image_captioning_InceptionV3_final.ipynb

{"nbformat":4,"nbformat_minor":0,"metadata":{"colab":{"name":"image_captioning_InceptionV3_final.ipynb","provenance":[],"collapsed_sections":[],"machine_shape":"hm"},"kernelspec":{"display_name":"Python 3","name":"python3"},"accelerator":"GPU"},"cells":[{"cell_type":"markdown","metadata":{"id":"K2s1A9eLRPEj"},"source":["#### Image Captioning - InceptionV3 + LSTM\n","#### <REFERENCE>\n","##### Copyright 2018 The TensorFlow Authors.\n","######@title Licensed under the Apache License, Version 2.0 (the \"License\") you may not use this file except in compliance with the License.\n","###### You may obtain a copy of the License at https://www.apache.org/licenses/LICENSE-2.0\n","###### Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on an \"AS IS\" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.\n","###### See the License for the specific language governing permissions and limitations under the License."]},{"cell_type":"markdown","metadata":{"id":"Cffg2i257iMS"},"source":["# Image Captioning이란?"]},{"cell_type":"markdown","metadata":{"id":"QASbY_HGo4Lq"},"source":["![image.png](data:image/png;base64,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)"]},{"cell_type":"markdown","source":["이미지 캡셔닝이란 이미지가 지닌 정보를 설명해주는 가장 적합한 문장을 생성하는 것입니다. 이러한 작업은 컴퓨터 비전과 자연어 처리라는 인공지능 주요 두 분야를 연결하기 때문에 중요한 의의를 갖는 연구입니다. 기존에 수행된 연구는 주로 이미지를 영어로만 문장을 생성해냈다면, 저희는 한국어로 문장을 생성하는 알고리즘을 제안합니다.\n","\n","이 노트북 파일은 End-To-End 파일입니다. [MS COCO 2014 Dataset](https://cocodataset.org/#download)을 프로그램에서 자체적으로 다운로드하고, [AI HUB에서 제공한 한국어 Annotation File을 (편집 과정을 일부 거쳤음)](https://aihub.or.kr/keti_data_board/visual_intelligence_)을 이용자가 직접 업로드하면, InceptionV3를 이용해 이미지를 전처리 및 캐시 처리한 후, 추출해낸 Feature값을 LSTM에 입력해  모델을 훈련합니다. 훈련된 모델을 활용해 새로운 이미지에 대한 캡션을 생성하고, 이를 BLEU 기준을 통해 평가하는 것까지가 하나의 프로그램이 될 것입니다. \n","\n"],"metadata":{"id":"8VOWIWue2OJj"}},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"DOhga2oXX4Sh","outputId":"c0f7ad42-ff53-42fe-f58b-52339b689a27"},"source":["!sudo apt-get install -y fonts-nanum\n","!sudo fc-cache -fv\n","!rm ~/.cache/matplotlib -rf"],"execution_count":null,"outputs":[{"output_type":"stream","name":"stdout","text":["Reading package lists... Done\n","Building dependency tree       \n","Reading state information... Done\n","The following NEW packages will be installed:\n","  fonts-nanum\n","0 upgraded, 1 newly installed, 0 to remove and 37 not upgraded.\n","Need to get 9,604 kB of archives.\n","After this operation, 29.5 MB of additional disk space will be used.\n","Get:1 http://archive.ubuntu.com/ubuntu bionic/universe amd64 fonts-nanum all 20170925-1 [9,604 kB]\n","Fetched 9,604 kB in 1s (6,636 kB/s)\n","debconf: unable to initialize frontend: Dialog\n","debconf: (No usable dialog-like program is installed, so the dialog based frontend cannot be used. at /usr/share/perl5/Debconf/FrontEnd/Dialog.pm line 76, <> line 1.)\n","debconf: falling back to frontend: Readline\n","debconf: unable to initialize frontend: Readline\n","debconf: (This frontend requires a controlling tty.)\n","debconf: falling back to frontend: Teletype\n","dpkg-preconfigure: unable to re-open stdin: \n","Selecting previously unselected package fonts-nanum.\n","(Reading database ... 155219 files and directories currently installed.)\n","Preparing to unpack .../fonts-nanum_20170925-1_all.deb ...\n","Unpacking fonts-nanum (20170925-1) ...\n","Setting up fonts-nanum (20170925-1) ...\n","Processing triggers for fontconfig (2.12.6-0ubuntu2) ...