/93 [==============================] - 3s 27ms/step - loss: 0.4113 - accuracy: 0.8553 - val_loss: 0.2464 - val_accuracy: 0.9198\n",
            "Epoch 211/300\n",
            "93/93 [==============================] - 2s 26ms/step - loss: 0.3867 - accuracy: 0.8613 - val_loss: 0.2603 - val_accuracy: 0.9120\n",
            "Epoch 212/300\n",
            "93/93 [==============================] - 4s 41ms/step - loss: 0.3846 - accuracy: 0.8540 - val_loss: 0.2879 - val_accuracy: 0.9159\n",
            "Epoch 213/300\n",
            "93/93 [==============================] - 2s 26ms/step - loss: 0.4706 - accuracy: 0.8367 - val_loss: 0.2425 - val_accuracy: 0.9224\n",
            "Epoch 214/300\n",
            "93/93 [==============================] - 2s 26ms/step - loss: 0.4634 - accuracy: 0.8300 - val_loss: 0.2350 - val_accuracy: 0.9314\n",
            "Epoch 215/300\n",
            "93/93 [==============================] - 3s 27ms/step - loss: 0.3695 - accuracy: 0.8500 - val_loss: 0.2388 - val_accuracy: 0.9340\n",
            "Epoch 216/300\n",
            "93/93 [==============================] - 2s 25ms/step - loss: 0.4082 - accuracy: 0.8540 - val_loss: 0.2505 - val_accuracy: 0.9146\n",
            "Epoch 217/300\n",
            "93/93 [==============================] - 3s 27ms/step - loss: 0.4102 - accuracy: 0.8473 - val_loss: 0.2494 - val_accuracy: 0.9250\n",
            "Epoch 218/300\n",
            "93/93 [==============================] - 2s 25ms/step - loss: 0.4111 - accuracy: 0.8513 - val_loss: 0.2395 - val_accuracy: 0.9276\n",
            "Epoch 219/300\n",
            "93/93 [==============================] - 2s 25ms/step - loss: 0.3869 - accuracy: 0.8620 - val_loss: 0.2416 - val_accuracy: 0.9314\n",
            "Epoch 220/300\n",
            "93/93 [==============================] - 4s 40ms/step - loss: 0.3785 - accuracy: 0.8533 - val_loss: 0.2430 - val_accuracy: 0.9263\n",
            "Epoch 221/300\n",
            "93/93 [==============================] - 3s 27ms/step - loss: 0.4221 - accuracy: 0.8467 - val_loss: 0.2623 - val_accuracy: 0.9107\n",
            "Epoch 222/300\n",
            "93/93 [==============================] - 3s 28ms/step - loss: 0.3756 - accuracy: 0.8707 - val_loss: 0.2382 - val_accuracy: 0.9353\n",
            "Epoch 223/300\n",
            "93/93 [==============================] - 3s 36ms/step - loss: 0.4081 - accuracy: 0.8527 - val_loss: 0.2328 - val_accuracy: 0.9366\n",
            "Epoch 224/300\n",
            "93/93 [==============================] - 3s 27ms/step - loss: 0.3878 - accuracy: 0.8627 - val_loss: 0.2544 - val_accuracy: 0.9185\n",
            "Epoch 225/300\n",
            "93/93 [==============================] - 3s 31ms/step - loss: 0.3950 - accuracy: 0.8567 - val_loss: 0.2474 - val_accuracy: 0.9392\n",
            "Epoch 226/300\n",
            "93/93 [==============================] - 2s 26ms/step - loss: 0.3856 - accuracy: 0.8513 - val_loss: 0.2468 - val_accuracy: 0.9301\n",
            "Epoch 227/300\n",
            "93/93 [==============================] - 2s 26ms/step - loss: 0.4046 - accuracy: 0.8507 - val_loss: 0.2471 - val_accuracy: 0.9301\n",
            "Epoch 228/300\n",
            "93/93 [==============================] - 3s 27ms/step - loss: 0.3654 - accuracy: 0.8740 - val_loss: 0.2490 - val_accuracy: 0.9224\n",
            "Epoch 229/300\n",
            "93/93 [==============================] - 3s 29ms/step - loss: 0.3898 - accuracy: 0.8607 - val_loss: 0.2446 - val_accuracy: 0.9211\n",
            "Epoch 230/300\n",
            "93/93 [==============================] - 3s 27ms/step - loss: 0.3836 - accuracy: 0.8580 - val_loss: 0.2354 - val_accuracy: 0.9327\n",
            "Epoch 231/300\n",
            "93/93 [==============================] - 3s 27ms/step - loss: 0.4064 - accuracy: 0.8613 - val_loss: 0.2412 - val_accuracy: 0.9237\n",
            "Epoch 232/300\n",
            "93/93 [==============================] - 3s 27ms/step - loss: 0.4256 - accuracy: 0.8520 - val_loss: 0.2478 - val_accuracy: 0.9276\n",
            "Epoch 233/300\n",
            "93/93 [==============================] - 3s 27ms/step - loss: 0.4061 - accuracy: 0.8540 - val_loss: 0.2620 - val_accuracy: 0.9301\n",
            "Epoch 234/300\n",
            "93/93 [==============================] - 3s 36ms/step - loss: 0.3851 - accuracy: 0.8613 - val_loss: 0.2559 - val_accuracy: 0.9172\n",
            "Epoch 235/300\n",
            "93/93 [==============================] - 3s 27ms/step - loss: 0.4156 - accuracy: 0.8640 - val_loss: 0.2816 - val_accuracy: 0.9211\n",
            "Epoch 236/300\n",
            "93/93 [==============================] - 3s 28ms/step - loss: 0.4259 - accuracy: 0.8433 - val_loss: 0.2356 - val_accuracy: 0.9314\n",
            "Epoch 237/300\n",
            "93/93 [==============================] - 3s 27ms/step - loss: 0.3969 - accuracy: 0.8553 - val_loss: 0.2416 - val_accuracy: 0.9301\n",
            "Epoch 238/300\n",
            "93/93 [==============================] - 2s 26ms/step - loss: 0.4062 - accuracy: 0.8607 - val_loss: 0.2570 - val_accuracy: 0.9340\n",
            "Epoch 239/300\n",
            "93/93 [==============================] - 3s 27ms/step - loss: 0.3828 - accuracy: 0.8653 - val_loss: 0.2577 - val_accuracy: 0.9146\n",
            "Epoch 240/300\n",
            "93/93 [==============================] - 3s 35ms/step - loss: 0.3878 - accuracy: 0.8547 - val_loss: 0.2374 - val_accuracy: 0.9379\n",
            "Epoch 241/300\n",
            "93/93 [==============================] - 3s 29ms/step - loss: 0.4281 - accuracy: 0.8420 - val_loss: 0.2302 - val_accuracy: 0.9366\n",
            "Epoch 242/300\n",
            "93/93 [==============================] - 3s 28ms/step - loss: 0.3836 - accuracy: 0.8653 - val_loss: 0.2283 - val_accuracy: 0.9340\n",
            "Epoch 243/300\n",
            "93/93 [==============================] - 2s 25ms/step - loss: 0.3713 - accuracy: 0.8600 - val_loss: 0.2391 - val_accuracy: 0.9276\n",
            "Epoch 244/300\n",
            "93/93 [==============================] - 3s 36ms/step - loss: 0.3943 - accuracy: 0.8560 - val_loss: 0.2409 - val_accuracy: 0.9276\n",
            "Epoch 245/300\n",
            "93/93 [==============================] - 2s 26ms/step - loss: 0.4204 - accuracy: 0.8427 - val_loss: 0.2420 - val_accuracy: 0.9198\n",
            "Epoch 246/300\n",
            "93/93 [==============================] - 3s 27ms/step - loss: 0.3992 - accuracy: 0.8533 - val_loss: 0.2376 - val_accuracy: 0.9327\n",
            "Epoch 247/300\n",
            "93/93 [==============================] - 3s 27ms/step - loss: 0.4111 - accuracy: 0.8600 - val_loss: 0.2475 - val_accuracy: 0.9314\n",
            "Epoch 248/300\n",
            "93/93 [==============================] - 3s 35ms/step - loss: 0.3862 - accuracy: 0.8633 - val_loss: 0.2406 - val_accuracy: 0.9314\n",
            "Epoch 249/300\n",
            "93/93 [==============================] - 2s 26ms/step - loss: 0.3750 - accuracy: 0.8727 - val_loss: 0.2598 - val_accuracy: 0.9250\n",
            "Epoch 250/300\n",
            "93/93 [==============================] - 3s 27ms/step - loss: 0.4177 - accuracy: 0.8427 - val_loss: 0.2463 - val_accuracy: 0.9276\n",
            "Epoch 251/300\n",
            "93/93 [==============================] - 2s 26ms/step - loss: 0.3659 - accuracy: 0.8660 - val_loss: 0.2407 - val_accuracy: 0.9288\n",
            "Epoch 252/300\n",
            "93/93 [==============================] - 3s 35ms/step - loss: 0.3857 - accuracy: 0.8533 - val_loss: 0.2496 - val_accuracy: 0.9185\n",
            "Epoch 253/300\n",
            "93/93 [==============================] - 3s 28ms/step - loss: 0.3798 - accuracy: 0.8593 - val_loss: 0.2449 - val_accuracy: 0.9301\n",
            "Epoch 254/300\n",
            "93/93 [==============================] - 2s 26ms/step - loss: 0.3897 - accuracy: 0.8613 - val_loss: 0.2362 - val_accuracy: 0.9340\n",
            "Epoch 255/300\n",
            "93/93 [==============================] - 3s 35ms/step - loss: 0.3875 - accuracy: 0.8647 - val_loss: 0.2442 - val_accuracy: 0.9314\n",
            "Epoch 256/300\n",
            "93/93 [==============================] - 2s 26ms/step - loss: 0.3602 - accuracy: 0.8607 - val_loss: 0.2348 - val_accuracy: 0.9301\n",
            "Epoch 257/300\n",
            "93/93 [==============================] - 3s 28ms/step - loss: 0.4068 - accuracy: 0.8513 - val_loss: 0.2299 - val_accuracy: 0.9366\n",
            "Epoch 258/300\n",
            "93/93 [==============================] - 2s 26ms/step - loss: 0.4000 - accuracy: 0.8573 - val_loss: 0.2416 - val_accuracy: 0.9288\n",
            "Epoch 259/300\n",
            "93/93 [==============================] - 3s 35ms/step - loss: 0.3828 - accuracy: 0.8580 - val_loss: 0.2337 - val_accuracy: 0.9301\n",