\n","/usr/share/fonts: caching, new cache contents: 0 fonts, 1 dirs\n","/usr/share/fonts/truetype: caching, new cache contents: 0 fonts, 3 dirs\n","/usr/share/fonts/truetype/humor-sans: caching, new cache contents: 1 fonts, 0 dirs\n","/usr/share/fonts/truetype/liberation: caching, new cache contents: 16 fonts, 0 dirs\n","/usr/share/fonts/truetype/nanum: caching, new cache contents: 10 fonts, 0 dirs\n","/usr/local/share/fonts: caching, new cache contents: 0 fonts, 0 dirs\n","/root/.local/share/fonts: skipping, no such directory\n","/root/.fonts: skipping, no such directory\n","/var/cache/fontconfig: cleaning cache directory\n","/root/.cache/fontconfig: not cleaning non-existent cache directory\n","/root/.fontconfig: not cleaning non-existent cache directory\n","fc-cache: succeeded\n"]}]},{"cell_type":"code","metadata":{"id":"U8l4RJ0XRPEm"},"source":["import tensorflow as tf\n","\n","import matplotlib.pyplot as plt\n","\n","import collections\n","import random\n","import numpy as np\n","import pandas as pd\n","import os\n","import time\n","import json\n","from PIL import Image\n","\n","from nltk.translate.bleu_score import sentence_bleu\n","from nltk.translate.bleu_score import SmoothingFunction\n","\n","plt.rc('font', family = 'NanumBarunGothic')"],"execution_count":null,"outputs":[]},{"cell_type":"markdown","metadata":{"id":"jBzavxSzmUiv"},"source":["## MS-COCO 데이터셋 다운로드 및 준비\n","\n","[MS-COCO 2014 데이터셋](http://cocodataset.org/#home)을 다운로드합니다. 그리고 다운로드받은 파일의 압축을 해제하고 각각 train2014, val2014 폴더로 저장합니다. 이 코드에서 caption file을 다운로드받기는 하지만, 이 파일은 영어 caption 파일로 이 코드 내에서는 활용되지 않습니다.\n","\n","만일 Colab에서 다음을 실행시킬 경우, 파일에 'captions_kor_train.json' 파일과 'captions_kor_val.json' 파일을 업로드해줍니다. 만일 로컬 PC에서 동작시키실 경우, 후에 두 annotation file의 경로를 지정해주면 됩니다."]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"krQuPYTtRPE7","outputId":"0b91d8b4-cb84-44b5-97c1-8d9287d77642"},"source":["# Download caption annotation files\n","annotation_folder = '/annotations/'\n","if not os.path.exists(os.path.abspath('.') + annotation_folder):\n","  annotation_zip = tf.keras.utils.get_file('captions.zip',\n","                                           cache_subdir=os.path.abspath('.'),\n","                                           origin='http://images.cocodataset.org/annotations/annotations_trainval2014.zip',\n","                                           extract=True)\n","  annotation_file = os.path.dirname(annotation_zip)+'/annotations/captions_train2014.json'\n","  os.remove(annotation_zip)\n","\n","# Download image files\n","image_folder = '/train2014/'\n","if not os.path.exists(os.path.abspath('.') + image_folder):\n","  image_zip = tf.keras.utils.get_file('train2014.zip',\n","                                      cache_subdir=os.path.abspath('.'),\n","                                      origin='http://images.cocodataset.org/zips/train2014.zip',\n","                                      extract=True)\n","  PATH = os.path.dirname(image_zip) + image_folder\n","  os.remove(image_zip)\n","else:\n","  PATH = os.path.abspath('.') + image_folder"],"execution_count":null,"outputs":[{"output_type":"stream","name":"stdout","text":["Downloading data from http://images.cocodataset.org/annotations/annotations_trainval2014.zip\n","252878848/252872794 [==============================] - 3s 0us/step\n","252887040/252872794 [==============================] - 3s 0us/step\n","Downloading data from http://images.cocodataset.org/zips/train2014.zip\n","13510574080/13510573713 [==============================] - 266s 0us/step\n","13510582272/13510573713 [==============================] - 266s 0us/step\n"]}]},{"cell_type":"code","source":["image_folder ='train2014/'\n","PATH = image_folder"],"metadata":{"id":"2HnQE194mkHt"},"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"id":"m8iBJCyVB2ud"},"source":["annotation_file = 'captions_kor_train.json'\n","with open(annotation_file, 'r') as f:\n","    annotations = json.load(f)"],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"id":"miER7EHMB3Ge"},"source":["# Group all captions together having the same image ID.\n","image_path_to_caption = collections.defaultdict(list)\n","for val in annotations['annotations']:\n","  caption = f\"<start> {val['caption']} <end>\"\n","  image_path = PATH + 'COCO_train2014_' + '%012d.jpg' % (val['image_id'])\n","  image_path_to_caption[image_path].append(caption)"],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"7vvqkqYGMhvm","outputId":"2dbb611f-d084-4cf9-a910-6c80f89f463c"},"source":["image_paths = list(image_path_to_caption.keys())\n","random.shuffle(image_paths)\n","\n","# 용량 제한을 두어 30,000개의 캡션(6,000개의 이미지)만 사용할 수도 있습니다.\n","train_image_paths = image_paths#[:6000]\n","print(len(train_image_paths))"],"execution_count":null,"outputs":[{"output_type":"stream","name":"stdout","text":["82783\n"]}]},{"cell_type":"code","metadata":{"id":"hrmdtMX8Lnyh"},"source":["train_captions = []\n","img_name_vector = []\n","\n","for image_path in train_image_paths:\n","  caption_list = image_path_to_caption[image_path]\n","  train_captions.extend(caption_list)\n","  img_name_vector.extend([image_path] * len(caption_list))"],"execution_count":null,"outputs":[]},{"cell_type":"markdown","metadata":{"id":"8cSW4u-ORPFQ"},"source":["## InceptionV3를사용하여 이미지 전처리하기\n","\n","다음으로, (Imagenet에 대해 사전 훈련된) InceptionV3을 사용하여 이미지를 분류하고 마지막 Convolutional Layer에서 특징을 추출해냅니다.\n","\n","- 이미지를 299px x 299px로 resize합니다.\n","\n","- [InceptionV3 설명](https://www.tensorflow.org/api_docs/python/tf/keras/applications/inception_v3/preprocess_input)"]},{"cell_type":"code","metadata":{"id":"zXR0217aRPFR"},"source":["def load_image(image_path):\n","    img = tf.io.read_file(image_path)\n","    img = tf.image.decode_jpeg(img, channels=3)\n","    img = tf.image.resize(img, (299, 299))\n","    img = tf.keras.applications.inception_v3.preprocess_input(img)\n","    return img, image_path"],"execution_count":null,"outputs":[]},{"cell_type":"markdown","metadata":{"id":"MDvIu4sXRPFV"},"source":["## InceptionV3 초기화 및 사전 훈련된 Imagenet 가중치 로드하기\n","\n","이제 Output Layer가 InceptionV3의 마지막 Convolutional Layer인 tf.keras 모델을 만듭니다.\n","\n","- 네트워크를 통해 각 이미지를 전달하고 결과 벡터를 사전 (image_name --&gt; feature_vector)에 저장합니다.\n","- 모든 이미지가 네트워크를 통해 전달된 후, 사전을 디스크에 저장합니다."]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"8H3Eo4SyDvCS","outputId":"bf127fd0-d56a-43a8-ad1a-f7263dfa62d7"},"source":["image_model = tf.keras.applications.InceptionV3(include_top=False,\n","                                                weights='imagenet')\n","new_input = image_model.input\n","hidden_layer = image_model.layers[-1].output\n","\n","image_features_extract_model = tf.keras.Model(new_input, hidden_layer)"],"execution_count":null,"outputs":[{"output_type":"stream","name":"stdout","text":["Downloading data from https://storage.googleapis.com/tensorflow/keras-applications/vgg19/vgg19_weights_tf_dim_ordering_tf_kernels_notop.h5\n","80142336/80134624 [==============================] - 1s 0us/step\n","80150528/80134624 [==============================] - 1s 0us/step\n"]}]},{"cell_type":"markdown","metadata":{"id":"rERqlR3WRPGO"},"source":["## InceptionV3에서 추출된 특성 캐시하기\n","\n","InceptionV3로 각 이미지를 사전 처리하고 출력을 디스크에 캐시합니다. 현재 코드보다 정교한 캐싱 전략(ex. 임의 액세스 디스크 I/O를 줄이기 위해 이미지 분할 등)으로 성능을 향상할 수 있습니다.\n"]},{"cell_type":"code","metadata":{"id":"Dx_fvbVgRPGQ"},"source":["# Get unique images\n","encode_train = sorted(set(img_name_vector))\n","\n","# Feel free to change batch_size according to your system configuration\n","image_dataset = tf.data.Dataset.from_tensor_slices(encode_train)\n","image_dataset = image_dataset.map(\n","  load_image, num_parallel_calls=tf.data.AUTOTUNE).batch(16)\n","\n","for img, path in image_dataset:\n","  batch_features = image_features_extract_model(img)\n","  batch_features = tf.reshape(batch_features,\n","                              (batch_features.shape[0], -1, batch_features.shape[3]))\n","\n","  for bf, p in zip(batch_features, path):\n","    path_of_feature = p.numpy().decode(\"utf-8\")\n","    np.save(path_of_feature, bf.numpy())"],"execution_count":null,"outputs":[]},{"cell_type":"markdown","metadata":{"id":"nyqH3zFwRPFi"},"source":["## 캡션 전처리 및 토큰화하기\n","\n","- 캡션을 토큰화합니다. 