            "Epoch 260/300\n",
            "93/93 [==============================] - 3s 27ms/step - loss: 0.3482 - accuracy: 0.8680 - val_loss: 0.2399 - val_accuracy: 0.9314\n",
            "Epoch 261/300\n",
            "93/93 [==============================] - 3s 27ms/step - loss: 0.3802 - accuracy: 0.8593 - val_loss: 0.2503 - val_accuracy: 0.9301\n",
            "Epoch 262/300\n",
            "93/93 [==============================] - 3s 36ms/step - loss: 0.3982 - accuracy: 0.8547 - val_loss: 0.2465 - val_accuracy: 0.9327\n",
            "Epoch 263/300\n",
            "93/93 [==============================] - 3s 28ms/step - loss: 0.3780 - accuracy: 0.8607 - val_loss: 0.2579 - val_accuracy: 0.9211\n",
            "Epoch 264/300\n",
            "93/93 [==============================] - 3s 27ms/step - loss: 0.3842 - accuracy: 0.8620 - val_loss: 0.2463 - val_accuracy: 0.9314\n",
            "Epoch 265/300\n",
            "93/93 [==============================] - 4s 41ms/step - loss: 0.3938 - accuracy: 0.8593 - val_loss: 0.2346 - val_accuracy: 0.9263\n",
            "Epoch 266/300\n",
            "93/93 [==============================] - 3s 27ms/step - loss: 0.4344 - accuracy: 0.8467 - val_loss: 0.2499 - val_accuracy: 0.9353\n",
            "Epoch 267/300\n",
            "93/93 [==============================] - 3s 27ms/step - loss: 0.3848 - accuracy: 0.8607 - val_loss: 0.2543 - val_accuracy: 0.9340\n",
            "Epoch 268/300\n",
            "93/93 [==============================] - 3s 28ms/step - loss: 0.3624 - accuracy: 0.8640 - val_loss: 0.2424 - val_accuracy: 0.9327\n",
            "Epoch 269/300\n",
            "93/93 [==============================] - 3s 33ms/step - loss: 0.4108 - accuracy: 0.8520 - val_loss: 0.2469 - val_accuracy: 0.9276\n",
            "Epoch 270/300\n",
            "93/93 [==============================] - 3s 28ms/step - loss: 0.3606 - accuracy: 0.8687 - val_loss: 0.2456 - val_accuracy: 0.9250\n",
            "Epoch 271/300\n",
            "93/93 [==============================] - 3s 28ms/step - loss: 0.4041 - accuracy: 0.8600 - val_loss: 0.2689 - val_accuracy: 0.9263\n",
            "Epoch 272/300\n",
            "93/93 [==============================] - 4s 39ms/step - loss: 0.3954 - accuracy: 0.8593 - val_loss: 0.2397 - val_accuracy: 0.9366\n",
            "Epoch 273/300\n",
            "93/93 [==============================] - 3s 28ms/step - loss: 0.3527 - accuracy: 0.8680 - val_loss: 0.2549 - val_accuracy: 0.9327\n",
            "Epoch 274/300\n",
            "93/93 [==============================] - 2s 26ms/step - loss: 0.3712 - accuracy: 0.8687 - val_loss: 0.2565 - val_accuracy: 0.9366\n",
            "Epoch 275/300\n",
            "93/93 [==============================] - 3s 27ms/step - loss: 0.3966 - accuracy: 0.8507 - val_loss: 0.2447 - val_accuracy: 0.9366\n",
            "Epoch 276/300\n",
            "93/93 [==============================] - 3s 27ms/step - loss: 0.3923 - accuracy: 0.8513 - val_loss: 0.2433 - val_accuracy: 0.9379\n",
            "Epoch 277/300\n",
            "93/93 [==============================] - 4s 41ms/step - loss: 0.3834 - accuracy: 0.8547 - val_loss: 0.2438 - val_accuracy: 0.9301\n",
            "Epoch 278/300\n",
            "93/93 [==============================] - 3s 28ms/step - loss: 0.3771 - accuracy: 0.8627 - val_loss: 0.2517 - val_accuracy: 0.9185\n",
            "Epoch 279/300\n",
            "93/93 [==============================] - 3s 27ms/step - loss: 0.4030 - accuracy: 0.8540 - val_loss: 0.2435 - val_accuracy: 0.9288\n",
            "Epoch 280/300\n",
            "93/93 [==============================] - 4s 42ms/step - loss: 0.3933 - accuracy: 0.8693 - val_loss: 0.2519 - val_accuracy: 0.9237\n",
            "Epoch 281/300\n",
            "93/93 [==============================] - 3s 27ms/step - loss: 0.4017 - accuracy: 0.8593 - val_loss: 0.2481 - val_accuracy: 0.9276\n",
            "Epoch 282/300\n",
            "93/93 [==============================] - 3s 27ms/step - loss: 0.3890 - accuracy: 0.8600 - val_loss: 0.2354 - val_accuracy: 0.9353\n",
            "Epoch 283/300\n",
            "93/93 [==============================] - 3s 36ms/step - loss: 0.4101 - accuracy: 0.8520 - val_loss: 0.2391 - val_accuracy: 0.9314\n",
            "Epoch 284/300\n",
            "93/93 [==============================] - 3s 28ms/step - loss: 0.3680 - accuracy: 0.8713 - val_loss: 0.2503 - val_accuracy: 0.9276\n",
            "Epoch 285/300\n",
            "93/93 [==============================] - 3s 27ms/step - loss: 0.3509 - accuracy: 0.8753 - val_loss: 0.2436 - val_accuracy: 0.9340\n",
            "Epoch 286/300\n",
            "93/93 [==============================] - 3s 31ms/step - loss: 0.4067 - accuracy: 0.8553 - val_loss: 0.2322 - val_accuracy: 0.9314\n",
            "Epoch 287/300\n",
            "93/93 [==============================] - 3s 27ms/step - loss: 0.3821 - accuracy: 0.8527 - val_loss: 0.2352 - val_accuracy: 0.9288\n",
            "Epoch 288/300\n",
            "93/93 [==============================] - 2s 26ms/step - loss: 0.3915 - accuracy: 0.8480 - val_loss: 0.2474 - val_accuracy: 0.9237\n",
            "Epoch 289/300\n",
            "93/93 [==============================] - 3s 27ms/step - loss: 0.3645 - accuracy: 0.8653 - val_loss: 0.2563 - val_accuracy: 0.9288\n",
            "Epoch 290/300\n",
            "93/93 [==============================] - 3s 28ms/step - loss: 0.4249 - accuracy: 0.8407 - val_loss: 0.2721 - val_accuracy: 0.9069\n",
            "Epoch 291/300\n",
            "93/93 [==============================] - 2s 26ms/step - loss: 0.3784 - accuracy: 0.8627 - val_loss: 0.2345 - val_accuracy: 0.9288\n",
            "Epoch 292/300\n",
            "93/93 [==============================] - 3s 26ms/step - loss: 0.3596 - accuracy: 0.8687 - val_loss: 0.2499 - val_accuracy: 0.9366\n",
            "Epoch 293/300\n",
            "93/93 [==============================] - 4s 42ms/step - loss: 0.3653 - accuracy: 0.8673 - val_loss: 0.2502 - val_accuracy: 0.9340\n",
            "Epoch 294/300\n",
            "93/93 [==============================] - 3s 28ms/step - loss: 0.3739 - accuracy: 0.8693 - val_loss: 0.2459 - val_accuracy: 0.9353\n",
            "Epoch 295/300\n",
            "93/93 [==============================] - 3s 27ms/step - loss: 0.4030 - accuracy: 0.8580 - val_loss: 0.2503 - val_accuracy: 0.9340\n",
            "Epoch 296/300\n",
            "93/93 [==============================] - 3s 27ms/step - loss: 0.3489 - accuracy: 0.8667 - val_loss: 0.2386 - val_accuracy: 0.9392\n",
            "Epoch 297/300\n",
            "93/93 [==============================] - 4s 41ms/step - loss: 0.3853 - accuracy: 0.8613 - val_loss: 0.2503 - val_accuracy: 0.9276\n",
            "Epoch 298/300\n",
            "93/93 [==============================] - 3s 28ms/step - loss: 0.3550 - accuracy: 0.8713 - val_loss: 0.2388 - val_accuracy: 0.9314\n",
            "Epoch 299/300\n",
            "93/93 [==============================] - 3s 28ms/step - loss: 0.3678 - accuracy: 0.8647 - val_loss: 0.2495 - val_accuracy: 0.9392\n",
            "Epoch 300/300\n",
            "93/93 [==============================] - 4s 40ms/step - loss: 0.3997 - accuracy: 0.8500 - val_loss: 0.2433 - val_accuracy: 0.9301\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAjcAAAHHCAYAAABDUnkqAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjcuMSwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/bCgiHAAAACXBIWXMAAA9hAAAPYQGoP6dpAACTO0lEQVR4nOzdd3xT1fsH8E+SNt17L1pGKWVvLHuDKCKoDFHAgYrgQlRQEcdXcf9UcC9QURAUREGUDULZexUolBboLt07ub8/Tu5N0gFtSRpaPu/Xq6+mNzc3J7dp75PnPOcclSRJEoiIiIgaCbWtG0BERERkSQxuiIiIqFFhcENERESNCoMbIiIialQY3BAREVGjwuCGiIiIGhUGN0RERNSoMLghIiKiRoXBDRERETUqDG6IyGISEhKgUqmwaNGiWj92y5YtUKlU2LJli8XbRUQ3FwY3RERE1KgwuCEiIqJGhcENEZEVFRQU2LoJRDcdBjdEjcirr74KlUqF06dP47777oOHhwf8/Pwwd+5cSJKEpKQkjBo1Cu7u7ggMDMQHH3xQ6RhpaWl46KGHEBAQAEdHR3To0AGLFy+utF92djamTJkCDw8PeHp6YvLkycjOzq6yXadOncLdd98Nb29vODo6omvXrli9enWdXuOFCxfw+OOPIyoqCk5OTvDx8cE999yDhISEKtv4zDPPICIiAg4ODggNDcWkSZOQkZGh7FNcXIxXX30VLVu2hKOjI4KCgjBmzBjEx8cDqL4WqKr6oilTpsDV1RXx8fEYMWIE3NzcMHHiRADA9u3bcc8996BJkyZwcHBAWFgYnnnmGRQVFVV5vsaOHQs/Pz84OTkhKioKL730EgBg8+bNUKlUWLlyZaXH/fzzz1CpVIiNja3taSVqVOxs3QAisrxx48YhOjoab7/9NtasWYP//e9/8Pb2xpdffomBAwfinXfewZIlSzBr1ix069YNffv2BQAUFRWhf//+OHv2LGbMmIGmTZti+fLlmDJlCrKzs/HUU08BACRJwqhRo/Dff//hscceQ3R0NFauXInJkydXasvx48fRq1cvhISEYPbs2XBxccGvv/6KO++8E7/99htGjx5dq9e2d+9e7Ny5E+PHj0doaCgSEhLw+eefo3///jhx4gScnZ0BAPn5+ejTpw9OnjyJBx98EJ07d0ZGRgZWr16NixcvwtfXFzqdDrfffjs2btyI8ePH46mnnkJeXh7Wr1+PY8eOoXnz5rU+9+Xl5Rg2bBh69+6N999/X2nP8uXLUVhYiGnTpsHHxwd79uzBggULcPHiRSxfvlx5/JEjR9CnTx/Y29vjkUceQUREBOLj4/Hnn3/izTffRP/+/REWFoYlS5ZUOndLlixB8+bNERMTU+