그리고 최대 단어의 크기를 상위 10,000개 단어로 제한합니다. \n","그 단어 이외의 단어들은 토큰 \"UNK\"(알 수 없음)로 대체됩니다.\n","- 그 후, 단어 대 인덱스 및 인덱스 대 단어 매핑을 만듭니다.\n","- 마지막으로, 모든 시퀀스를 가장 긴 시퀀스와 같은 길이로 채웁니다."]},{"cell_type":"code","metadata":{"id":"HZfK8RhQRPFj"},"source":["# Find the maximum length of any caption in the dataset\n","def calc_max_length(tensor):\n","    return max(len(t) for t in tensor)"],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"id":"oJGE34aiRPFo"},"source":["# Choose the top 10,000 words from the vocabulary\n","top_k = 10000\n","tokenizer = tf.keras.preprocessing.text.Tokenizer(num_words=top_k,\n","                                                  oov_token=\"<unk>\",\n","                                                  filters='!\"#$%&()*+.,-/:;=?@[\\]^_`{|}~')\n","tokenizer.fit_on_texts(train_captions)"],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"id":"8Q44tNQVRPFt"},"source":["tokenizer.word_index['<pad>'] = 0\n","tokenizer.index_word[0] = '<pad>'"],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"id":"0fpJb5ojRPFv"},"source":["# Create the tokenized vectors\n","train_seqs = tokenizer.texts_to_sequences(train_captions)"],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"id":"AidglIZVRPF4"},"source":["# Pad each vector to the max_length of the captions\n","# If you do not provide a max_length value, pad_sequences calculates it automatically\n","cap_vector = tf.keras.preprocessing.sequence.pad_sequences(train_seqs, padding='post')"],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"id":"gL0wkttkRPGA"},"source":["# Calculates the max_length, which is used to store the attention weights\n","max_length = calc_max_length(train_seqs)"],"execution_count":null,"outputs":[]},{"cell_type":"markdown","metadata":{"id":"M3CD75nDpvTI"},"source":["## 데이터를 훈련(Train) 및 평가(Validation) 데이터셋로 분할하기\n"]},{"cell_type":"code","metadata":{"id":"iS7DDMszRPGF"},"source":["img_to_cap_vector = collections.defaultdict(list)\n","for img, cap in zip(img_name_vector, cap_vector):\n","  img_to_cap_vector[img].append(cap)\n","\n","# Create training and validation sets using an 80-20 split randomly.\n","img_keys = list(img_to_cap_vector.keys())\n","random.shuffle(img_keys)\n","\n","slice_index = int(len(img_keys)*0.8)\n","img_name_train_keys, img_name_val_keys = img_keys[:slice_index], img_keys[slice_index:]\n","\n","img_name_train = []\n","cap_train = []\n","for imgt in img_name_train_keys:\n","  capt_len = len(img_to_cap_vector[imgt])\n","  img_name_train.extend([imgt] * capt_len)\n","  cap_train.extend(img_to_cap_vector[imgt])\n","\n","img_name_val = []\n","cap_val = []\n","for imgv in img_name_val_keys:\n","  capv_len = len(img_to_cap_vector[imgv])\n","  img_name_val.extend([imgv] * capv_len)\n","  cap_val.extend(img_to_cap_vector[imgv])"],"execution_count":null,"outputs":[]},{"cell_type":"markdown","metadata":{"id":"uEWM9xrYcg45"},"source":["## 훈련을 위한 tf.data 데이터세트 생성하기\n"]},{"cell_type":"markdown","metadata":{"id":"horagNvhhZiy"},"source":["이미지와 캡션이 준비되었습니다! 다음으로 모델 훈련에 사용할 `tf.data` 데이터세트를 만들어 봅시다."]},{"cell_type":"code","metadata":{"id":"Q3TnZ1ToRPGV"},"source":["# Feel free to change these parameters according to your system's configuration\n","\n","BATCH_SIZE = 64\n","BUFFER_SIZE = 1000\n","embedding_dim = 256\n","units = 512\n","vocab_size = top_k + 1\n","num_steps = len(img_name_train) // BATCH_SIZE\n","# Shape of the vector extracted from InceptionV3 is (64, 2048)\n","# These two variables represent that vector shape\n","features_shape = 2048\n","attention_features_shape = 64"],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"id":"SmZS2N0bXG3T"},"source":["# Load the numpy files\n","def map_func(img_name, cap):\n","  img_tensor = np.load(img_name.decode('utf-8')+'.npy')\n","  return img_tensor, cap"],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"id":"FDF_Nm3tRPGZ"},"source":["dataset = tf.data.Dataset.from_tensor_slices((img_name_train, cap_train))\n","\n","# Use