t2EzUqEhE1GvPmzZMASI888oiyrby8XAoNDZVUKpX09ttvK9uvXLkiOTk5SZMnT1a2ffTRRxIA6aefflK2lZaWSjExMZKrq6uUm5srSZIkrVq1SgIgvfvuu2bP06dPHwmA9P333yvbBw0aJLVr104qLi5Wtun1eqlnz55SZGSksm3z5s0SAGnz5s1XfY2FhYWVtsXGxkoApB9++EHZ9sorr0gApN9//73S/nq9XpIkSfruu+8kANKHH35Y7T7Vtev8+fOVXuvkyZMlANLs2bNr1O758+dLKpVKunDhgrKtb9++kpubm9k20/ZIkiTNmTNHcnBwkLKzs5VtaWlpkp2dnTRv3rxKz0N0s2G3FFEj9PDDDyu3NRoNunbtCkmS8NBDDynbPT09ERUVhXPnzinb1q5di8DAQEyYMEHZZm9vjyeffBL5+fnYunWrsp+dnR2mTZtm9jxPPPGEWTuysrKwadMmjB07Fnl5ecjIyEBGRgYyMzMxbNgwnDlzBpcuXarVa3NyclJul5WVITMzEy1atICnpycOHDig3Pfbb7+hQ4cOVWaGVCqVso+vr2+ldpvuUxem56WqdhcUFCAjIwM9e/aEJEk4ePAgACA9PR3btm3Dgw8+iCZNmlTbnkmTJqGkpAQrVqxQti1btgzl5eW477776txuosaCwQ1RI1Txwujh4QFHR0f4+vpW2n7lyhXl5wsXLiAyMhJqtfm/hujoaOV++XtQUBBcXV3N9ouKijL7+ezZs5AkCXPnzoWfn5/Z17x58wCIGp/aKCoqwiuvvIKwsDA4ODjA19cXfn5+yM7ORk5OjrJffHw82rZte9VjxcfHIyoqCnZ2luuht7OzQ2hoaKXtiYmJmDJlCry9veHq6go/Pz/069cPAJR2y4HmtdrdqlUrdOvWDUuWLFG2LVmyBLfccgtatGhhqZdC1GCx5oaoEdJoNDXaBoj6GWvR6/UAgFmzZmHYsGFV7lPbi/ETTzyB77//Hk8//TRiYmLg4eEBlUqF8ePHK89nSdVlcHQ6XZXbHRwcKgWHOp0OQ4YMQVZWFl544QW0atUKLi4uuHTpEqZMmVKndk+aNAlPPfUULl68iJKSEuzatQsLFy6s9XGIGiMGN0SkCA8Px5EjR6DX680u0KdOnVLul79v3LgR+fn5ZtmbuLg4s+M1a9YMgOjaGjx4sEXauGLFCkyePNlspFdxcXGlkVrNmzfHsWPHrnqs5s2bY/fu3SgrK4O9vX2V+3h5eQFApePLWayaOHr0KE6fPo3Fixdj0qRJyvb169eb7Sefr2u1GwDGjx+PmTNn4pdffkFRURHs7e0xbty4GreJqDFjtxQRKUaMGIGUlBQsW7ZM2VZeXo4FCxbA1dVV6UYZMWIEysvL8fnnnyv76XQ6LFiwwOx4/v7+6N+/P7788kskJydXer709PRat1Gj0VTKNi1YsKBSJuWuu+7C4cOHqxwyLT/+rrvuQkZGRpUZD3mf8PBwaDQabNu2zez+zz77rFZtNj2mfPvjjz8228/Pzw99+/bFd999h8TExCrbI/P19cWtt96Kn376CUuWLMHw4cMrdTsS3ayYuSEixSOPPIIvv/wSU6ZMwf79+xEREYEVK1Zgx44d+Oijj+Dm5gYAGDlyJHr16oXZs2cjISEBrVu3xu+//25W8yL79NNP0bt3b7Rr1w5Tp05Fs2bNkJqaitjYWFy8eBGHDx+uVRtvv/12/Pjjj/Dw8EDr1q0RGxuLDRs2wMfHx2y/5557DitWrMA999yDBx98EF26dEFWVhZWr16NL774Ah06dMCkSZPwww8/YObMmdizZw/69OmDgoICbNiwAY8//jhGjRoFDw8P3HPPPViwYAFUKhWaN2+Ov/76q1a1Qq1atULz5s0xa9YsXLp0Ce7u7vjtt9/M6p1kn3zyCXr37o3OnTvjkUceQdOmTZGQkIA1a9bg0KFDZvtOmjQJd999NwDgjTfeqNV5JGrUbDVMi4gsTx4Knp6ebrZ98uTJkouLS6X9+/XrJ7Vp08ZsW2pqqvTAAw9Ivr6+klarldq1a2c23FmWmZkp3X///ZK7u7vk4eEh3X///dLBgwcrDY+WJEmKj4+XJk2aJAUGBkr29vZSSEiIdPvtt0srVqxQ9qnpUPArV64o7XN1dZWGDRsmnTp1SgoPDzcb1i63ccaMGVJISIik1Wql0NBQafLkyVJGRoayT2FhofTSSy9JTZs2lezt7aXAwEDp7rvvluLj45V90tPTpbvuuktydnaWvLy8pEcffVQ6duxYlUPBqzrPkiRJJ06ckAYPHiy5urpKvr6+0tSpU6XDhw9Xeb6OHTsmjR49WvL09JQcHR2lqKgoae7cuZWOWVJSInl5eUkeHh5SUVHRVc8b0c1EJUlWrCYkIiKrKS8vR3BwMEaOHIlvv/3W1s0humGw5oaIqIFatWoV0tPTzYqUiQhg5oaIqIHZvXs3jhw5gjfeeAO+vr5mkxcSETM3REQNzueff45p06bB398fP/zwg62bQ3TDYeaGiIiIGhVmboiIiKhRYXBDREREjcpNN4mfXq/H5cuX4ebmdl2r/hIREVH9kSQJeXl5CA4OrrR+W0U3XXBz+fJlhIWF2boZREREVAdJSUkIDQ296j43XXAjTx+flJQEd3d3G7eGiIiIaiI3NxdhYWHKdfxqbrrgRu6Kcnd3Z3BDRETUwNSkpIQFxURERNSoMLghIiKiRoXBDRERETUqDG6IiIioUWFwQ0RERI0KgxsiIiJqVBjcEBERUaPC4IaIiIgaFQY3RERE1KgwuCEiIqJGhcENERERNSoMboiIiKhRYXBDRNTYlBUBkmTrVhDZDIMbIqLGJD0O+L+2wI93Vr9PacG1j1NWZLEm1VhpAbDxDSDjjOWPLUlAWbHlj3sjKMmzzXNuehO4tN+47fx2YNv7gK68/ttTAYMbIqLGorQA+OFOoDADOLcFKMmvvE/sZ8BbIcDpf6o/ztZ3gbeCgRN/GI5baI3WimDj5F/A+ldEQLN/MbD9feC3hyyfefrrGeDNAODbYcDxldbJbOn1wM6FwMk/a//Yqn5XpnIuAv++DGTGA1nngb9fAFZOAxZ2B+aHit/r1RTn1r5NgHhPbXkbiN9kPI5eB/wxHdj2LvDHDMP2HGDZfcCmN4BDP9XtuSzIztYNIKIGJPmw+Mfd9znAzsHWrbnxlRYAdk6Aup4+R254Dci7bPw5/RQQ2hW4sBM4vQ7o9TSQsB2AJH6PLYdVPsbpf4HNb4rbh5eKi+7qGcCQN4CeM67dhrSTwKElQNcHAe9mV9938Ujg4h5xOzMesHcWt5MPA0l7gCY9zPfX6wBdGWDveO12AMDZjeI4fZ8DTv0ltiXtEl/RI4HRXwFaZ+P+kiQyVqbbrubSfuD4KqDbQ4BXhHjd/74EqO2BJw8Ank0qP6a8FLDTmm/b/SXw9/OAbxTQehTQZjQQ0Np8nw2vAUd/BQ4vA9QaIC/Z/P4tbwMd7wWcPCs/5/FVwPLJQMf7gDs+EY8/+ReQeQboNhVwcK3+NW6ZD+xcIG77RIrHOHoCxdliW9oJIO0UcHK1cdvuL4HOkwGVqvrjWhmDG6Ibzck/xT/6mBmAxop/oqUFgJ2j+EdXE+WlwNKJQE4S4BEKdJlicpx6vIBXJEnAvu8AtR3QZbJ1nkNXDuhKAK3LVfYpA+I3A+e3Aq3vFOf1u+FAmzuBMV9V/5jYhYB3c3Gxre5iUFoAaByu/n7IThTnwVTqMZF92fmJ+NnFH8g1BD+m3QmAOI+HfxEZAVleCrDvW0DSi4t2cCcgolf1bZAv0gBQmAXcaZJNyLkk2tdhPOAbKbo15MAGAJJ2A84+xp93fgJ4vAO4hxjPy9J7xX6PbAW8wsU2XTmgUhvff+e3A2c3iIv8hlfFNs8mQEG6uH3L48Cer8Tfmfc7wJDXjM+5bjaw+wsgpIsINJy8gAFzAAe3yq/1/Hbg57FAWSGw91ug5xPA3q/FffoyYPuHwKBXRFCVekIc8/AvIqM2YSng0xw4sBjo+pDoygGAjDiRDdn2LjDsLSBmuthenCOCBwAoSBPf/VoBHSaI87P9AyD9JLDrc9FeU7py43k49BOgKxVB6q+TAEknfid3LACa9a/8GvPTgD3fGH/ONHQXykGMkxdQdAXY/714bbK0EyKIbtq38jHriUqSbq6qs9zcXHh4eCAnJwfu7u62bg6RubIi4J2mQHkR0GaM+Ofo4nf1T1aA+HS97T1xgQztevX9tC7Axb3A4juA6NuBu76pfn9T+74TqX0AaD8eGPOluJgvux8I6wbcv7L6x5YWGgIpCwZAe78F8lOBggxxAQaAJw6Ii4Yl6fXA4tuBxFggvJcI7Jx9gKgRQHhP44X3l3uBuDXitnsIENgeOP23+Pne5SIoqBgE7v0GWPOsuN1yuPgU3ay/CGLkoDH9FPDtUCCwHTDpD/Gp/+xG8Tvs9ZQIhA//Irp1zvwjLiiB7UXQFNwZuHzA+HzdHwFOrAbyUwCogDlJ4sKdcwn48yng7Hqxn7Ov6NqydxEXsNyLYrtbEDB9N+DoUfk8ZcYDC7oAMFxSfKOAGYbgJTsRWHQ7kH0B8AgDHt0GXDkPfD0QcPAQAYK+rOrz32MacOvboivmk45i2y3TgYEviwzTrs/FxXrI6yJY/KCVaLup8F7AhR2ARxPgmaPAqTUiUFLbAZP/BPyjRdbo3eZAaYX6lW4PA7e+J37/Z9cD4b2BgDbitZYVGC/wMtdAcX5Vhg8Nkq7ya2o+SAS/Z/41Pt4tGBg8Dzj2m9iu0gBT/hLvsX3fA389Dfi0EIGargy4+3vA1U8c7/hKYPkUwMEdmHlSZOkK0oEej4ls0h/TxX1lhYC+XLzWskIRFEp6k9f5rnhsZrwI1v59WbyPQroCt/8fkHEaaBIjgpeyQvEeXTXN+Lr8Won7938PRN0GTPi56t9pHdXm+s3MDZGtFWWLi4VKJboPyg2FnMd/F1+OnuKC4hZY/TG2vQvs+BjYvwh4bLv4h2nvYryQlhUDW98GdnwCRPQWAUF5EXB0OXDLNHEh0rpUnzkoLwG2fWD8OTFWfHL9ZTxQXiz647POVd0NkR4HfH+reI4H/67++HYO4hPuoSXiou3qX/3rPbgEWDOz8vbjK4G+s6p/nFm7TotPzh3Gi8DBVGa8oS5DL4KlCzvE9oTtxn1iF4puniGviayWHBho3YDcS+JL9uskcb47TwJGfCB+F+E9RaZDdnqd+GozBhj6hqilCGgNQCUuuIk7gdVPiIvTsRXiMSqNyAwkHzIep/+LInAAzAMbALiSYPzkD0lkb7ITgX9eAkpyAY0W6D9HZDfmh4qLd5mh+NjFT3SFbJ4vgo2K9nwljhnYDkg5Kj7llxYA57cBfz5tCKggMn+rponsFgAEtQdK84HLB8XPzr5A1K3iIl9WKLIhtzwGnFhlfK6DP4rfRcoR8fO+RcDAuaKOqDBD/A4c3UXmQV9m/P35txLfW90GtLxVBJ/f3yqCnL7Pi/PsGiACpysJIiOy91vRtSU/155vxPuzrEAEkQ/+I7JAR5YBmWeBu78DNv0POLdZ7B/QDgjtAlyIFe/phO0ieyMHgXJg1O0h8V5sPw5Y+ag43tJ7gUHzgD2GjFDnSeK5K4oeJbrFriSI9/S/L4v3bn6q6MYCgH7Pi32WTxHn1d4ZeGSL+L3t/UZ8FaQb66wieoluSfmxQe3FFwB4hBjani3eM7pSEfDf9S2gsRf/h/RlImtkzezzVTBzQ2RJkiT+oV8tUABEH/Xx38U/kvRT4h/rwJfERSZ2IRDUQaT185LFxUy+vyr56cDH7cU/LEB8QivJFZ/EbjMEJEvuEZ8Gq+IZLi7EYT2AUQuBAz+K7EGzfsZ94jcBP44WF56iLPGP0y1Y1HeoNOLT6YCXxb5qtbhw+TQX5+LrQSJlDgDPHAeOrhB1GQ6u4kJ67DfRldH6TvFPvzhb/AMf8nrV7U09IT7xlxeJzEReMhDUUVyoAtoB0/6r/rzL9i8G1j4nupoc3IH7fhfZJ0AUhP77MpSLj/z6uj0sPpmWFYmL99FfxQVi1mkR2H3ZVwSpA14G/n5OPDbsFhE8yHUwGi0w4CVgwzxjW7RuIuu1/3sR2Dl6AiPeA36fatxH4yDaWpH8qV9tJ36PEb1FTUXyYdEeWbuxor1yRkbmHmrMyoR0Fd1IflHi588Mn9ABceyRH4n3gEoNTN0MBHc0Hqc4F/iwtQgO7vtdZArykkXWZffnYh+/VqJuZ9lEcTGMHCrek10fElmMPYauu2YDgEmrxO0f7hRBQudJ4jUlHzbPNrj4iSxGcTYwabUIFuPWiKzD0P8BF/cD3ww0trPnkyJwBMTv5Yc7RbClKzXu03myOIcAsOIhYyCpdRXtLM4x/j5GvA90N/k9yQoygbi1IoCtmEn8ord4/wDGvyE7J+CZY4CLr9heki+yhXLAB4jnfPoo4BZQ+fkAUZPz34dVv1e8mwOP/SdqiU6tBTa+BvR+RgRTAHDoZ/MMDCDep5vfFOd7zqXq65A2vArErRMZ4MC2YlvOJWMAZEG1uX5ztBSRpez5GljQGZgfIi5SsoNLjJ+85J8/6wFsfUcENoD4hJh63DgiodfT4p/dXYbuln3fVT+MdefHIrDxixYXxhLDqIj9i0R9xfnt4iKi0Yp/+A6Gfwq9nhIX7uwLIoC6sAP4pJP4B/nPi+bPkWz41Nq0r0jJA+KfstbNWLOw9W1g8/+Aja+LlP3OBcDvjxgDG0DUIWyYBxxZKj4pbntfBHOA+GQu9+Wf3w5kJ4lRNAWZxseX5IvCyPIikdp/eCPw7ClxUVbbAalHrz2MuCRPdAXpSgAnb3G+fhwtPllveUfUlkACIvqIcybpxDkb+LK4kPV6UtTQ+ESK8358lfEiFNRRXIjdQ8XPfWaKroXRX4ouBV0psPkt8/Z0migCq9s+AKAS50DONMj6zxaZCfdQURQ6+U9xsZU/9bccLgpY5Yuyb5SxW0SlMdZHVeyukQObwa8CD/1rDGwAEYzIAtsBzQeKAFTSi2A57ZTx/n3ficDGN0rsF9RRbJcDm44TRZag5VBjHcYZQ6bLt6UIrEyfSzbA8D488KMhsNEA/WaLbS7+wJQ1QKvbxc/7F4luOUCcI0C8V9UmmQP/aONtzybinD15UBQAy6JuNd4e9qaolYm+A5i+R3SHAYbgQSW2V8XFB+h8f9VdpG1GG2/3ny2CsilrjIENIAL/B/8F+r0g3jfRdwCTV1cf2JgeVw5s3A3BhXdz8R6Ug5NWI0QmWA5sAFGI3LlCvdqRXw2Pb3b1AuvBrwLTdxkDG8AqgU1tsVuK6FpK8q9d83J2I7DWpDvkyDKg031Awg7gj8fFNr8oEYD8Yyj4a9ZfFAQeXyWyDr8/YvikrDIW97W6XfyTyr0kPkF2MvzTPvmnuLh0mmgs+BvyOuAZJgKRvV+Lmoy934ouJEBcdHs+Iepy0k8DkUNEvcWx38XFZ9t7xuAi9Zj4risT7ZFT8oHtRPpZ/uTZ6T6gw73A+nkiQALEherSPkP2AyJACOsh0vH7vxfbPJsYC2B1JSID4h8tMl7nt4lullXTDN1AKnFx3vGReN6M0+IT75ivjN1uzt7inJ3dII45fL7570dXDhxYJI7lFihS5l4RwLSdwM/jxPN8P9y4f/854sJzZoP4vfZ8QmRJZCqVuCBsfE186pWDguBOYiTP5NUiyJJHI/k0FzUjW98Wr1elATpOEN0dMYYRSPZOhvNyQdSEAOK+8F4ieFGrzbvc2twJHPzJ+HswZe8oLooZcSJ7YHrhAUT9U7khWO75pPgUX5F/NHDccDuog/h++/8BWfHi9/DDKOCJfcZ6L0AEfiqVOA9yvZFKI7pW7J3EzxG9xe9Jzoz5tRSZIVlge+PtsO4iY7bX8B5vPlCcA68I0W3iESqCkUM/iUwoIN5/cveTvaN4HfL71TRgk3mEir+Nfd+K89LUJGPpFghM3WT8ueMEMXoIkngdVws2qtNmtJgfxt5Z3HasJgNhpxXB3YAXq76/osB2IpDJiheZoEe3ib+lZv3F38e13PaB+J+QnSg+3MjFw/KHmQaGwQ3ZVlmx6KOt6Yid6yVJor6jpkNJd3wCrJ8LjP0RaG34lFacK+oc0k4C7e4W/zDlESIthojai6Q94p++acCz+S3APViktQPbARN/E/3R4T3FCBs5oAjpbPxnpLET2YINr4qvFoNF5mH5A+ICvX+RyGKEdBX/mFQq8c/czkFkOP77P5F50GiB3oYaFa8I8QWIIbJ9DV0orW4TF/nVT4if0+NE90/L4caLQ2B7cQHe+zUAlWibiw/QYpDIDnWbKrpUNv1PzFeicQDGLxHnPWG7sTth2HxRM5OfKn5uORy4xxD4fNJJdPPI9S2X9ouMg3whV2lEXYPpJ11AXATPbgB2fSYugpFDxPa8VDGqRa5LkbMKzQeJYOreX8X9CdtFJmrYm8ZRV5GDgacOVf3e6DBezOmRuFO0FzB21fg0r/ypvc2dIrgBRJfMqE8rH9MvSgQ38sieZgNEG6rSebI4J25B4n1RUVg3Edy0HiW6y+TuSvm45UVi+6BXqj5+xcwNIN6Xk1YDXw8Q9R17vhLvjdJ8ILSbCHRNzwMgfg+mQUBEhRE0vlHi78I9VHRlVSyIv+0DEYAlxop2qzVAh3HG+5sPMHbFOPuKrlVTQR1NgpsoVKnf86I+KXLo1bMUnk3Eeyt+I9Dunur3uxrvZoZMikv1gU1dqFRA+7Ei+Go7Rvx9tB1T88dr7MUHn+TD5tsD2la9/w2OwQ3ZTsYZ0f/c8V7xibA+/Peh6DbpeJ+4iFU1J4SsMEt0HQGiXqH1HaKG5Kv+4tMRAMR+KjIrOYmiEPHub4EFXUXR5h+Pi0yMk5cYLSRnUADgtg+NhXaeTURaevcXItsif5KX9XhM1KmkHhOTZKntjCNL5G6FAXPMa3xa3W7MjqjU4gJ2rVSxd1Px9e/LIgA7ulxctI79ZtwnqL34dBvWA2hyi/ECfscC0ZUSPUq0Y+DLYtSEe5D45Feca6xdcfAQF7z244xDlE1T9RG9jcECIDJRcnFut6kiS2F68ZRF3Wr8lL/iQWC4Yd6P//7PvOBWvt1ikPiudQYmLhfzfoT3rHlK3T1YXAxO/GEslg3uVP3+/tHGYtvqhqz7tjSvjfKNrP54Yd1FrY57qLgwVTTkDXGxbjXS0N4QIN0Q3Hg2AUa8W/2x5fbKTLMpzt6iaHnlIyIDIenEe2zE+8ZMmhxAAiIraCqogwgiS/NE15p7sHjP3P+7KHT3blq5LV7hxqHfFWldxGip0/+Iug/TdgPid3LwR5Edqm4ov1ug6DaridFfGt7r1XRJ1UR4z7o/9mp6zxTZm6jh1963OnKXpjzKy7/11fe/QbHmhupHWbGYgMvU6XUiNX7oZ8vOgFqSZ+hOqUCvA3Z9IW4f+kkMGf11suiiqcrOBeLiDojuJb1eDKPNihddM80HiX8AOYliZNIdC8Qn4Yje4jHyqIOBc8WEZoCoibn7O3FhMhXSWXSzPHmw8qcteyfgnsXiQnBxrwiSNA7GgCCsh2iLKY0dcP8qUbMz64zoVqkpuV4kQS7MlcSXa4AY7eHoLuozTAt+3QKBtncZAzaVSmQc5JS2o7sxIGl1m8gsdbpPXBQdPY1ZFqDyJ/uSHEOwYwiaqgpsZMPeEkN1S3JFcLlzgSjsBIwFz4D45x3Rx/izvRPQ/p7a1woM/Z/oAgDE6/Cs5gIsG7cEmLDMvK7DlGlmwd5ZDJu+muYDRbdOVZy9RdZGDjg8Qo33uQdd/biAuEi2GS2yMe7B5ve1u1vUHMkXwNs+NP+9uAWI32/kMJGVM6WxA8JjxG3fSGNQ7hd19Tl0rmbQK8C0HZUDG0Ccg5CuYgSYJbj6iSycreZ1uho7rXgfVzUvT03JXZoydkvRTaU4V/wBVTt0uFR0S7S6TRTrfdlXdGGY1kLIRZjlxWKUTKsR137ekjzxT1/uxrp8UAQR/WaLP8rsROCznqL25P5VYp+Tq8XQ3iY9REbF0UOk8tNPiSLWuL9FV0qXyYZuMq3oGjAdpluUJQpj5ULhHtOAfs+JItSiLNGvLX8qbNrH2P/v0QTodL+43aSHqJ+42hDn6vi2AB74W7QpYbuobeg8RWQOwntV/XuoqmukJjxCgLTjwMV95tsrDpeurZ5PiIJdeZZbvyiRsXJwN9ZjAOL8abTiyz1EdK0A4hPk1TJtgAiaJv0hUvPb3xeztpYZJsCLeVycu/NbRXBpiS4BzybifbDxdfM5b6pztQwEID41y3xaWPYCahrcuNUguFGrgXsWVXOfBrj1HTHvUb/nK9f8AFV3u8nkkVIhXa7djuvl4gtM3Wj952lMAlqLvzt7l2sH7DcoBjdUewk7gEUjxD+0OxaKuR3kOSh6zxT/+A7/LLp0LuwUn9z0ZWKU0JA3jJ/uLx8yHvP038bgpiRPjB7KThIZgUGviItW4i7gxzEikJjws8ikrHhIZFLcQ0TwtPtLke5OOwF82l10r8iFi/JQ0+g7RJYl+bCoDTm7XtTGRA4Vk6WVF4tPvWUFYiiv1lmMYjr4k7HAteMEcSz5E6gp04xAv+eMU62bdr3URVB74M4KF4y2d13fMasif0qvOKmaaddEXbQZXfkcVJWedwsUQY+do5izQw5uKk7FXx2NnSgGPvSzcQh2s34i+Bw0TxQq1yaTdS29nqk84qeuTLuhfKvJyNRVbYOba2kxCHj6SN0e2/VBkQls2ufa+1L9C2gj5nkKaH1jZqhqgMEN1Z7c3XLwJ+DKBfOJzbybiQtukmFm0rQTxuxBSY4o2vNtKQpL5boVQPSX6/WiG2jJ3WKKdVloN9HVs/wBEXDErRV985f2G49xfpsYeXTgR/Gz1s048iekixhZJE9G1uZOw4iOjqLW4uMOIlOz/hXRxQSIDI+Tt6ihOfKrCG52G7q0mvU3v1BU5NNCFBuW5Is2NTTuFV6bi784H01uqb82yN12pjUsTaoIJKujsReTom0yzGkid42EdjHOmmsparXIoFmCs7eYu6Ugvfri17oy7eKq2M1U39QaY4E+3XjajRX/57s/auuW1BmDG6o9+ZM0YAxsvJuJuohdX4jg5uJesb0w0xjoAGLEUMJ/xqJB9xDRxZWfCiQfFI9PMkzv7tVUFH9mnRMpbGVBQEn8LM/DAIhjHvpZBFDyvA7xm8WcGp5hwJHlwO8Piwu16VBPlUrULez/XhQNA2J0gK5MFFx6hBrn5ZD0onujz7NXPz8qVc2XNLgRVaw7mbxajIyJHFr/bTEtTA2rYeZG1uUBMa+OrrT6GpcbUVBHkU00fe2WYPp7vdps10Re4WLSvwaMwQ1d26k1oovolsfEkF7TYcFqO1FY6tsS+L82YiG8+M1iLhKZPMspYJySXL4/tKsoJj67XtTPyPff/b0IcpIPiQurPNIorIfYvnOBcU4YOwdR97Lpf2KfHo+KT6adTEZptL9H9L27BVYeWSIHN7LbPjTvAgnqaCge1ou1V6or4GwsTD/VO3qIIcFVFWrWB/9o0ZVl71L1CstX4+IDPLhODP23daaiNkZ+LOYJMi2ytgSfFqKQ2tX/+gpOiRoABjd0dXo98Pujoo6lyS3iH2Nhpvgn+dB68/li2o4Rk9f9+WTNjx/cSaxPcnY9cG6rYX4PleiCkOdAyTgjur8AMdnYL+ONAVOH8WK/+E0ia+Pibyzgraj5gKq3N+1rHPro7FN5ng2NnRimerMw7ZbyanrtIllrUmuqL2qtiaDrrBOyBY8Q68zw6hYoljUwXXmbqJFqmJVCVH+yLxhXyb241zgNv19U5Ynwej4phvZmJ1Y+joufWPcHAG7/yNj/36SncajhacPU6fJ03/JEcxf3isBD6yqGlroaUuo+LcQoJ9MC3t5PX30Srqo4eRoDmshh9Teh4I3KNMtR1Zwj1HCZLp9B1IgxuKHKSvJE9xNgXPsIEEGG0iVVxbDgwLZiETyZ6QynPpHA+J+A+34Duj4ghjVPXCG6f+R/tvKaKAGGSaPkIYjyqB15aGzfWeL5x/4g0uvyZGxuQcb5ZGqr19OAfxsxXPhmp3U2LjXgxeCGiBoedks1Zrs+F8On7Z2APrPEXCnXcmoNsHSi6IIa9alYYkB2cZ9YERmofs6TgS+LKvuCNBFoyMsS+LYwn/bfM0x8ASLwUdsZ1yaSp/t2CxJzncgr9soTS3Wfar4Sb1AHMSW8Z5j5fCm10WpEzebZuVm4h4pFGb2b2bolRES1xuCmsTq7AVg32/hzWREwdrEYQu3sU30dxc4FACRRwPv1QPNJtrLiDfPGoPo5T5w8xey12RdEkCIHNz5XmUbeTismL0szrNInT/etVovsjbyA29Wmom/Wr/r7qPa6TBZzzFi6qJWIqB6wW6oxKisG1hoWQ5TrUU7/IzI57zUXa+2YSj4CrHlWrF+UGCvqZjzDxTwx8RVm9izMEPUzFYtuTXk3NaxE62Ps3vC5RtbItA7A9LbpbK7XOgZZTvepYigohwwTUQPE4KYx2vedmBvGNRAY/7PoCiovEsvYA2L234JM4NIB4O/ZYoXfvd8AvxnqZZoPqly7EtrNeHvo/6pfgM6USiWWKQjpalxvqTpynY29s3mdh9yNBTC4ISKiGmFw0xjJSyH0eVasnyNPeS/pxffSfLGy9dcDgN2fi1oXV5NP6B3vFbP4yuwcjcOrw3uJ1Zxrqv8LYl2Xa63jE2aY/Ta0m/l03wxuiIiollhz09jkpxtnBG51m/jeZrSxKyr6DrGQZE6iKOKNvkMsERDRG/hjuqipkVdtDu5sWC4hUqwj5ewtZve1xrwn4THA5D8r1+bII6bcggEHV8s/LxHd0OJS8pCcU4T+UXVYcJaqVFBSjid+OYjU3GKseKwnnLSNb/oLBjeNzZl/AUhiBJE8EVhge6DjRBG43PWtWMk377KYWdh01NM935sfq8N4EdyEdhdzv1hq/ZzqyMscVNwW0hWIvt26z01E9a64TIdyvQRXh+ovRY/9tB/nMwqwekYvtA/1rL/GAcgqKMWDi/aiVwsfPDes1bUf0AAUl+nw0OK92HUuCwBwMPEKerbwva5jSpKEv4+loFMTTwR51HHEqoUxuGlsTv8tvrc0WUtHpQLu/Mz4c8WVpavTbaroFqrtmj6W5OQpurWIqMaKy3RQqQAHuxv7E/l93+xGQmYBNszsB09nbaX7c4rKcD5DLHi7/kRqvQc37/x9CoeSsnEoKbvOwY0kScgpKqvy9V1LUakOW0+nYWCrAGjtRHd9fkk5DidlIyO/BIOjA+BSTWCYmV+CjzeewaSYCLTwN2a9f92XpAQ2AHDscs5Vgxu9XsK/J1IQ6uUMvSTh9wOXMKpjMDo18VL2+ePQZTy97BCa+blg3VN9lbbaku1bQJZTmAWc3SRuW2KhQLUaaDlMBBhE1CDkFJYhZv5GTPx6NyR5Ms4bUE5RGfZduIKM/FLsOZ9V5T5nUvOU25tOpdX5uQpKyqHXG8+FJEkoLC2/6mPKdXqsPZas/JxbXIZDSdm4lF1Uq+desOksOr6+HmuPJl975wo+3XwWj/10AG+tFfONFZXqMPTDrZj4zW48tfQQPt18ttrHfrXtHH6IvYC5q46Zbd9/4QoAwNFeXP6PXco1u1+nl1Cu0ys/f7LpDB776QBuX/Af7li4A4t2JuD5FUdQWFqOt9aexN6ELPy8W8xKfy69AIt3JtT6dVoDg5uGTpKAnQuBbe8D/7wElBWI+WWCOti6ZURkITq9BJ2+ZoHKkUvZuFIoAocdZzPN7isq1WH/hSvYEpdmdryScp3Zxb8q8en5OJeeX/vGV+NsmvFYB5OyIUkSSsv1ZvucTjXuc/xyLlJzi6s9XlpeMd77R2RaTG0+lYbOb6zHrBWHlW0vrTqGzm+sx46zGdUeb/2JVOQVGwOgbafTMeazHbj/W/OgMSmrEJO+24N5fxwzCwoAkUH7bsd5AMBHG07XKNg8k5qH+X+fRGpuMbYb2rdsbxKyC0ux4sBFXM4xnoPYc5mVHi+fw53xmco+CYbsFwAcvSjmKpvYQ9QzHrucY/b4R37Yh1vmb0R2YSlOpeQqAZRWo4ZaBWjUKpxJy8dzy4/gq23nMPm7PdiTYAxOP954Bhn5Jdd8ndbG4KahO/kn8O9LwKY3gMM/i223fWDbxQ6JyGJSc4vR+Y31ePTH/TW6OJpeyBbHJii3C0rK0e+9zbjr852Y8v1eLN0rPm3HpeSh77ubcfuC/6oNoDLyS3DHgv8w+rOdKCrVAQDScovxwb9xSMmpPuAwdTYtD2O/jMX2M+nKz7KDiVcw4etd6PXOJiRlFSrbT5tkbgARqGQXluKjDadx4rIx47A5Lg3d39yITzfH4/U/jyvb41Ly8MQvB1FSrscfhy4rF92fdyeiuEyPid/sRnpe5QvxsUs5eNXkOADw5+HL0EsiO3EiOReSJGHzqTSM/mwHtp1Ox+LYC3h2+WGzc/jviVRkF5YZXks+tp5OV47/8qqjmPjNLqw+fFn5vW49nY7Rn+3El1vP4Z2/T+H4JRF4FJXp8GPsBXy7/RwAYErPCAAi4Csp1ynPt+ZIMqLm/o1PNp7BcZOg5dd9SSgt1yO3uAznDO+PCd3FDPHnMwqQXyKCuJzCMmw8lYaM/FIcTMzGa6tPoEwnYUjrAByaNwQH5w7FoFaisHuNIRNVaHg/DGzlj1aBbsgvKceGE6mVzml9Y3DTEEkSsOMTkalZN0ds0xj6czsalk4gukml5BTXuutAlpFfgsTMwmvvaEHXClg2nkxDTlEZNpxMxeY40TVTWFpudnE3dc4kuNl4MlU5F6dS8pBmciHfdjod5zMKMPGbXUjNLcGJ5Fyzrh9JkvDKH8fw2p/H8fuBiygo1SGnqAxHLmZDr5cw/ecDWLDpLJ7/7UiNXueCTWex53wWXvnjOPR6CWdMsjJ7zmdh17kspOeV4MWVR5VzcsYQAIX7iMVw3/83Dnd+ugMfbTiDZ5YdgiRJyC4sxZO/HFSOdSAxGzq9hLziMjz64z7lwq3TS/jz8GXkFJWZtWvGzweUgA0QGaqxX8YiNbcEkf6u6G2oR9l+xpjlWXXwEiZ+sxsPLNqLjPxSNPNzgZ1ahT8OXcbPe0TQmFVQip92XQAAeDnbAwA+2xKPoxdzMPqzHfhpVyJ2nM3Ek78cxMurjiG7sBSP/bhfae/KQ5dQbhIofbjhNBIyC+HuaIfnhkXBy9kepeV6nEwW56hMp8fb605CkoAP15+GXhJZFgD4fGs8Wr78N2YuOwQACPF0Qgt/NwR5OEKSgJPJ4r10IOmK8nwnknOx15CReXFENJy1dvBwtsfIDsaFdQPdHeHuKGp+7u3eBMPaiClF/rtKRqy+MLhpiJJ2A+vnArELgdyLgEcT4OmjwISlYsVtoptMam4xknOKkJZbjH7vbcZtn2w3u2DVhCRJGPdlLIZ9tA2pucVYvDMBL648qqT5y3V6/BibYBY4vbjyKPq9t7nKT/81cT6jALfM34g315yodp/d541dD++ui4NOL2HeH8cx4pPtWHXwUqX95cyNRq2CXoKyj5wRsdeIC96+hCt4759TyMgvhVYjLgU/mGR69iZcwQ+xF/D9jgR88O9pZfv+xCtYsvsC9iaIC+G20+nYGX/1i1l+STn+OZ6ivOYNJ1Nx1qSLyzRhtP1MBn7dlwTA2C31yu2tEenvioz8UiQYgs+41DzsOZ+Fz7bEI6+4HFEBbsprO5+Rj5dXHUNCZiFCPJ3w9GAxxcTvBy6ZZYxctBrsPp+FR37ch9JyPSRJwmt/nkBhqQ7dIrzw2+M90S7UA4AxQwEAX28/j53xmXCwU+Ph3k3x54zeeGG4KDhevDMBn2+JR5f/rcee81lQqYAv7usCrUaNPeezMOHrXSjTSeje1BuP9m0GlQpYsjsRv+xJQlGZDs18XeDhZK+sXXxr20C0D/VQfn6wd1O4ONihY5gnAJH1AkRRb1KWeVB/V+cQhHg6KY/dcFIEr+0Nr6lNsPh+zJAhOnDBGNz8eyIV5XoJbo52iDAElwAwKNofTvaiUP3hPk2x7NEYfDSuIwZF+6N3pAgEd8ZnXrOb09oY3DREuz4X331aiHWY7vxMTJMfdatYp4noJlJcpsOIj7djwPtbMMPQBZFdKEbZ5BSWIbOG/f+nU/MRn16AojIdtsSl4c01J/Hz7kSsOXoZALDy4CXM/eM4nl4qsgSSJGHVwUu4kFmI1YcvV3vc1Ycv44UVR1BcVjnY+mjDaaTmlmDF/otVZnAkScJuw8gWlUpkX9afSMWGkyLt/38bTleq85BHFw1vKz5Fy9mYC4ag4LZ2QdDaqZFZUIp/jovjfDS+I1QqEVjI3Rm/H7ioHLPEpBZmS1w63lkXBwBo6itmKn93XZxZ+3V6CQs3ncE7605Bp5fw99FkFJcZj/H19nNK5sbNZLTPiHaiza/9eQIHEq8oQWOPZj7484neeHJgC9zRIVjZ791/4rDIUMA6+9ZWaBciLtZfbTuHPw5dhkatwsfjO2JSTATs1CocvZSDtUdFkNW3pR8WPdgdzloNtp/JwJqjl7HhZBq2nU6HVqPGe3d3gLujPZp4Gy/sFX0zuStevr01XBzsML57GFy0GpxNy8c7605BkoAm3s6YObglejTzwZujxYLA+SXl8HS2x2cTO2POiGh0j/AGACzcJNbQGxTtjyGtA5Tn6NncB39M74XYOQPxx/ReeGKgCNTk0UqHkrJx8Uoh/m+9CECjAtyUx/aJ9MPqGb3w54zeZtvbGs5T2xAxuep3O85j/4UspdgYAA4bapdaBbpBZVLm4Ky1w2uj2uCeLqGY2CMc0UHuuLNTCFQqFTqGecJFq0FWQSlOJFedWawvDG4ampyLos4GAO5ZDDweCzTtY9s20U3lVEouBr6/Ba+uPn7tnetBbHwmMgtKUVymNxt1cyGzAKM/24FhH21HXrF5V8SxSzl4dfVxPPrjPuy/IB5jmkr/7r8ElBqChsU7RdfCEUMh5t6EKzibJrp45E/z645VPRImr7gMc347gmX7krD6kHkAdD6jAH8agqIrhWVIqaJY9uKVIqTkFsNeo8KE7k0AiC6GK4Y6jguZhXhn3Sks3ZOIcp0eZTo9kq6IT+8P9hLLmBxMvIIrBaW4kCWCnsgAN3Q0DKnW6SWEejnh1raBGGiYJO/OT3fg1dXHseaIeE2B7o4AxEUOEF1I+SXlaBngimWP3AKtRo1DSdlK8KTXS5jz+xG8/+9pfL4lHr/uS8LyfSJQmhQTDnuNCnsTrigZsJEdRTdHMz8XLJjQGTHNfFBYqsNDi/YCEF0org52cLTXYObQKHwyoRNmDBAX+P0XrqC0XI9bmnmjf5QfWgeLi/Xy/eL5xnYNQ9cIb3i7aNG3pR8AYMlu8fts6e+KbhHeynnaeDINH28UAcJDfZoiwhC4hXmZBzdykDC2ayj6RPop290c7TG6c4jy8+3tg7Dt+QF4YpBo6z1dwzBzSEt4ONnj7THt4OvqAABKIFNgeC/1a+mPW9saZ4zvGuENlUqFIA8ndAjzVLqaOjURv8Mtcem489MduJRdhGAPR/z4UHf4umrhotWgZ3Mf+Lg6oF2oB2YMNM7wLmduxnYNQ6C7I5KyijDuy13YZxLcyFoFVp5dfmzXMLx3T4dKk//Za9To0cwHAK5arF0fGNw0NLu/BCSdWBAzsK2tW0M3mbNpeRj+0XacyyjAop0JdU49/xibgD7vbsKRi9nX3SY5M1Fxbo1d5zJxLqMAGfkl2Jdg/KedklOMu7/YiUU7E/DP8VTMWn4EZTq92T/jOJNC1kNJ2Th6MQenUoyfRJftTVIyJACw78IVpFURnKw8eEm5aP17IhX/HE/B3FXHkFNUho82nDbrjjmUmI3X/jxuFijtMoyGaR/qidvbBQEwfqKWL3Jfbz+P2b8fxeLYC0jKKoROL8HJXoNOYZ5oFegGvSQKVeVaoibezugaYZyj5Na2gVCpVHjjzrbo1cIHZToJi3YmIK+kHCGeTlg5vSeeHhyJRQ90NzvH98dEwN/dUem2OWDoHlm+Pwm/7jNmfV5edQx7ErJgp1bh0X7Ncf8tEcp97o52mD6gBXq18MGrI9tAo1bho/Ed4eOiVQK46CBjxkHWOtgdI9oFwtFedAt9eV9XqFQqpZtFTiKN7BCkPGaoIYiQM0iRAWLulwGGAtl/T6Ti2KVc2GtUmNqnmfK4MG/jpHS+rlp8OrETnh8ehXkjTRb4NZhsyBB5u2jx6h2V739yUCQOzB2C4W1N22UMZBzt1ega4YXekb6ICnBDuxAPtAyo/PoB8Z7QqFXIKSpDRn4pooPc8dvjPeHv7ojVM3pjzZN94GMIoABgRLsgdI/wRoinEzobsj7Bnk7455m+GNI6AOV6MVrNRasxG48SFVj181enl6FGydZ1N5zE70YX+6lYWbvjvWJphb3fiO0xM2zbLmoQist0+L/1pzGkdQC6GtLf1+Ptv+PMfr6UXYQwk7T9f2cyEOjhgBb+1f9D/Od4Cl5ZfRySBKzYf/GqE7Pp9BKe+OUAnOzt8P497aFSqVBYWo5Np9IwIMofzlqNEtx8Mr4jfF0dsOFkGr7YGo+NJsWxu85nKhexjzeeQXGZHtFB7kjLLcb5jAL8sidRCSRMuTnYIa+kHEt2X8CpFGPA89uBS2avW5KAb3ecx8O9m8HPzcGwTcIPsReUfbafSUdsfAYKSnXYEZ+Bc+kFUKnEJ+OTybl47984nEsvwJJdiVj7lCua+7kq3U/dm3qja4Q3XB3slILTh3s3RVxqHhIzC3HO8Brk2ohwH2eo1SoMbOWPUyl52HQqDRcMNTfhPs6GGYHjAUC50AZ7OmHJw7dg2+l0zP7tCC7nFGNctzAEeTjh6cEtAQDtQzyw78IVuDrYYXQnkaXo3MQT+y9cwf4LVzCmc6gy/PzRvs2w8VQazqblQ6UC5o9phxBPJzw1OFIZHl1SrkeI4XllAe6OWPNkH6w7lozLOcW4u0tole+NhRM6Q6WCWZdJ6yBjlsHbRat0+QDAoOgAqFRHlcAn0hA0dAzzhLeLFlkFpQCAAVH+8HYxdu8HezpBrRJ1QS38XdHC363a93dkgBv+mNELns5aJTNTkRyUypr4OCMqwA1xqXmIaeYDR0M9y7qn+5i9too8nOzx8fiOOHopB1EBbri1bZCSSQn2rDxLsEatwi+P3AJ1hXPm4WSPBRM6YcLXu3AwMRtdIrxxLj0fFw0ZwKqCy6vpE+kLVwc7uDvZQ5Kkq74Ga2Lm5kaWnSRW8l41DSjIAHZ+DJQVAsGdxOR6RAaSJGHb6XTlH7Rs+b4kfLntHF6o4YiW7MJSTPhqF77Zfg56vYRPN59FbLzxon8mzXxorulQ3bNp+bjv290Y/OE2/HHIWOiq10v4Yms81h1LQU5hGWYuO6RcYPYmVE6DmzqZnIu1R1Pw24GLiE8XmZL3/zmNGT8fxH3f7sZBw6RqWjs1+rb0Q9cIbzT1FRd4+Z8zAKVuZWe8sVj1jVFtlFT9//46icJSHXxctAjycFQeN3OouKivPnwZecXlsFOr4OvqgKyCUizdI47jYrigfLn1HHq9swmJmYUoLtPh9b9O4GxaPpy1GgS6O6KkXK9kcc4ZXssTA1rgLkNXhrytVKfHzF8P48WVR5WamCGtxQy1fSKNM8kOaOWPRQ90xx8zesHJXtR6/Gaok2nmJ7pUBhoCus2n0pT6lXBvF3Rr6o0gD0e0C/FAJ0NhqqxvSz/880xf/PRQD8wYYL5YrfypfGzXMGXJhC7hIgsg12vI86bENPfBB/d0QJdwL3xwTwfc01UMPZYvymoV8MyQlpV+5wAQ6OGIKb2a4sUR0dVmLtRqVaULZ1SgmxI8DGsTADuN8RLn5+Zg9lojDbP2atQq9Gtp7F4a09k8mLLXqJUlBapri6k2wR4IqSK4uJr7bhFdjvI5AlCjoOD29sGYc2s0xnQOrdH6UJoqzhkAONpr8M2krnisX3PMubWV2YzGNXnNpiL9XXHwlSH49N7ONgtsAAY3N7Z8kxk5Dy8F9n4rbvd/kfPYkJm3/z6FSd/tqTQbqZy9iE8vMJuArbrupD8OXUbsuUx8vPEM/j6Wgvf+icOLK48CEKOFLhkChh5NxSdi02yG3C0BAE8tPYT/DENnf4hNwNt/n8Kzvx7C/sQsFBiCCPH4XORWqIcxtddkcrBtp9NRWq7HyoPiAn4wMRtjPtsJAIhp5gNnrbjYhlVRAHr0Ug6m/bQf9369Gzq9hMHR/uga4Y17ezRBdJC7Ul/TL8pPuVh7u2hxb48mcNFqlNqa5n6u6NtSXODlgsnHB7TA2K6h8HV1QGm5Hn8fS8YTvxzE9zsSAAAzBrZQinsBUQsjBypPDoo0yzYAgIOdGkcu5uAXQ/D0vzvbKt0IcvZJq1Ero2XcHO1xW3uRfZGLZeVC305NvODv5oA8Q7bHw8keHs72cHWww5bn+mPFtBio1ZX/l7g52qN3pG+l+6b1b46v7u+C2bcalyKQ2xaXmodUQyYMEBf5DmGe+G1az0oBw6iOITg8byge7dsMluRor1GKik2HLMuGGLqAgj0c4eZor2yXz6uHkz0GtPKr9LgIQ8Bc2wt9Td13SziOvjoUI9oFXXtnK/FxdcDsW1shOsgdzf1EcBPq5WR2nmpCpVLBXmP70ML2LaDqFaQbb298XWRtAtsDkUNs1yayio83nMGzvx5GWYWRLzWxMz4DX24Tk3utOZqsjFopLC1XZikFxIyrAPDSyqPo+Pq/uJApLkLlOj2+2haPA4lXlC6evOJyZWbShMwCFJXqkJxTjHK9BK1GrRRnmmZu5AnHZL8duIikrEK8+4/oyioo1eHPw6KeJKa5D8J9nCFJ5sNPKzKtldl6Oh1b4tJwpbAMXs72yvwarg52mNwzXNmvqtEtOr1Y2M9OrcKE7mF4/x4xg7eDnQYrH++JJQ/3wHt3t8fLt7VGz+YieOkT6QsHO41Z0WirIDfEGAomZe1CPPDu3R3whCELtGxvEtafSIVKBXw7uSse798CYzqHQKNWYXC0P14Z2Rp7XxyMHx7sDjuNWimCBYBmvi744r4uGNI6AENaB2DBhE647xbjaxveNhBdw73wUJ+mSvcFAKXYWNY1XASfGrVKCXwA43wx8muv7dpTjvYaDG0TaFZ74+/uiFAvMdz4592JkCRRhCx3z1XHzdHeKp/sF0zohB8f6q78Hk3d1SUE7UM9MNkwCZ7s1raBeKxfc3w4tkOV5+TpwS1xb48muKNj5YDJElQqVa2DCGuSR1F1qOe1vCyJNTc3MtPgRmcYznrLNGZtbiDJOSKTUZeVcNNyi1FYqkOghyM+3iiKSwdH++PWdkGQJAnHL+ciMsC10j9b037solIdnltu3uV08Yqog9l5NtNsOvv1J1LxSN9m+OtIMnKLy7Hy4CU8Pbgllu+/iLfWnoK3i1ap5wCMmQlJEl1O8oijUG8nZeRMnEnm5phhUrl7ezTBz7sTselUGvKKy83mB5HX12kT7AEHOw0uZBZiX8IV9DeM1Kn4Ok0zN7vOZSpv/bu7hOLx/i2QW1yGUC9nszqGIA8n2GtUKNNJhudyx3FD2z6d2FmZaEzmaK9RulsAYFy3MLg4GIOaga38sc4wR0tUoBtimpsHN3KWZGArf8xbfVyZRK9ncx8MihZFrO1DPRE7e6CyeKKHs/FC5umsRYinEy5lF2F420AMaOWvZBIqcne0x4ppPStt7xLuhW8mdUV2URnah5oXoY7sEKxkka42rPl6dAn3wsUrRcoEdvLF0RbCvJ2rzN4BgL+bKLatyF6jNstGVdQtwhvdLFCz1lDc3j4Yej3M/i4aGmZubmSmwQ0AuPgBbe+yTVuokuIyHUYu+A8jF/x3zUX4Kjqblo/BH27FiE+2Y8/5LGXUzNK9oiti9eHLuH3Bf3hzzUmzx209nY72r/6Lz7aIrMrnW+NxKbvIMOOoSCUfNoxA2mSYzXaw4QK7P/EKjl3KVWZoXX8i1azoNaugtNLaPrLTqXlKQWoTb2fl4nkuvQBlOj10ekmZMXdyTAS8nO2VWXVVKqB/lAgU5PlSWge7o5thxI7pujSACGri0/PFCKS8EthrVPBzc0BJuR5b4sTfxJjOofBy0SLcx6VSgaZGrUKoYfiuRq3C88NbwdtFi1dub10psKmKRq3CqI4hSlFpf5NuiuhAd4R6OStBglajVoo3w7ydzWoVxnQy74rxd3esdrXke3s0QbiPc6UMTG0Mbh2Au7uEVuo66RTmiVAv0UbTzI0lye8xua5HHrVEDZO9Ro27uoQi0KT+rKFhcHMjK6gwlK7bVMDu6qleqj9nUvORkV+KjPxS7DxbeaRNXnEZftp1AdmF5kW+VwpK8dDivcg1ZDWWGQIaANh2Jh2XsouUOVFWHbykBBypucV4Ztkh5JWU45vt53EuPR9fbBUjXl6+LRq3NBOfLI9czMGxSznKJGwTb2mCNsHukCTgK8PaNIBYl2b14cvK1Osy06JV2em0PCSaBDchnk5w0WpQqtPj9k/+wyt/HENRmQ7OWg1a+LtiYCvjJGQj2gUpI2tkrYPclfkw9iZkYbmhyPevI5cx+MOtGPTBVtzzRSwAkfWQL54qlRglFB109cyA/Mk93NsZ/Vr64cDcIXiwd9OrPqY6/m6OGNM5BC38XZUh1HLXVBMf86yRvO6Ok73GrM7mWqYPaIGtzw2oNuNwPVQqFWYMaAEPJ3uzYceWdFu7IHQ2zLsCGCeJI7IVBjc3MjlzM+Bl4L7fgL6zbNseMmM6F4qcJTG1aEcCXl51DLd98p9SwFtarsdjP+1XJjwDjLUwgOgCWrwzATsM09nnFpcrU9vP/u2IMhoqq6AUDyzai9JyPWKa+WB420Clf3zdsRQ8vHgfisv06NvSD30j/ZT0stwtJHtppShAvqdLqDIp2PQBLZTRHnKXy9nUfLPgRq1WoaXcNZWahyW7RXdE6yB3aNQqsxlWnxjYwmy4t7+bA/zcHNDU1wUP9moKSQKe/+0I+ry7CTN+Poj49AJlGn1AFC+/MDwKL42Ixj9P98XLt7eu6tdhpolhbhLTTMr1+HBsR2yY2U+pi5ADl24m88UAorvMzdEOD/aOgIvDjdPrP757ExyeNxQdKoyMshS1WoVX72hjGJptnCSOyFZunL8+qqzAcMH0bAK0GGzbtlAlpsW0m0+lVZrTQV7k8FJ2EX7Zm4iJPcLx6p/Hsft8Flwd7HBnp2D8tCtRGanTq4UPdpzNxNfbz8F0Jv51x1LQOtgdmw1dMkNaB2D9iVQlQHrptmioVCrlwiUHIZH+rlh4bydo1CrENPPBV9vOKSsWB3s44nJOMfJLyuGs1eCRvs0Q4OGIxMxCtA3xwPPDo/Dn4cu4u0sYHvtpP06n5cHLUC8id8lM69cc3+9IgFoNZW4Tee6QAa38MKpjMJr7uaJVoDv0hjVq8orLzQpo594eDb0kJo1LyiqCRq3C9P7NMbVvM5xOzceuc5mYFBMON0d7TK3FyJq+kX5YtjdJyfhY2oBW/ljzZG9E+LiYbY8McMPRV2/OaRrah3rii/u6oLRcjwD3htudQY0Dg5sbmdwt5dJwi7oauqJSHbaeTsegaP9KwxtNi2mTc4pxKiXPrLvEtDh3/tpTCPZ0ws+7E6FSAZ9M6AgfFwf8tCtR2eepQS1xpeC4Usgb6e+KM2n5+Od4ihK4tAvxwDODWyrZntvbByldAPLwTUDMvbLowe5wN2QaujX1hkatUoKb2SOi8dv+iwjxcsLj/ZsrNSrysUZ1DMGojiFKpigpqwipuaKeoomhbmNom0AMbROIknKxtlN8eoHSpeVgp8HH4zsp7VGrVWgb7IHYc5loYxLcqFTiE/8jfZvhdGoeInxclGnvu4R7KcOya2tom0Ace21YrUcD1QbrSiqrSU0TUX1gt9SNaPuHwMElxm4pl8rzLpD1FJSUY3NcGnR6CW+uPYHHftqPb/87X2k/OXPjbxjy+tt+45TzOr2EhAyRQZFHIT3ywz4AwMj2wRjYKgCtgtyU1ZgBMcz4+eFRys/PDo1SpqF/zzCcekArf7QOdkfP5j5wd7TDs0ON+2vUKswa2hLdm3pj3dN9zSYSc3WwM+sqiGnmg8UPdsdbo9spgU1VvF20ypw0cu1PxbV2HOw0WDW9FxY90M1sTZyKHunbDJ2beGKsyURlsmBPJ/SP8lcCG0uwZmBDRDc2Bjc3mvQ4YONrwJ9PmmRuGNxYwz/HU3DnpztwtsKsu/+3/jQe+H4vPlwfpxT2bj6Vhr0JWRjz2Q4cv5yDnMIyJOeItYTkIaTf7jiP9SdScaWgFBevFKJUp4eDnRpfT+oKAMrQZHlWXAc7DaINWYxgD0e4O9qjX0s/3NujCfq29EP/KD88YuiKkTMo8oyzix/sjtg5g5SaGNmMgZH49dGYKgtT5SJYX1eHa85BYqqNSXGoj4u2yloSN0d79I/yv+q8JQNa+eP3x3sh3MdyAQwRUVUY3NxoUg0rLevLxQKZAODsU/3+VCOSJOGnXRewxVAHk1VQiudXHMGhpGx8+1+C2b67zov6kc+3xCO3WHQtHUzMxv/WnMSBxGz8sPMCThsCohBPJ4zpHIpxXcMgScDUH/ah0xvr8dZaMYS7qa8LuoR74R7D+jgj2gWaDdXtYMimyMW5KpUKb41uhx8e7A5Hew0m94xQVmX2ddWivSHQsNeoa12wOqJdEOw1KmVET029NbotpvSMQDM/F9wfE37tBxAR2Rhrbm406afMf3b0BOy0Ve56s8orLsMFQ+FrYWk5TqfmK1PRV2dLXDpeXnUMWjs1/nt+AD7eeEaZ72XDyVS8qW8LtVqF4jKdUktjukJBqU6vrMa8P/GKshJyS8PKwvPuaI349HwcSLwCvQRlTSC5DuaNO9uiW1PvSjUJY7uGYUtcOsZ3q9xVA4gJ5l64NQrPLDuMOzqEVDlVfk21DfHA7hcHK2sC1VSol3OVKxwTEd2oGNzcaNLMJ21jl1Rlzyw7jA0nU/HUoEhsOpWGo5dy8OHYDpXWrzH1tWF+l9JyPZ5cehC7z4uJ47QaNdLzSnD4YjY6NfFCXEqe0n0ka+HvirNpxnWZzqblY7NhmYKoQNGt5Ky1w4ppPXGloBTd3tyAckNkJC9g6GivqbLWpG2IB7Y9P+Cqr3d0p1B0Dfe2yIRapqsdExE1Vjbvlvr0008REREBR0dH9OjRA3v27Lnq/h999BGioqLg5OSEsLAwPPPMMyguLq6n1taDipkbBjdmkrIKsfGUyIp8vPEMjhrWM3r/nzgUl+kq7S9JEnadyzRbY2nXuSxIEjChexiGtBFDheXRR/LxuoZ7oXWQO3q38MXUPpUnf5MXpBwUbd7F4+WiVdZdAozBzfUK83a+IRajIyJqCGz633LZsmWYOXMm5s2bhwMHDqBDhw4YNmwY0tIqT4gGAD///DNmz56NefPm4eTJk/j222+xbNkyvPjii/XccispLwEy4823cRi4meX7L0KSxMrJgMi8+LhocTmnGD/tumC2b0Z+CW79eDvGf7ULAHBHh2ClxqV7U2+8dkdbDDVMNrfq4CWcSc3D0YsiuLmlmQ/WPtUHPz3cQ1mAT6NWoZ9J4BLs4YguTSoPVR7ZwbhQoenwbCIiqh827Zb68MMPMXXqVDzwwAMAgC+++AJr1qzBd999h9mzZ1faf+fOnejVqxfuvfdeAEBERAQmTJiA3bt312u7rSbzrCgidvAA7B2B/FTAtXbFn41ZuU6vTNP/9l3tkFdcjuggd8Sn5WP270fx/Y4EPNzHONHbe+vicColD3ZqFVoFueHpwZEAgL+OJGNSTDi0dmoMbOUPfzcHXM4pxm0L/oOzVgwfNp0+PszbGQvv7QQnew1Sc0uw9bQYon97h+Aqa2CGtA6Ep/MJ6PUSgxsiIhuwWeamtLQU+/fvx+DBxpl31Wo1Bg8ejNjY2Cof07NnT+zfv1/pujp37hzWrl2LESNGVPs8JSUlyM3NNfu6Ycn1Nv6tgGDDBGiNqFvqUnYRdpzNuOZ+GfkluPPTHfh081ll28UrhRj7ZSySc4rh6WyPW9sGYVJMBLpFeGNkh2CoVeL48irdh5Oy8et+EQgte/QW/PVEHzTzc0UzP1c8OShSWZ3ZzdEeq2f0Rr+Wfigt1yO7UBQZV5w+/vb2wRgUHWA2qdzI9sFVtt/VwQ6rp/fGHzN631BT8BMR3Sxs9p83IyMDOp0OAQHm06MHBATg1KlTVT7m3nvvRUZGBnr37g1JklBeXo7HHnvsqt1S8+fPx2uvvWbRtluNXG/j1wpoMxooyQNa32nTJlmKTi9h4te7kJBZiJWP90SnKrpzZCsPXMKhpGycTM7F/THhcHe0x5zfj+JAYjZcHezwzl3t4