map to load the numpy files in parallel\n","dataset = dataset.map(lambda item1, item2: tf.numpy_function(\n","          map_func, [item1, item2], [tf.float32, tf.int32]),\n","          num_parallel_calls=tf.data.AUTOTUNE)\n","\n","# Shuffle and batch\n","dataset = dataset.shuffle(BUFFER_SIZE).batch(BATCH_SIZE)\n","dataset = dataset.prefetch(buffer_size=tf.data.AUTOTUNE)"],"execution_count":null,"outputs":[]},{"cell_type":"markdown","metadata":{"id":"nrvoDphgRPGd"},"source":["## 모델\n","\n","모델은 Encoder로 CNN을 Decoder로 RNN을 활용합니다. \n","- CNN에서는 이미지의 Feature들을 추출해낸 후, \n","- 추출한 Feature들을 RNN을 통해서 단어 토큰으로 출력하는 구조로 이루어집니다. "]},{"cell_type":"code","metadata":{"id":"ja2LFTMSdeV3"},"source":["class BahdanauAttention(tf.keras.Model):\n","  def __init__(self, units):\n","    super(BahdanauAttention, self).__init__()\n","    self.W1 = tf.keras.layers.Dense(units)\n","    self.W2 = tf.keras.layers.Dense(units)\n","    self.V = tf.keras.layers.Dense(1)\n","\n","  def call(self, features, hidden):\n","    # features(CNN_encoder output) shape == (batch_size, 64, embedding_dim)\n","\n","    # hidden shape == (batch_size, hidden_size)\n","    # hidden_with_time_axis shape == (batch_size, 1, hidden_size)\n","    hidden_with_time_axis = tf.expand_dims(hidden, 1)\n","\n","    # attention_hidden_layer shape == (batch_size, 64, units)\n","    attention_hidden_layer = (tf.nn.tanh(self.W1(features) +\n","                                         self.W2(hidden_with_time_axis)))\n","\n","    # score shape == (batch_size, 64, 1)\n","    # This gives you an unnormalized score for each image feature.\n","    score = self.V(attention_hidden_layer)\n","\n","    # attention_weights shape == (batch_size, 64, 1)\n","    attention_weights = tf.nn.softmax(score, axis=1)\n","\n","    # context_vector shape after sum == (batch_size, hidden_size)\n","    context_vector = attention_weights * features\n","    context_vector = tf.reduce_sum(context_vector, axis=1)\n","\n","    return context_vector, attention_weights"],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"id":"AZ7R1RxHRPGf"},"source":["class CNN_Encoder(tf.keras.Model):\n","    # Since you have already extracted the features and dumped it\n","    # This encoder passes those features through a Fully connected layer\n","    def __init__(self, embedding_dim):\n","        super(CNN_Encoder, self).__init__()\n","        # shape after fc == (batch_size, 64, embedding_dim)\n","        self.fc = tf.keras.layers.Dense(embedding_dim)\n","\n","    def call(self, x):\n","        x = self.fc(x)\n","        x = tf.nn.relu(x)\n","        return x"],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"id":"V9UbGQmERPGi"},"source":["class RNN_Decoder(tf.keras.Model):\n","  def __init__(self, embedding_dim, units, vocab_size):\n","    super(RNN_Decoder, self).__init__()\n","    self.units = units\n","\n","    self.embedding = tf.keras.layers.Embedding(vocab_size, embedding_dim)\n","    self.gru = tf.keras.layers.GRU(self.units,\n","                                   return_sequences=True,\n","                                   return_state=True,\n","                                   recurrent_initializer='glorot_uniform')\n","    self.fc1 = tf.keras.layers.Dense(self.units)\n","    self.fc2 = tf.keras.layers.Dense(vocab_size)\n","\n","    self.attention = BahdanauAttention(self.units)\n","\n","  def call(self, x, features, hidden):\n","    # defining attention as a separate model\n","    context_vector, attention_weights = self.attention(features, hidden)\n","\n","    # x shape after passing through embedding == (batch_size, 1, embedding_dim)\n","    x = self.embedding(x)\n","\n","    # x shape after concatenation == (batch_size, 1, embedding_dim + hidden_size)\n","    x = tf.concat([tf.expand_dims(context_vector, 1), x], axis=-1)\n","\n","    # passing the concatenated vector to the GRU\n","    output, state = self.gru(x)\n","\n","    # shape == (batch_size, max_length, hidden_size)\n","    x = self.fc1(output)\n","\n","    # x shape == (batch_size * max_length, hidden_size)\n","    x = tf.reshape(x, (-1, x.shape[2]))\n","\n","    # output shape == (batch_size * max_length, vocab)\n","    x = self.fc2(x)\n","\n","    return x, state, attention_weights\n","\n","  def reset_state(self, batch_size):\n","    return tf.zeros((batch_size, self.units))"],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"id":"Qs_Sr03wRPGk"},"source":["encoder = CNN_Encoder(embedding_dim)\n","decoder = RNN_Decoder(embedding_dim, units, vocab_size)"],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"id":"-bYN7xA0RPGl"},"source":["optimizer = tf.keras.optimizers.Adam()\n","loss_object = tf.keras.losses.SparseCategoricalCrossentropy(\n","    from_logits=True, reduction='none')\n","\n","\n","def loss_function(real, pred):\n","  mask = tf.math.logical_not(tf.math.equal(real, 0))\n","  loss_ = loss_object(real, pred)\n","\n","  mask = tf.cast(mask, dtype=loss_.dtype)\n","  loss_ *= mask\n","\n","  return tf.reduce_mean(loss_)"],"execution_count":null,"outputs":[]},{"cell_type":"markdown","metadata":{"id":"6A3Ni64joyab"},"source":["## 체크포인트"]},{"cell_type":"code","metadata":{"id":"PpJAqPMWo0uE"},"source":["checkpoint_path = \"./checkpoints/train\"\n","ckpt = tf.train.Checkpoint(encoder=encoder,\n","                           decoder=decoder,\n","                           optimizer=optimizer)\n","ckpt_manager = tf.train.CheckpointManager(ckpt, checkpoint_path, max_to_keep=5)"],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"id":"fUkbqhc_uObw"},"source":["start_epoch = 0\n","if ckpt_manager.latest_checkpoint:\n","  start_epoch = int(ckpt_manager.latest_checkpoint.split('-')[-1])\n","  # restoring the latest checkpoint in checkpoint_path\n","  ckpt.restore(ckpt_manager.latest_checkpoint)"],"execution_count":null,"outputs":[]},{"cell_type":"markdown","metadata":{"id":"PHod7t72RPGn"},"source":["## 훈련하기\n","\n","- 각 `.npy` 파일에 저장된 특성을 추출한 다음, 해당 특성을 endcoder를 통해 전달합니다.\n","- encoder의 출력값, decoder의 hidden state(0으로 초기화됨) 및 입력값(start token)이 decoder로 전달됩니다.\n","- decoder는 예측값 및 decoder hidden state를 반환합니다.\n","- 그 후 decoder hidden state가 모델로 다시 전달되고, 예측값은 손실을 계산하는 데 사용됩니다. \n","- teacher forcing를 사용하여 decoder에 대한 다음 입력을 결정합니다. teacher forcing은 출력값이, 다음 입력값으로 decoder에 전달되는 기법입니다.\n","- 마지막으로 Gradient를 계산하고 이를 Optimizer에 적용하여 Back-propagation합니다.\n","\n"]},{"cell_type":"code","metadata":{"id":"Vt4WZ5mhJE-E"},"source":["# adding this in a separate cell because if you run the training cell\n","# many times, the loss_plot array will be reset\n","loss_plot = []"],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"id":"sqgyz2ANKlpU"},"source":["@tf.function\n","def train_step(img_tensor, target):\n","  loss = 0\n","\n","  # initializing the hidden state for each batch\n","  # because the captions are not related from image to image\n","  hidden = decoder.reset_state(batch_size=target.shape[0])\n","\n","  dec_input = tf.expand_dims([tokenizer.word_index['<start>']] * target.shape[0], 1)\n","\n","  with tf.GradientTape() as tape:\n","      features = encoder(img_tensor)\n","\n","      for i in range(1, target.shape[1]):\n","          # passing the features through the decoder\n","          predictions, hidden, _ = decoder(dec_input, features, hidden)\n","\n","          loss += loss_function(target[:, i], predictions)\n","\n","          # using teacher forcing\n","          dec_input = tf.expand_dims(target[:, i], 1)\n","\n","  total_loss = (loss / int(target.shape[1]))\n","\n","  trainable_variables = encoder.trainable_variables + decoder.trainable_variables\n","\n","  gradients = tape.gradient(loss, trainable_variables)\n","\n","  optimizer.apply_gradients(zip(gradients, trainable_variables))\n","\n","  return loss, total_loss"],"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":["- Epoch는 다음 학습을 몇 번 반복할 지를 결정합니다. \n","- Default 값으로는 20번으로 지정되어있습니다.\n","- 이용자가 원하는 만큼 지정해서 성능을 향상시킵니다."],"metadata":{"id":"9Qf-xrud3tWs"}},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"UlA4VIQpRPGo","outputId":"8c7f13be-75fd-44a9-c922-31ef7fa51143"},"source":["EPOCHS = 100\n","\n","for epoch in range(start_epoch, EPOCHS):\n","    start = time.time()\n","    total_loss = 0\n","\n","    for (batch, (img_tensor, target)) in enumerate(dataset):\n","        batch_loss, t_loss = train_step(img_tensor, target)\n","        total_loss += t_loss\n","\n","        if batch % 100 == 0:\n","            average_batch_loss = batch_loss.numpy()/int(target.shape[1])\n","            print(f'Epoch {epoch+1} Batch {batch} Loss {average_batch_loss:.4f}')\n","    # storing the epoch end loss value to plot later\n","    loss_plot.append(total_loss / num_steps)\n","\n","    if epoch % 5 == 0:\n","      ckpt_manager.save()\n","\n","    print(f'Epoch {epoch+1} Loss {total_loss/num_steps:.6f}')\n","    print(f'Time taken for 1 epoch {time.time()-start:.2f} sec\\n')"],"execution_count":null,"outputs":[{"output_type":"stream","name":"stdout","text":["Epoch 1 Batch 0 Loss 2.0545\n","Epoch 1 Batch 100 Loss 1.3776\n","Epoch 1 Batch 200 Loss 1.1991\n","Epoch 1 Batch 300 Loss 1.1407\n","Epoch 1 Batch 400 Loss 1.0351\n","Epoch 1 Batch 500 Loss 1.1145\n","Epoch 1 Batch 600 Loss 0.9742\n","Epoch 1 Batch 700 Loss 1.0252\n","Epoch 1 Batch 800 Loss 0.9339\n","Epoch 1 Batch 900 Loss 0.8882\n","Epoch 1 Batch 1000 Loss 0.9143\n","Epoch 1 Batch 1100 Loss 0.9673\n","Epoch 1 Batch 1200 Loss 0.9382\n","Epoch 1 Batch 1300 Loss 0.8988\n","Epoch 1 Batch 1400 Loss 0.9003\n","Epoch 1 Batch 1500 Loss 0.9165\n","Epoch 1 Batch 1600 Loss 0.8899\n","Epoch 1 Batch 1700 Loss 0.8482\n","Epoch 1 Batch 1800 Loss 0.8784\n","Epoch 1 Batch 1900 Loss 0.9278\n","Epoch 1 Batch 2000 Loss 0.8425\n","Epoch 1 Batch 2100 Loss 0.8488\n","Epoch 1 Batch 2200 Loss 0.7794\n","Epoch 1 Batch 2300 Loss 0.9629\n","Epoch 1 Batch 2400 Loss 0.7330\n","Epoch 1 Batch 2500 Loss 0.8496\n","Epoch 1 Batch 2600 Loss 0.8093\n","Epoch 1 Batch 2700 Loss 0.7902\n","Epoch 1 Batch 2800 Loss 0.7908\n","Epoch 1 Batch 2900 Loss 0.9197\n","Epoch 1 Batch 3000 Loss 0.7580\n","Epoch 1 Batch 3100 Loss 0.8537\n","Epoch 1 Batch 3200 Loss 0.8030\n","Epoch 1 Batch 3300 Loss 0.8703\n","Epoch 1 Batch 3400 Loss 0.8037\n","Epoch 1 Batch 3500 Loss 0.8391\n","Epoch 1 Batch 3600 Loss 0.8454\n","Epoch 1 Batch 3700 Loss 0.7650\n","Epoch 1 Batch 3800 Loss 0.8673\n","Epoch 1 Batch 3900 Loss 0.8127\n","Epoch 1 Batch 4000 Loss 0.7791\n","Epoch 1 Batch 4100 Loss 0.8237\n","Epoch 1 Loss 0.909130\n","Time taken for 1 epoch 648.03 sec\n","\n","Epoch 2 Batch 0 Loss 0.8564\n","Epoch 2 Batch 100 Loss 0.7681\n","Epoch 2 Batch 200 Loss 0.7685\n","Epoch 2 Batch 300 Loss 0.7603\n","Epoch 2 Batch 400 Loss 0.7916\n","Epoch 2 Batch 500 Loss 0.8286\n","Epoch 2 Batch 600 Loss 0.8067\n","Epoch 2 Batch 700 Loss 0.8025\n","Epoch 2 Batch 800 Loss 0.8019\n","Epoch 2 Batch 900 Loss 0.8329\n","Epoch 2 Batch 1000 Loss 0.7958\n","Epoch 2 Batch 1100 Loss 0.7886\n","Epoch 2 Batch 1200 Loss 0.8258\n","Epoch 2 Batch 1300 Loss 0.7488\n","Epoch 2 Batch 1400 Loss 0.7921\n","Epoch 2 Batch 1500 Loss 0.8029\n","Epoch 2 Batch 1600 Loss 0.7716\n","Epoch 2 Batch 1700 Loss 0.8089\n","Epoch 2 Batch 1800 Loss 0.7292\n","Epoch 2 Batch 1900 Loss 0.7346\n","Epoch 2 Batch 2000 Loss 0.7583\n","Epoch 2 Batch 2100 Loss 0.7755\n","Epoch 2 Batch 2200 Loss 0.8007\n","Epoch 2 Batch 2300 Loss 0.7100\n","Epoch 2 Batch 2400 Loss 0.7287\n","Epoch 2 Batch 2500 Loss 0.7459\n","Epoch 2 Batch 2600 Loss 0.8289\n","Epoch 2 Batch 2700 Loss 0.8602\n","Epoch 2 Batch 2800 Loss 0.7600\n","Epoch 2 Batch 2900 Loss 0.7828\n","Epoch 2 Batch 3000 Loss 0.7880\n","Epoch 2 Batch 3100 Loss 0.7687\n","Epoch 2 Batch 3200 Loss 0.7658\n","Epoch 2 Batch 3300 Loss 0.7417\n","Epoch 2 Batch 3400 Loss 0.7335\n","Epoch 2 