WhvnKDNxcEO0UHuOH45FwcuZCPUqwQPLd4HSQLGdApBl3Dvq7Yt0MMRix7ohqV7k/C/v06giY8Lgqop3o30d8VdnUOhtVOhbUj1izc2sdLqy0REdG0NqkJxy5YteOutt/DZZ5/hwIED+P3337FmzRq88cYb1T5mzpw5yMnJUb6SkpKq3dfmzm0R3wPbAc0HAA+sFVmcRuCf4ylIMKyFVDF7s/9CFt775xRi4zOh10tKcW9JuR5/H01GUakOu86JguBfH42pcop3OaPyx6FLmPD1LmTkl6BVoBtevC26Ru1TqVSY0L0J9r48GKum96x2Mjq1WoUPxnbA/DHtrzphHRER2Y7NMje+vr7QaDRITU01256amorAwKqncJ87dy7uv/9+PPzwwwCAdu3aoaCgAI888gheeuklqNWVYzUHBwc4ONR8NlabSY8DLu4FVBog+g5bt8aiJEnCV9vOKT/vTbhidv9zy4/gXEYBPt0cjz6Rvth3IUu57/cDlxDi6YwynYQgD0dEB7mhKl3CvfBD7AX8awiMOoZ54qeHe9R6ThdnLbuRiIgaOptlbrRaLbp06YKNGzcq2/R6PTZu3IiYmJgqH1NYWFgpgNFoRPeEJElVPaThOLREfI8cCrhZZyVjWzmYlI1DhgnwAODAhSvKAo7JOUU4l1EAtQrQ2qmx/UwG9JKY+VelAnafz8IyQxFxTDOfarMlnSt0cz07tGWtAxsiImocbNotNXPmTHz99ddYvHgxTp48iWnTpqGgoEAZPTVp0iTMmTNH2X/kyJH4/PPPsXTpUpw/fx7r16/H3LlzMXLkSCXIaZB05cDhZeJ2p4m2bYsVbDwpsim3tQ+Cq4Md8krKcSpFFHbHGuafaRfqiVdHGmfBvbtLKHq3EEOw/zws1ne6pXn1y1CEejkpC1i2CnRTHktERDcfm360HTduHNLT0/HKK68gJSUFHTt2xLp165Qi48TERLNMzcsvvwyVSoWXX34Zly5dgp+fH0aOHIk333zTVi/BMi4fAPJTxFILkcNs3RqL+++MqLEZEOWPvOJybDudjn0JV9Am2EOZXC+mmQ8mdA/D0UvZ+Pd4Ku7qHIrb2gfhtk+2KzMG97xKcKNSqTCwlT+W7k3C9AEtWA9DRHQTU0kNvj+ndnJzc+Hh4YGcnBy4u1c/2qVebXsf2PSGqLUZ96OtW2NROYVl6PjGv5AkYNecQVi+LwkfrD+N7hHe+OGh7hj0wVZcyi7C4ge7KxPkSZKkBCefbDyDD9efRriPM7Y+d/VlCopKdUjMKkRUYNV1OURE1HDV5vrNooQbQcJ/4ntEH9u2wwpiz2VAksT6TIEejhjWNhAfbzyDPQlZGPHJdlzKLoKdWoVuEcaaGdOsy7T+zeHqYIeOTTyv+VxOWg0DGyIialhDwRul8lIgyTDDctOGG9wkZhZiX0JWpe3bDV1Scg1MywA3/PBgd7g72uFcegEAoEcz72pHKdlr1Hiwd9NKBcNERETVYebG1i4fAMoKAWdfMXlfA1FcpoODnVrJsjzy4z6cTs3DP0/3RWSAyJ6sO5aC3w5cBACzAt+eLXzx7zP98N/ZDKgA9GnJ4l8iIrIcBje2dn67+B7RG2ggRbA7zmZgyvd78MTASDw5KBL5JeU4lZIHQAzddrTX4O11p7DmSDIAoH+UH/pHmS8jEejhiLu7hNZ724mIqPFjt5StJZgENw3Egk1nUKaTsPGUWL39TGqect/hpGxM/WGfEthM6RmBbyZ1hZ2GbzUiIqofzNzYUnmJsd6mgRQTn07Nw65zorYmPi0fkiThTGq+cv/6k6nILiyDvUaFP6b3RuvgG2REGhER3TT4cdqWLu0HyovFyt9+UbZuTY38GHtBuZ1fUo7U3BKcNsncyKtq92jqw8CGiIhsgsGNLTWwepusglKlQFhr6GaKT8/HmbT8SvsObOVfr20jIiKSMbixJaXepmF0SX373zkUlurQNsQdfQ0jnM6m5Ss1NwHuxgVKGdwQEZGtMLixlbJiIGmPuH2DBDdZBaWYv/Ykery1AR+uPw293jh5dU5hGRbvFF1SMwZEooW/GO59OCkbl3OKAQCjO4nRT818XRDh61LPrSciIhJYUGwrqccAXYmY38Y30tatQXZhKUYu+A+XsosAiGUPLmQW4P/GdoRarcLb604iv6QcUQFuGNo6AHnForZm3fEUAIC/mwMe7B2BIxezcd8t4TZ7HURERAxubCX1uPge2M7m9TaSJOG5FUdwKbsIoV5OGNMpBJ9vjccfhy7j1raBKNNJ+GVPElQqYN4draFWq9DC3xUAUFiqAwBEBbrB380RP0+9xZYvhYiIiMGNzcjBTUAb27YDwJ9HkrH+RCq0GjW+uK8L2oZ4AAA+2XQW76yLQ1qu6Haa3r8FejYXtTbNDcGN7OE+zeq30URERNVgzY2tpJ0Q32+A4OafY6Jr6eE+TZXA5oFeTeGi1eB8RgEKSnWIaeaDpwcbu8/cHe0R7uMMAHhuWJSyojcREZGtMXNjC5Ikam4Amwc3Or2EHfFicctB0QHKdi8XLe6PicAXW+PR1NcFn9/XudIsw1/e3wUXs4owKJojo4iI6MbB4MYW8lKAoiuASgP42nbyvhOXc5FdWAZXBzt0CPUwu+/pwZEI8XLCkOgAeDprKz22VaA7WgVyoj4iIrqxMLixBbnexqcFYO9okyYUleqw8uAlnErJBQDc0synUmbG0V6D+znyiYiIGhgGN7aQJhcTt7ZZE15dfRzL9iUpP/du4WOzthAREVkSC4ptIe2U+O5vm3qbs2n5WL4/yWxb70gWBBMRUePAzI0t5KeK7+7B9faUxy/noFwnoUOYJz74Nw56CRgc7Y8hrQNQqpOUeWuIiIgaOgY3tlCcI747eVrtKSRJwvMrjqCwVIe5t7fGXZ/vhE4vYfED3fH3sRSoVMDzw1uhZYCb1dpARERkCwxubKE4W3x39LTaU1y8UoTl+8UK3mfT8lFcpgcATP/5AACgdwtfBjZERNQosebGFoqyxXcrZm4OJF5RbscZVu0GgCuFYk2o8d2aWO25iYiIbInBTX2TpHrJ3BxMzDb7OczbCcEeYti5t4sWg1tz4j0iImqcGNzUt9ICQF8ublsxc3MoKRsAcEeHYHg522P28Ghl/af7ejSBg53Gas9NRERkS6y5qW9yMbHaDrB3tspTlJTrcOKymJxv1tAofDKhEwBRZNyrhS9HRhERUaPG4Ka+mXZJqVRWeYrjl3NRqtPD20WLMG8nZbtKpUJUIIuIiYiocWO3VH2zcjFxam4xvtp6DgDQKcwTKisFUERERDcqZm7qmxWLiQtKyjFq4Q6k5BYDAO7oWH+TBBIREd0oGNzUNytmblYevISU3GIEeTjig7Ed0LO5r8Wfg4iI6EbHbqn6pmRuPCx6WEmS8GPsBQDAw32aMbAhIqKbFoOb+iaPlrJAt9TRizm49+tdOJ2ahz3nsxCXmgcnew3u7hJ63ccmIiJqqNgtVd8s2C21ODYBO+Mz8d1/52GnEYXDozoGw8PJ/rqPTURE1FAxuKlvFiwovpBZAAA4fDEHasOgqL4t/a77uERERA0Zg5v6ZsHMzYXMQgBAXEquMuS7Y9j1H5eIiKghY3BT3yyUuSksLUdaXgkAQC8BkCQEuDsgyLB+FBER0c2KBcX1TSkovr7RUnLWxlRHTtpHRETE4KbeWahbSq63MdWpidd1HZOIiKgxYHBT3yzULSVnbnxdHZRtrLchIiJicFO/yoqBcrE0Ql0yN5IkKbcTDMHN7e2D4KzVwNXBDu1CLDsxIBERUUPEguL6JGdtVGpAW7vVuYvLdBjxyXbkFpXhsX7NcTYtDwDQLsQDd3UOhVoNuDjw10lERMSrYX2S620c3AF17ZJmxy7l4Fy6qLP535qTyvYIX2e0C2XGhoiISMZuqfqQfATISzGOlKpDl9ShpGzltpO9Rrkd7uNynY0jIiJqXBjcWFtmPPBlH+CDqOsqJj58UQRGs4a2xKrpvdDczwU9mnrDx0VrubYSERE1AuyWsrbkQ8bbdRgGnlNYBrUaOHJRPLZDmCeiAt2wYWY/AOC8NkRERBUwuLE2B3fj7fxU8b2GmZvC0nL0f38z7DVqZTbi9iHisQxqiIiIqsbgxtrsnYy3r5wX32uYuUnIKMSVwjLl56a+LvBw5orfREREV8OaG2vT64y3swzBTQ2XXkjNLTb7uQNHRREREV0Tgxtr0xszL0rmpobdUhWDmyGtAy3UKCIiosaL3VLWZpq5yU4S32vYLZWaK+psxncLw3PDouDNkVFERETXxODG2nQmmRvJEOjUMHOTYsjcBLg7wsdkDSkiIiKqHrulrM20W0pWw8xNmklwQ0RERDXD4MbaTLulZDUtKM6TgxtmbYiIiGqKwY216arI3NS0WypH1Nwwc0NERFRzDG6sTV9eeZuT1zUfVqbTI7OAwQ0REVFtMbixtqpqbmrQLZWRXwJJAuzUKq4fRUREVAsMbqytYs2Ngzug1lS9r4mUHFFv4+/mALWaSy0QERHVFIMba6tYc1PjCfxEl5Q/u6SIiIhqhcGNtVXslqrhSKk0jpQiIiKqEwY31laxoLiGc9zI3VKBzNwQERHVCoMba9NVCG5qvGgmu6WIiIjqgsGNtdUxcyN3SzFzQ0REVDsMbqytUs2NZ40eJndLcY4bIiKi2mFwY211Hi3FgmIiIqK6YHBjbRXnuXHxueZDikp1yC0W3VkBHszcEBER1YadrRvQ6MndUs0HAZ5NgOhR13yInLVxstfAzYG/IiIiotqweebm008/RUREBBwdHdGjRw/s2bPnqvtnZ2dj+vTpCAoKgoODA1q2bIm1a9fWU2vrQC4obnILMPKjGmVuTLukVCrOTkxERFQbNk0LLFu2DDNnzsQXX3yBHj164KOPPsKwYcMQFxcHf3//SvuXlpZiyJAh8Pf3x4oVKxASEoILFy7A09Oz/htfU/JQ8BosuSBLzeOCmURERHVl0+Dmww8/xNSpU/HAAw8AAL744gusWbMG3333HWbPnl1p/++++w5ZWVnYuXMn7O3tAQARERH12eTakzM3avsaPySVI6WIiIjqzGbdUqWlpdi/fz8GDx5sbIxajcGDByM2NrbKx6xevRoxMTGYPn06AgIC0LZtW7z11lvQ6XRV7g8AJSUlyM3NNfuqV3LNjfracWRqbjF+iE1A0pVCABwpRUREVBc2y9xkZGRAp9MhICDAbHtAQABOnTpV5WPOnTuHTZs2YeLEiVi7di3Onj2Lxx9/HGVlZZg3b16Vj5k/fz5ee+01i7e/xuSh4JprZ25m/3YEm+PSlZ+ZuSEiIqo9mxcU14Zer4e/vz+++uordOnSBePGjcNLL72EL774otrHzJkzBzk5OcpXUlJSPbYYxqHgNcjcmAY2AIMbIiKiurBZ5sbX1xcajQapqalm21NTUxEYGFjlY4KCgmBvbw+NxlicGx0djZSUFJSWlkKr1VZ6jIODAxwcbNi9U4tuqQB3B2VNKfEzgxsiIqLaslnmRqvVokuXLti4caOyTa/XY+PGjYiJianyMb169cLZs2eh1+uVbadPn0ZQUFCVgc0NQS4orkG3VFGpee0Q15UiIiKqPZt2S82cORNff/01Fi9ejJMnT2LatGkoKChQRk9NmjQJc+bMUfafNm0asrKy8NRTT+H06dNYs2YN3nrrLUyfPt1WL+HadDXL3BSXGWcllvmzoJiIiKjWbDoUfNy4cUhPT8crr7yClJQUdOzYEevWrVOKjBMTE6FWG+OvsLAw/PPPP3jmmWfQvn17hISE4KmnnsILL7xgq5dwbTWsuUnPM3ZH+bs5INjTCY72NZ8bh4iIiASbz+0/Y8YMzJgxo8r7tmzZUmlbTEwMdu3aZeVWWZC+ZqOl0vLE3DYhnk7495m+0No1qFpvIiKiG4bNg5tGr4bdUmmGQmJ/dwe4cD0pIiKiOqtTemDz5s2WbkfjVcNuqTR5yQU3FhETERFdjzoFN8OHD0fz5s3xv//9r/7njWloajgUXO6WYhExERHR9alTcHPp0iXMmDEDK1asQLNmzTBs2DD8+uuvKC0ttXT7Gr4aDgVXuqXcGNwQERFdjzoFN76+vnjmmWdw6NAh7N69Gy1btsTjjz+O4OBgPPnkkzh8+LCl29lwKTU31yooloMbdksRERFdj+sektO5c2fMmTMHM2bMQH5+Pr777jt06dIFffr0wfHjxy3RxoZNWRX86sO65eDGj91SRERE16XOwU1ZWRlWrFiBESNGIDw8HP/88w8WLlyI1NRUnD17FuHh4bjnnnss2daGqYbdUulyzQ27pYiIiK5LncYcP/HEE/jll18gSRLuv/9+vPvuu2jbtq1yv4uLC95//30EBwdbrKENVg2Ggpfr9MgsEPVK7JYiIiK6PnUKbk6cOIEFCxZgzJgx1S5K6evryyHjgMlQ8OozN6l5JZAkQKNWwcflBl0ji4iIqIGoU3BjuthltQe2s0O/fv3qcvjGRRkKXn3NzcaTYmX06CA3qNWq+mgVERFRo1Wnmpv58+fju+++q7T9u+++wzvvvHPdjWpUalBz8/uBSwCA0Z1C66NFREREjVqdgpsvv/wSrVq1qrS9TZs2+OKLL667UY3KNYaCx6fn41BSNjRqFe7owBolIiKi61Wn4CYlJQVBQUGVtvv5+SE5Ofm6G9Vo6HUAJHG7moLiVQdF1qZfSz/4caQUERHRdatTcBMWFoYdO3ZU2r5jxw6OkDIld0kBgKbq4Gb/hSsAgOFtAuujRURERI1enQqKp06diqeffhplZWUYOHAgAFFk/Pzzz+PZZ5+1aAMbNLlLCqg2cxOXkgcAiA5yr48WERERNXp1Cm6ee+45ZGZm4vHHH1fWk3J0dMQLL7yAOXPmWLSBDZpp5qaKmpuM/BJkFpRCpQJa+LvWY8OIiIgarzoFNyqVCu+88w7mzp2LkydPwsnJCZGRkdXOeXPTMgtuKp/q04asTbi3M5y0V1+egYiIiGqmTsGNzNXVFd26dbNUWxofObhRqQF15fKmuFQR3LQMcKvPVhERETVqdQ5u9u3bh19//RWJiYlK15Ts999/v+6GNQrXGAZ+2hDcRAUyuCEiIrKUOo2WWrp0KXr27ImTJ09i5cqVKCsrw/Hjx7Fp0yZ4eHhYuo0Nl/7q60qdSmFwQ0REZGl1Cm7eeust/N///R/+/PNPaLVafPzxxzh16hTGjh2LJk2aWLqNDZe8rlQVw8AlSVJqbqLYLUVERGQxdQpu4uPjcdtttwEAtFotCgoKoFKp8Mwzz+Crr76yaAMbtKt0S13KLkJBqQ72GhUifF3quWFERESNV52CGy8vL+TliaxDSEgIjh07BgDIzs5GYWGh5VrX0MkFxVWNlDLU2zT3c4W9pk6/BiIiIqpCnQqK+/bti/Xr16Ndu3a455578NRTT2HTpk1Yv349Bg0aZOk2NlxyzU0Vi2bK9TYcKUVERGRZdQpuFi5ciOLiYgDASy+9BHt7e+zcuRN33XUXXn75ZYs2sEGTa27UleewOc1iYiIiIquodXBTXl6Ov/76C8OGDQMAqNVqzJ492+INaxSuUnMTl5oPgMXEREREllbrYg87Ozs89thjSuaGrqKaoeDlOj3i0wzBDTM3REREFlWnStbu3bvj0KFDFm5KIyQXFFcYCp6QWYhSnR7OWg1CPJ1s0DAiIqLGq041N48//jhmzpyJpKQkdOnSBS4u5kOZ27dvb5HGNXg6ebSUebeUvBJ4ZIAb1GpVfbeKiIioUatTcDN+/HgAwJNPPqlsU6lUkCQJKpUKOp3OMq1r6KoZCi6vKRUVwJXAiYiILK1Owc358+ct3Y7GqZqh4OfSRb0Nh4ETERFZXp2Cm/DwcEu3o3FSuqXMh4JfKRQLjfq6OtR3i4iIiBq9OgU3P/zww1XvnzRpUp0a0+joq665yS4UGR0P56pXCyciIqK6q1Nw89RTT5n9XFZWhsLCQmi1Wjg7OzO4kVXTLZVTJLZ7OjG4ISIisrQ6DQW/cuWK2Vd+fj7i4uLQu3dv/PLLL5ZuY8NVTUFxjpy5YXBDRERkcRZbsTEyMhJvv/12pazOTU1XObgp1+mRVyK2ezprbdEqIiKiRs2iy1Hb2dnh8uXLljxkw1ZF5ia3uFy57e5Yp15BIiIiuoo6XV1Xr15t9rMkSUhOTsbChQvRq1cvizSsUaii5ibbMFLKzcEOdhqLxpZERESEOgY3d955p9nPKpUKfn5+GDhwID744ANLtKtx0FVeWyq7iCOliIiIrKlOwY1er7d0OxonvWGmZpPgRh4pxWJiIiIi62C/iDVV0S0lj5TyZOaGiIjIKuoU3Nx111145513Km1/9913cc8991x3oxqNKgqK5ZobTyeOlCIiIrKGOgU327Ztw4gRIyptv/XWW7Ft27brblSjUUXNTU6RCHjc2S1FRERkFXUKbvLz86HVVs482NvbIzc397ob1WjINTemo6WKDJkbdksRERFZRZ2Cm3bt2mHZsmWVti9duhStW7e+7kY1GvoqMjeFXHqBiIjImuo0Wmru3LkYM2YM4uPjMXDgQADAxo0b8csvv2D58uUWbWCDpnRLmRQUc7QUERGRVdUpuBk5ciRWrVqFt956CytWrICTkxPat2+PDRs2oF+/fpZuY8OlDAXXKJvkeW7YLUVERGQddZ7//7bbbsNtt91mybY0PleZodiDo6WIiIisok41N3v37sXu3bsrbd+9ezf27dt33Y1qNKoYCi6PlmK3FBERkXXUKbiZPn06kpKSKm2/dOkSpk+fft2NajQqDAWXJAk5HC1FRERkVXUKbk6cOIHOnTtX2t6pUyecOHHiuhvVaMiZG0O3VGGpDmU6CQCDGyIiImupU3Dj4OCA1NTUStuTk5NhZ1fnMp7Gp7xYfNc4ADCOlLLXqOBkr6nuUURERHQd6hTcDB06FHPmzEFOTo6yLTs7Gy+++CKGDBliscY1eCX54ruDKwDgikkxsUqlslWriIiIGrU6pVnef/999O3bF+Hh4ejUqRMA4NChQwgICMCPP/5o0QY2aKUF4rtWBDdZBSK48XHhSCkiIiJrqVNwExISgiNHjmDJkiU4fPgwnJyc8MADD2DChAmwt2ctiaI0T3w3BDeZ+YbgxpXBDRERkbXUuUDGxcUFvXv3RpMmTVBaKi7af//9NwDgjjvusEzrGjo5c2Polso0ZG68mbkhIiKymjoFN+fOncPo0aNx9OhRqFQqSJJkVkOi0+ks1sAGTemWcgEAZOaXAAB8XR1s1SIiIqJGr04FxU899RSaNm2KtLQ0ODs749ixY9i6dSu6du2KLVu2WLiJDZSu3DhaqkLNDTM3RERE1lOnzE1sbCw2bdoEX19fqNVqaDQa9O7dG/Pnz8eTTz6JgwcPWrqdDU9pvvG21rxbijU3RERE1lOnzI1Op4ObmxsAwNfXF5cvXwYAhIeHIy4uznKta8jk4EajBexEMCN3S3G0FBERkfXUKXPTtm1bHD58GE2bNkWPHj3w7rvvQqvV4quvvkKzZs0s3caGqUK9DWAyFJw1N0RERFZTp+Dm5ZdfRkGBuHi//vrruP3229GnTx/4+Phg2bJlFm1ggyVP4GfokgKMQ8FZc0NERGQ9dQpuhg0bptxu0aIFTp06haysLHh5eXHmXVmpeXBTUq5DXolYa8rXhZkbIiIia7HYQlDe3t6WOlTjoAQ3oltK7pKyU6vg7sT1t4iIiKylTgXFVAMVJ/AzdEl5uXBdKSIiImticGMtJRWWXuC6UkRERPXihghuPv30U0RERMDR0RE9evTAnj17avS4pUuXQqVS4c4777RuA+ui0qKZhmHgnOOGiIjIqmwe3CxbtgwzZ87EvHnzcODAAXTo0AHDhg1DWlraVR+XkJCAWbNmoU+fPvXU0lqqUHOjLJrJYmIiIiKrsnlw8+GHH2Lq1Kl44IEH0Lp1a3zxxRdwdnbGd999V+1jdDodJk6ciNdee+3GnVeHi2YSERHZhE2Dm9LSUuzfvx+DBw9WtqnVagwePBixsbHVPu7111+Hv78/HnrooWs+R0lJCXJzc82+6kWFmpssQ+bGl91SREREVmXT4CYjIwM6nQ4BAQFm2wMCApCSklLlY/777z98++23+Prrr2v0HPPnz4eHh4fyFRYWdt3trpGKNTeFxtFSREREZD0275aqjby8PNx///34+uuv4evrW6PHzJkzBzk5OcpXUlKSlVtpUKHmJr9YTODn6sA5boiIiKzJpldaX19faDQapKammm1PTU1FYGBgpf3j4+ORkJCAkSNHKtv0ej0AwM7ODnFxcWjevLnZYxwcHODgYIMi3go1NwWlIrhxc2RwQ0REZE02zdxotVp06dIFGzduVLbp9Xps3LgRMTExlfZv1aoVjh49ikOHDilfd9xxBwYMGIBDhw7VX5dTTVSouck3LL3gomVwQ0REZE02v9LOnDkTkydPRteuXdG9e3d89NFHKCgowAMPPAAAmDRpEkJCQjB//nw4Ojqibdu2Zo/39PQEgErbba5CzY3cLeXCbikiIiKrsvmVdty4cUhPT8crr7yClJQUdOzYEevWrVOKjBMTE6FWN6jSIKFCzU1BCbuliIiI6sMNcaWdMWMGZsyYUeV9W7ZsuepjFy1aZPkGWYJSc+MGvV5CQakOADM3RERE1tYAUyINgCSZZW7kYmKAo6WIiIisjcGNNZQVAZIYxQWtKwpKRNbGTq2Cgx1PORERkTXxSmsNctYGAOydkV9SBkB0SalUKhs1ioiI6ObA4MYalC4pV0CtRr4hc8MuKSIiIutjcGMNyjBwzk5MRERU3xjcWENZsfhu5wjAOIGfK4eBExERWR2DG2vQlYjvdmLZB3mOGw4DJyIisj4GN9agEyuAQyNWAFcyNw4aW7WIiIjopsHgxhrKqwtumLkhIiKyNgY31lBN5obdUkRERNbH4MYalODGHoDJulIMboiIiKyOwY01yMGNoaCYK4ITERHVHwY31sBuKSIiIpthcGMNFQqK5YUz3TjPDRERkdUxuLGGipkbuVtKy+CGiIjI2hjcWEN189wwc0NERGR1DG6sQSko5jw3RERE9Y3BjTVUyNwUGFYFZ0ExERGR9TG4sYZyw9pSGi30ekkpKGbmhoiIyPoY3FiDrkx812hRWKaDJIkfGdwQERFZH4MbazDplpJnJ9aoVXC05+kmIiKyNl5trcGkoDhPGQaugUqlsmGjiIiIbg4MbqzBJHOTVyy6qNwc7W3YICIiopsHgxtrMCkozioQgY6Pq9aGDSIiIrp5MLixBpOCYjm48XJmcENERFQfGNxYg0m31JVCcdvbhcENERFRfWBwYw0mBcWZBQxuiIiI6hODG2swzdwwuCEiIqpXDG6sQQluHJBVIOpvWHNDRERUPxjcWEO5HNzYm9TccCg4ERFRfWBwYw0m3VJZSreUgw0bREREdPNgcGMNOsM8N3YOJsENMzdERET1gcGNNRjmuSmHHXKKWHNDRERUnxjcWIOhWyq3TKwlpVIBngxuiIiI6gWDG2swFBTnGIIbTyd7aNRcNJOIiKg+MLixBkPm5kqx+NGLc9wQERHVGwY3liZJSkHxlRKRrfFmlxQREVG9YXBjafpy5WamIXPD2YmJiIjqD4MbS5PnuAGQWSQBYHBDRERUnxjcWFp5iXIzvVgEN6y5ISIiqj8MbizNMMcNoEJmgQ4A4MPghoiIqN4wuLE0k9mJc4pF/Y27E2cnJiIiqi8MbixNztxotCgoEcGNm4OdDRtERER0c2FwY2k644rg+SWiW8qZwQ0REVG9YXBjaXJBscZBydy4Omhs2CAiIqKbC4MbS1O6pexRWCqCGxdmboiIiOoNgxtLMykozjdkbly0DG6IiIjqC4MbSzPU3EhqexSX6QEwc0NERFSfGNxYmqFbSq82Dv92Yc0NERFRvWFwY2mGgmKdWkzcZ69RwcGOwQ0REVF9YXBjaYZuqXKV6IpyZr0NERFRvWJwY2lKcCO6pVxZb0NERFSvGNxYmhzcQAQ3rLchIiKqXwxuLM1QUFwKEdRwpBQREVH9YnBjaYaC4lI5c8OaGyIionrF4MbSDN1SpZIIatgtRUREVL8Y3FiaEtywW4qIiMgWGNxYmiG4KTZkbjhaioiIqH4xuLE0ObjRi8wN57khIiKqXwxuLK1cDm7kzA1rboiIiOoTgxtLM2RuCvXi1LLmhoiIqH4xuLE0Q3BTpDMUFLNbioiIqF4xuLE0OXOj42gpIiIiW2BwY2mG4KZAJ3dLseaGiIioPjG4sTRDQXFBuQhqOBSciIiofjG4sbSyQgBAbjmHghMREdnCDRHcfPrpp4iIiICjoyN69OiBPXv2VLvv119/jT59+sDLywteXl4YPHjwVfevd4VZAIDUchcAzNwQERHVN5sHN8uWLcPMmTMxb948HDhwAB06dMCwYcOQlpZW5f5btmzBhAkTsHnzZsTGxiIsLAxDhw7FpUuX6rnl1SjMAABcLhPBDWtuiIiI6pdKkiTJlg3o0aMHunXrhoULFwIA9Ho9wsLC8MQTT2D27NnXfLxOp4OXlxcWLlyISZMmXXP/3NxceHh4ICcnB+7u7tfdfjOSBLzhB+jL0LP4E1yGL069MRyO9gxwiIiIrkdtrt82zdyUlpZi//79GDx4sLJNrVZj8ODBiI2NrdExCgsLUVZWBm9vb2s1s+ZKcgF9GQAgE+7QqFVwsLN5coyIiOimYtOCkIyMDOh0OgQEBJhtDwgIwKlTp2p0jBdeeAHBwcFmAZKpkpISlJSUKD/n5ubWvcHXUiC6pPR2ziiBFu5aDVQqlfWej4iIiCpp0GmFt99+G0uXLsXKlSvh6OhY5T7z58+Hh4eH8hUWFma9BhVmAgDKHEUWicXERERE9c+mwY2vry80Gg1SU1PNtqempiIwMPCqj33//ffx9ttv499//0X79u2r3W/OnDnIyclRvpKSkizS9ioZMjclWhHceDprrfdcREREVCWbBjdarRZdunTBxo0blW16vR4bN25ETExMtY9799138cYbb2DdunXo2rXrVZ/DwcEB7u7uZl9WYxgpVWjnAQDwcWVwQ0REVN9s3m8yc+ZMTJ48GV27dkX37t3x0UcfoaCgAA888AAAYNKkSQgJCcH8+fMBAO+88w5eeeUV/Pzzz4iIiEBKSgoAwNXVFa6urjZ7HQCUzE2uWgQ3XszcEBER1TubBzfjxo1Deno6XnnlFaSkpKBjx45Yt26dUmScmJgItdqYYPr8889RWlqKu+++2+w48+bNw6uvvlqfTa/MUHOTrRLZIW8XBjdERET1zebBDQDMmDEDM2bMqPK+LVu2mP2ckJBg/QbVlSG4yZREcOPD4IaIiKjeNejRUjccQ7dUWrnoHvNicENERFTvGNxYkrz0gmFdKWZuiIiI6h+DG0sqEN1SF0tEcMPMDRERUf1jcGNJhszNhWInAMzcEBER2QKDG0spLQTKCgEACUUiuGHmhoiIqP4xuLEUQ9ZGUtsjH05QqTjPDRERkS0wuLEUw0gpnZM3ABU8neyhUXPRTCIiovp2Q8xz0yhIEhDQFvlqTyCTXVJERPVBr9ejtLTU1s0gC9FqtWYT99YVgxtLCe0CTNuBnUeTgfMHWExMRGRlpaWlOH/+PPR6va2bQhaiVqvRtGlTaLXXdw1lcGNhmQXiEwTrbYiIrEeSJCQnJ0Oj0SAsLMwin/bJtvR6PS5fvozk5GQ0adIEKlXdSzsY3FjYFUNwwxXBiYisp7y8HIWFhQgODoazs7Otm0MW4ufnh8uXL6O8vBz29vZ1Pg5DXQvLYuaGiMjqdDodAFx39wXdWOTfp/z7rSsGNxYmBzdcEZyIyPqup+uCbjyW+n0yuLGwK4UiuPFk5oaIiKwoIiICH330UY3337JlC1QqFbKzs63WphsFa24srKCkHADg6sBTS0RE5vr374+OHTvWKiipzt69e+Hi4lLj/Xv27Ink5GR4eHhc93Pf6HgFtrDCUtFPyOCGiIhqS5Ik6HQ62Nld+xri5+dXq2NrtVoEBgbWtWkNCrulLCzfkLlxdtDYuCVERHQjmTJlCrZu3YqPP/4YKpUKKpUKixYtgkqlwt9//40uXbrAwcEB//33H+Lj4zFq1CgEBATA1dUV3bp1w4YNG8yOV7FbSqVS4ZtvvsHo0aPh7OyMyMhIrF69Wrm/YrfUokWL4OnpiX/++QfR0dFwdXXF8OHDkZycrDymvLwcTz75JDw9PeHj44MXXngBkydPxp133mnNU3XdGNxYGDM3RET1T5IkFJaW2+RLkqQatfHjjz9GTEwMpk6diuTkZCQnJyMsLAwAMHv2bLz99ts4efIk2rdvj/z8fIwYMQIbN27EwYMHMXz4cIwcORKJiYlXfY7XXnsNY8eOxZEjRzBixAhMnDgRWVlZ1e5fWFiI999/Hz/++CO2bduGxMREzJo1S7n/nXfewZIlS/D9999jx44dyM3NxapVq2r0em2JV2ALkzM3LgxuiIjqTVGZDq1f+ccmz33i9WFw1l77f76Hhwe0Wi2cnZ2V7qFTp04BAF5//XUMGTJE2dfb2xsdOnRQfn7jjTewcuVKrF69GjNmzKj2OaZMmYIJEyYAAN566y188skn2LNnD4YPH17l/mVlZfjiiy/QvHlzAMCMGTPw+uuvK/cvWLAAc+bMwejRowEACxcuxNq1a6/5Wm2NmRsLKtPpUVoupgF30bJbioiIaqZr165mP+fn52PWrFmIjo6Gp6cnXF1dcfLkyWtmbtq3b6/cdnFxgbu7O9LS0qrd39nZWQlsACAoKEjZPycnB6mpqejevbtyv0ajQZcuXWr12myB6QULKiwxTjrEzA0RUf1xstfgxOvDbPbc16viqKdZs2Zh/fr1eP/999GiRQs4OTnh7rvvvuYioRVn9VWpVFdde6uq/WvazXYj4xXYgvJLRZeU1k4New2TYkRE9UWlUtWoa8jWtFptjWbf3bFjB6ZMmaJ0B+Xn5yMhIcHKrTPn4eGBgIAA7N27F3379gUgZg4+cOAAOnbsWK9tqa0b/53QgMhz3LBLioiIqhIREYHdu3cjISEBrq6u1WZVIiMj8fvvv2PkyJFQqVSYO3euTVY/f+KJJzB//ny0aNECrVq1woIFC3DlypUbfmZophcsqIDFxEREdBWzZs2CRqNB69at4efnV20NzYcffggvLy/07NkTI0eOxLBhw9C5c+d6bi3wwgsvYMKECZg0aRJiYmLg6uqKYcOGwdHRsd7bUhsqqTF0rtVCbm4uPDw8kJOTA3d3d4se+78zGbjv292ICnDDP8/0teixiYjIqLi4GOfPn0fTpk1v+AttY6LX6xEdHY2xY8fijTfesPjxr/Z7rc31mykGCzIOA2e3FBERNXwXLlzAv//+i379+qGkpAQLFy7E+fPnce+999q6aVfFbikLKixltxQRETUearUaixYtQrdu3dCrVy8cPXoUGzZsQHR0tK2bdlW8CluQsaCYp5WIiBq+sLAw7Nixw9bNqDVmbiyowLD0AjM3REREtsPgxoIKWHNDRERkcwxuLIjrShEREdkegxsLkpdf4IrgREREtsPgxoLk5RecOUMxERGRzTC4sSDOUExERGR7DG4siN1SRERkTREREfjoo4+Un1UqFVatWlXt/gkJCVCpVDh06NB1Pa+ljlNfeBW2ILmgmN1SRERUH5KTk+Hl5WXRY06ZMgXZ2dlmQVNYWBiSk5Ph6+tr0eeyFgY3FlRgqLlh5oaIiOpDYGBgvTyPRqOpt+eyBHZLWVCBoVvKmTMUExFRBV999RWCg4Oh1+vNto8aNQoPPvgg4uPjMWrUKAQEBMDV1RXdunXDhg0brnrMit1Se/bsQadOneDo6IiuXbvi4MGDZvvrdDo89NBDaNq0KZycnBAVFYWPP/5Yuf/VV1/F4sWL8ccff0ClUkGlUmHLli1Vdktt3boV3bt3h4ODA4KCgjB79myUl5cr9/fv3x9PPvkknn/+eXh7eyMwMBCvvvpq7U9cHfAqbEFyQTEzN0RE9UySgLJC2zy3vTOgUl1zt3vuuQdPPPEENm/ejEGDBgEAsrKysG7dOqxduxb5+fkYMWIE3nzzTTg4OOCHH37AyJEjERcXhyZNmlzz+Pn5+bj99tsxZMgQ/PTTTzh//jyeeuops330ej1CQ0OxfPly+Pj4YOfOnXjkkUcQFBSEsWPHYtasWTh58iRyc3Px/fffAwC8vb1x+fJls+NcunQJI0aMwJQpU/DDDz/g1KlTmDp1KhwdHc0CmMWLF2PmzJnYvXs3YmNjMWXKFPTq1QtDhgy55uu5HrwKW4hOL6GoTF5+gTU3RET1qqwQeCvYNs/94mVA63LN3by8vHDrrbfi559/VoKbFStWwNfXFwMGDIBarUaHDh2U/d944w2sXLkSq1evxowZM655/J9//hl6vR7ffvstHB0d0aZNG1y8eBHTpk1T9rG3t8drr72m/Ny0aVPExsbi119/xdixY+Hq6gonJyeUlJRctRvqs88+Q1hYGBYuXAiVSoVWrVrh8uXLeOGFF/DKK69ArRYdQ+3bt8e8efMAAJGRkVi4cCE2btxo9eCG3VIWIq8IDnAoOBERVW3ixIn47bffUFJSAgBYsmQJxo8fD7Vajfz8fMyaNQvR0dHw9PSEq6srTp48icTExBod++TJk2jfvj0cHR2VbTExMZX2+/TTT9GlSxf4+fnB1dUVX331VY2fw/S5YmJioDLJWPXq1Qv5+fm4ePGisq19+/ZmjwsKCkJaWlqtnqsueBW2ELneRqNWwcGOMSMRUb2ydxYZFFs9dw2NHDkSkiRhzZo16NatG7Zv347/+7//AwDMmjUL69evx/vvv48WLVrAyckJd999N0pLSy3W1KVLl2LWrFn44IMPEBMTAzc3N7z33nvYvXu3xZ7DlL29vdnPKpWqUs2RNTC4sZACk9mJVTXoeyUiIgtSqWrUNWRrjo6OGDNmDJYsWYKzZ88iKioKnTt3BgDs2LEDU6ZMwejRowGIGpqEhIQaHzs6Oho//vgjiouLlezNrl27zPbZsWMHevbsiccff1zZFh8fb7aPVquFTqe75nP99ttvkCRJuebt2LEDbm5uCA0NrXGbrYUpBgthMTEREdXExIkTsWbNGnz33XeYOHGisj0yMhK///47Dh06hMOHD+Pee++tVZbj3nvvhUqlwtSpU3HixAmsXbsW77//vtk+kZGR2LdvH/755x+cPn0ac+fOxd69e832iYiIwJEjRxAXF4eMjAyUlZVVeq7HH38cSUlJeOKJJ3Dq1Cn88ccfmDdvHmbOnKnU29iS7VvQSJSW6+Gi1TC4ISKiqxo4cCC8vb0RFxeHe++9V9n+4YcfwsvLCz179sTIkSMxbNgwJatTE66urvjzzz9x9OhRdOrUCS+99BLeeecds30effRRjBkzBuPGjUOPHj2QmZlplsUBgKlTpyIqKgpdu3aFn58fduzYUem5QkJCsHbtWuzZswcdOnTAY489hoceeggvv/xyLc+GdagkSZJs3Yj6lJubCw8PD+Tk5MDd3d3ixzdN0RERkXUUFxfj/PnzaNq0qVkBLTVsV/u91ub6zcyNhTGwISIisi0GN0RERNSoMLghIiKiRoXBDRERETUqDG6IiIioUWFwQ0REDdZNNuC30bPU75PBDRERNTgajVig2JJLE5Dtyb9P+fdbV5xxjoiIGhw7Ozs4OzsjPT0d9vb2N8SsuHR99Ho90tPT4ezsDDu76wtPGNwQEVGDo1KpEBQUhPPnz+PChQu2bg5ZiFqtRpMmTa57zjgGN0RE1CBptVpERkaya6oR0Wq1FsnCMbghIqIGS61Wc/kFqoSdlERERNSoMLghIiKiRoXBDRERETUqN13NjTxBUG5uro1bQkRERDUlX7drMtHfTRfc5OXlAQDCwsJs3BIiIiKqrby8PHh4eFx1H5V0k81drdfrcfnyZbi5uV33OPqKcnNzERYWhqSkJLi7u1v02I0Nz1Xt8HzVHM9VzfFc1Q7PV81Z41xJkoS8vDwEBwdfc7j4TZe5UavVCA0NtepzuLu7841fQzxXtcPzVXM8VzXHc1U7PF81Z+lzda2MjYwFxURERNSoMLghIiKiRoXBjQU5ODhg3rx5cHBwsHVTbng8V7XD81VzPFc1x3NVOzxfNWfrc3XTFRQTERFR48bMDRERETUqDG6IiIioUWFwQ0RERI0KgxsiIiJqVBjcWMinn36KiIgIODo6okePHtizZ4+tm3RDePXVV6FSqcy+WrVqpdxfXFyM6dOnw8fHB66urrjrrruQmppqwxbXn23btmHkyJEIDg6GSqXCqlWrzO6XJAmvvPIKgoKC4OTkhMGDB+PMmTNm+2RlZWHixIlwd3eHp6cnHnroIeTn59fjq6gf1zpXU6ZMqfQ+Gz58uNk+N8u5mj9/Prp16wY3Nzf4+/vjzjvvRFxcnNk+Nfm7S0xMxG233QZnZ2f4+/vjueeeQ3l5eX2+lHpRk/PVv3//Su+vxx57zGyfm+F8ff7552jfvr0yMV9MTAz+/vtv5f4b6X3F4MYCli1bhpkzZ2LevHk4cOAAOnTogGHDhiEtLc3WTbshtGnTBsnJycrXf//9p9z3zDPP4M8//8Ty5cuxdetWXL58GWPGjLFha+tPQUEBOnTogE8//bTK+99991188skn+OKLL7B79264uLhg2LBhKC4uVvaZOHEijh8/jvXr1+Ovv/7Ctm3b8Mgjj9TXS6g31zpXADB8+HCz99kvv/xidv/Ncq62bt2K6dOnY9euXVi/fj3KysowdOhQFBQUKPtc6+9Op9PhtttuQ2lpKXbu3InFixdj0aJFeOWVV2zxkqyqJucLAKZOnWr2/nr33XeV+26W8xUaGoq3334b+/fvx759+zBw4ECMGjUKx48fB3CDva8kum7du3eXpk+frvys0+mk4OBgaf78+TZs1Y1h3rx5UocOHaq8Lzs7W7K3t5eWL1+ubDt58qQEQIqNja2nFt4YAEgrV65Uftbr9VJgYKD03nvvKduys7MlBwcH6ZdffpEkSZJOnDghAZD27t2r7PP3339LKpVKunTpUr21vb5VPFeSJEmTJ0+WRo0aVe1jbtZzJUmSlJaWJgGQtm7dKklSzf7u1q5dK6nVaiklJUXZ5/PPP5fc3d2lkpKS+n0B9azi+ZIkSerXr5/01FNPVfuYm/l8eXl5Sd98880N975i5uY6lZaWYv/+/Rg8eLCyTa1WY/DgwYiNjbVhy24cZ86cQXBwMJo1a4aJEyciMTERALB//36UlZWZnbtWrVqhSZMmN/25O3/+PFJSUszOjYeHB3r06KGcm9jYWHh6eqJr167KPoMHD4Zarcbu3bvrvc22tmXLFvj7+yMqKgrTpk1DZmamct/NfK5ycnIAAN7e3gBq9ncXGxuLdu3aISAgQNln2LBhyM3NVT6lN1YVz5dsyZIl8PX1Rdu2bTFnzhwUFhYq992M50un02Hp0qUoKChATEzMDfe+uukWzrS0jIwM6HQ6s18WAAQEBODUqVM2atWNo0ePHli0aBGioqKQnJyM1157DX369MGxY8eQkpICrVYLT09Ps8cEBAQgJSXFNg2+Qcivv6r3lXxfSkoK/P39ze63s7ODt7f3TXf+hg8fjjFjxqBp06aIj4/Hiy++iFtvvRWxsbHQaDQ37bnS6/V4+umn0atXL7Rt2xYAavR3l5KSUuV7T76vsarqfAHAvffei/DwcAQHB+PIkSN44YUXEBcXh99//x3AzXW+jh49ipiYGBQXF8PV1RUrV65E69atcejQoRvqfcXghqzq1ltvVW63b98ePXr0QHh4OH799Vc4OTnZsGXUmIwfP1653a5dO7Rv3x7NmzfHli1bMGjQIBu2zLamT5+OY8eOmdW5UfWqO1+mtVnt2rVDUFAQBg0ahPj4eDRv3ry+m2lTUVFROHToEHJycrBixQpMnjwZW7dutXWzKmG31HXy9fWFRqOpVBGempqKwMBAG7XqxuXp6YmWLVvi7NmzCAwMRGlpKbKzs8324bmD8vqv9r4KDAysVLReXl6OrKysm/78NWvWDL6+vjh79iyAm/NczZgxA3/99Rc2b96M0NBQZXtN/u4CAwOrfO/J9zVG1Z2vqvTo0QMAzN5fN8v50mq1aNGiBbp06YL58+ejQ4cO+Pjjj2+49xWDm+uk1WrRpUsXbNy4Udmm1+uxceNGxMTE2LBlN6b8/HzEx8cjKCgIXbp0gb29vdm5i4uLQ2Ji4k1/7po2bYrAwECzc5Obm4vdu3cr5yYmJgbZ2dnYv3+/ss+mTZug1+uVf743q4sXLyIzMxNBQUEAbq5zJUkSZsyYgZUrV2LTpk1o2rSp2f01+buLiYnB0aNHzQLC9evXw93dHa1bt66fF1JPrnW+qnLo0CEAMHt/3SznqyK9Xo+SkpIb731l0fLkm9TSpUslBwcHadGiRdKJEyekRx55RPL09DSrCL9ZPfvss9KWLVuk8+fPSzt27JAGDx4s+fr6SmlpaZIkSdJjjz0mNWnSRNq0aZO0b98+KSYmRoqJibFxq+tHXl6edPDgQengwYMSAOnDDz+UDh48KF24cEGSJEl6++23JU9PT+mPP/6Qjhw5Io0aNUpq2rSpVFRUpBxj+PDhUqdOnaTdu3dL//33nxQZGSlNmDDBVi/Jaq52rvLy8qRZs2ZJsbGx0vnz56UNGzZInTt3liIjI6Xi4mLlGDfLuZo2bZrk4eEhbdmyRUpOTla+CgsLlX2u9XdXXl4utW3bVho6dKh06NAhad26dZKfn580Z84cW7wkq7rW+Tp79qz0+uuvS/v27ZPOnz8v/fHHH1KzZs2kvn37Kse4Wc7X7Nmzpa1bt0rnz5+Xjhw5Is2ePVtSqVTSv//+K0nSjfW+YnBjIQsWLJCaNGkiabVaqXv37tKuXbts3aQbwrhx46SgoCBJq9VKISEh0rhx46SzZ88q9xcVFUmPP/645OXlJTk7O0ujR4+WkpOTbdji+rN582YJQKWvyZMnS5IkhoPPnTtXCggIkBwcHKRBgwZJcXFxZsfIzMyUJkyYILm6ukru7u7SAw88IOXl5dng1VjX1c5VYWGhNHToUMnPz0+yt7eXwsPDpalTp1b6cHGznKuqzhMA6fvvv1f2qcnfXUJCgnTrrbdKTk5Okq+vr/Tss89KZWVl9fxqrO9a5ysxMVHq27ev5O3tLTk4OEgtWrSQnnvuOSknJ8fsODfD+XrwwQel8PBwSavVSn5+ftKgQYOUwEaSbqz3lUqSJMmyuSAiIiIi22HNDRERETUqDG6IiIioUWFwQ0RERI0KgxsiIiJqVBjcEBERUaPC4IaIiIgaFQY3RERE1KgwuCGim96WLVugUqkqrYtDRA0TgxsiIiJqVBjcEBERUaPC4IaIbE6v12P+/Plo2rQpnJyc0KFDB6xYsQKAsctozZo1aN++PRwdHXHLLbfg2LFjZsf47bff0KZNGzg4OCAiIgIffPCB2f0lJSV44YUXEBYWBgcHB7Ro0QLffvut2T779+9H165d4ezsjJ49eyIuLs66L5yIrILBDRHZ3Pz58/HDDz/giy++wPHjx/HMM8/gvvvuw9atW5V9nnvuOXzwwQfYu3cv/Pz8MHLkSJSVlQEQQcnYsWMxfvx4HD16FK+++irmzp2LRYsWKY+fNGkSfvnlF3zyySc4efIkvvzyS7i6upq146WXXsIHH3yAffv2wc7ODg8++GC9vH4isiwunElENlVSUgJvb29s2LABMTExyvaHH34YhYWFeOSRRzBgwAAsXboU48aNAwBkZWUhNDQUixYtwtixYzFx4kSkp6fj33//VR7//PPPY82aNTh+/DhOnz6NqKgorF+/HoMHD67Uhi1btmDAgAHYsGEDBg0aBABYu3YtbrvtNhQVFcHR0dHKZ4GILImZGyKyqbNnz6KwsBBDhgyBq6ur8vXDDz8gPj5e2c808PH29kZUVBROnjwJADh58iR69epldtxevXrhzJkz0Ol0OHToEDQaDfr163fVtrRv3165HRQUBABIS0u77tdIRPXLztYNIKKbW35+PgBgzZo1CAkJMbvPwcHBLMCpKycnpxrtZ29vr9xWqVQARD0QETUszNwQkU21bt0aDg4OSExMRIsWLcy+wsLClP127dql3L5y5QpOnz6N6OhoAEB0dDR27NhhdtwdO3agZcuW0Gg0aNeuHfR6vVkNDxE1XszcEJFNubm5YdasWXjmmWeg1+vRu3dv5OTkYMeOHXB3d0d4eDgA4PXXX4ePjw8CAgLw0ksvwdfXF3feeScA4Nlnn0W3bt3wxhtvYNy4cYiNjcXChQvx2WefAQAiIiIwefJkPPjgg/jkk0/QoUMHXLhwAWlpaRg7dqytXjoRWQmDG6L/b9+ObRSGwTAMfxV1GgoaZkCeiJqG3kUKUiQbZIe0GScSG6CsQHVscOLOep4FbHevbP983TAMOR6PGccxz+czXdellJJa6+dZaJqm3O/3bNuWy+WSdV1zOBySJKWULMuSvu8zDENOp1Mej0eu1+tnjXmeU2vN7XbL6/XK+XxOrfUbxwV+mWkp4E/7mWTa9z1d1317O8A/4M8NANAUcQMANMWzFADQFDc3AEBTxA0A0BRxAwA0RdwAAE0RNwBAU8QNANAUcQMANEXcAABNETcAQFPeELAagsGAY8AAAAAASUVORK5CYII=\n"
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "print(X_train.shape)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "AsPGw5BImaNw",
        "outputId": "193cfeae-a621-4ae7-902f-2bfe6079ecdf"
      },
      "execution_count": 133,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "(150, 28, 28)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "drawAndGuess(model)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 378
        },
        "id": "YDioalpoj06A",
        "outputId": "01b6d59c-5d9a-48a8-da4e-3590a519fdd1"
      },
      "execution_count": 132,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Draw a score and then click on finish to try to recognize it.\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              "\n",
              "<canvas width=280 height=280 style=\"border:1px solid #000000;\"></canvas><br/>\n",
              "<button>Finish</button>\n",
              "<script>\n",
              "var canvas = document.querySelector('canvas')\n",
              "var ctx = canvas.getContext('2d')\n",
              "ctx.lineWidth = 20\n",
              "ctx.lineCap = 'round';\n",
              "var button = document.querySelector('button')\n",
              "var mouse = {x: 0, y: 0}\n",
              "canvas.addEventListener('mousemove', function(e) {\n",
              "  mouse.x = e.pageX - this.offsetLeft\n",
              "  mouse.y = e.pageY - this.offsetTop\n",
              "})\n",
              "canvas.onmousedown = ()=>{\n",
              "  ctx.beginPath()\n",
              "  ctx.moveTo(mouse.x, mouse.y)\n",
              "  canvas.addEventListener('mousemove', onPaint)\n",
              "}\n",
              "canvas.onmouseup = ()=>{\n",
              "  canvas.removeEventListener('mousemove', onPaint)\n",
              "}\n",
              "var onPaint = ()=>{\n",
              "  ctx.lineTo(mouse.x, mouse.y)\n",
              "  ctx.stroke()\n",
              "}\n",
              "var data = new Promise(resolve=>{\n",
              "  button.onclick = ()=>{\n",
              "    resolve(canvas.toDataURL('image/png'))\n",
              "  }\n",
              "})\n",
              "</script>\n"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "1/1 [==============================] - 0s 21ms/step\n",
            "This is a C with confidence 100.0%.\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "print(indexes_bad_prediction(X_test, Y_test))"
      ],
      "metadata": {
        "id": "hfWuQgIaxQpI"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "7zEewh1rBgMi"
      },
      "source": [
        "# This is how you can save your model and download it\n",
        "model.save('model.h5')\n",
        "files.download('model.h5')"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "koHaxgxnB2lQ"
      },
      "source": [
        "# This is how you can upload a previously saved model into your notebook. Only works with Chrome.\n",
        "files.upload()\n",
        "model = keras.models.load_model('model.h5')"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# This is how you can display your samples. Replace 'test' by your tag.\n",
        "for document in grades.find({'tags':{\n",
        "            '$in':['test']\n",
        "        }}):\n",
        "    nparr = np.frombuffer(b64decode(document['data']), np.uint8)\n",
        "    img_np = (cv2.resize(cv2.imdecode(nparr, cv2.IMREAD_UNCHANGED), (64, 64))[:,:,3])\n",
        "    plt.imshow(img_np)\n",
        "    plt.show()\n",
        "\n",
        "    print(document['label'])"
      ],
      "metadata": {
        "id": "97nMixDMBXbH"
      },
      "execution_count": null,
      "outputs": []
    }
  ]
}