Batch 3500 Loss 0.6775\n","Epoch 2 Batch 3600 Loss 0.7028\n","Epoch 2 Batch 3700 Loss 0.7385\n","Epoch 2 Batch 3800 Loss 0.8046\n","Epoch 2 Batch 3900 Loss 0.7304\n","Epoch 2 Batch 4000 Loss 0.6885\n","Epoch 2 Batch 4100 Loss 0.7608\n","Epoch 2 Loss 0.774365\n","Time taken for 1 epoch 578.28 sec\n","\n","Epoch 3 Batch 0 Loss 0.7598\n","Epoch 3 Batch 100 Loss 0.7218\n","Epoch 3 Batch 200 Loss 0.6984\n","Epoch 3 Batch 300 Loss 0.7704\n","Epoch 3 Batch 400 Loss 0.7165\n","Epoch 3 Batch 500 Loss 0.6894\n","Epoch 3 Batch 600 Loss 0.7404\n","Epoch 3 Batch 700 Loss 0.6932\n","Epoch 3 Batch 800 Loss 0.7093\n","Epoch 3 Batch 900 Loss 0.7698\n","Epoch 3 Batch 1000 Loss 0.6905\n","Epoch 3 Batch 1100 Loss 0.6949\n","Epoch 3 Batch 1200 Loss 0.7574\n","Epoch 3 Batch 1300 Loss 0.8199\n","Epoch 3 Batch 1400 Loss 0.7039\n","Epoch 3 Batch 1500 Loss 0.7488\n","Epoch 3 Batch 1600 Loss 0.8100\n","Epoch 3 Batch 1700 Loss 0.6962\n","Epoch 3 Batch 1800 Loss 0.8406\n","Epoch 3 Batch 1900 Loss 0.7465\n","Epoch 3 Batch 2000 Loss 0.6656\n","Epoch 3 Batch 2100 Loss 0.7477\n","Epoch 3 Batch 2200 Loss 0.6546\n","Epoch 3 Batch 2300 Loss 0.7301\n","Epoch 3 Batch 2400 Loss 0.7732\n","Epoch 3 Batch 2500 Loss 0.7214\n","Epoch 3 Batch 2600 Loss 0.7161\n","Epoch 3 Batch 2700 Loss 0.7653\n","Epoch 3 Batch 2800 Loss 0.7496\n","Epoch 3 Batch 2900 Loss 0.7409\n","Epoch 3 Batch 3000 Loss 0.7051\n","Epoch 3 Batch 3100 Loss 0.6829\n","Epoch 3 Batch 3200 Loss 0.6804\n","Epoch 3 Batch 3300 Loss 0.7344\n","Epoch 3 Batch 3400 Loss 0.7337\n","Epoch 3 Batch 3500 Loss 0.7030\n","Epoch 3 Batch 3600 Loss 0.6767\n","Epoch 3 Batch 3700 Loss 0.7040\n","Epoch 3 Batch 3800 Loss 0.7425\n","Epoch 3 Batch 3900 Loss 0.6971\n","Epoch 3 Batch 4000 Loss 0.7688\n","Epoch 3 Batch 4100 Loss 0.7241\n","Epoch 3 Loss 0.729162\n","Time taken for 1 epoch 578.27 sec\n","\n","Epoch 4 Batch 0 Loss 0.7686\n","Epoch 4 Batch 100 Loss 0.7280\n","Epoch 4 Batch 200 Loss 0.7493\n","Epoch 4 Batch 300 Loss 0.7067\n","Epoch 4 Batch 400 Loss 0.7021\n","Epoch 4 Batch 500 Loss 0.7125\n","Epoch 4 Batch 600 Loss 0.6431\n","Epoch 4 Batch 700 Loss 0.7379\n","Epoch 4 Batch 800 Loss 0.6866\n","Epoch 4 Batch 900 Loss 0.7241\n","Epoch 4 Batch 1000 Loss 0.7074\n","Epoch 4 Batch 1100 Loss 0.6640\n","Epoch 4 Batch 1200 Loss 0.6915\n","Epoch 4 Batch 1300 Loss 0.7724\n","Epoch 4 Batch 1400 Loss 0.6815\n","Epoch 4 Batch 1500 Loss 0.6475\n","Epoch 4 Batch 1600 Loss 0.6936\n","Epoch 4 Batch 1700 Loss 0.7569\n","Epoch 4 Batch 1800 Loss 0.7473\n","Epoch 4 Batch 1900 Loss 0.7100\n","Epoch 4 Batch 2000 Loss 0.7429\n","Epoch 4 Batch 2100 Loss 0.7189\n","Epoch 4 Batch 2200 Loss 0.6639\n","Epoch 4 Batch 2300 Loss 0.7122\n","Epoch 4 Batch 2400 Loss 0.6971\n","Epoch 4 Batch 2500 Loss 0.6910\n","Epoch 4 Batch 2600 Loss 0.6995\n","Epoch 4 Batch 2700 Loss 0.7293\n","Epoch 4 Batch 2800 Loss 0.6733\n","Epoch 4 Batch 2900 Loss 0.7109\n","Epoch 4 Batch 3000 Loss 0.7440\n","Epoch 4 Batch 3100 Loss 0.7615\n","Epoch 4 Batch 3200 Loss 0.6574\n","Epoch 4 Batch 3300 Loss 0.7111\n","Epoch 4 Batch 3400 Loss 0.6818\n","Epoch 4 Batch 3500 Loss 0.7503\n","Epoch 4 Batch 3600 Loss 0.7217\n","Epoch 4 Batch 3700 Loss 0.6693\n","Epoch 4 Batch 3800 Loss 0.7519\n","Epoch 4 Batch 3900 Loss 0.6522\n","Epoch 4 Batch 4000 Loss 0.7327\n","Epoch 4 Batch 4100 Loss 0.7350\n","Epoch 4 Loss 0.698789\n","Time taken for 1 epoch 579.50 sec\n","\n","Epoch 5 Batch 0 Loss 0.7007\n","Epoch 5 Batch 100 Loss 0.6326\n","Epoch 5 Batch 200 Loss 0.7310\n","Epoch 5 Batch 300 Loss 0.7362\n","Epoch 5 Batch 400 Loss 0.6200\n","Epoch 5 Batch 500 Loss 0.7001\n","Epoch 5 Batch 600 Loss 0.6604\n","Epoch 5 Batch 700 Loss 0.6909\n","Epoch 5 Batch 800 Loss 0.7094\n","Epoch 5 Batch 900 Loss 0.6102\n","Epoch 5 Batch 1000 Loss 0.6560\n","Epoch 5 Batch 1100 Loss 0.6200\n","Epoch 5 Batch 1200 Loss 0.7428\n","Epoch 5 Batch 1300 Loss 0.7750\n","Epoch 5 Batch 1400 Loss 0.5830\n","Epoch 5 Batch 1500 Loss 0.6895\n","Epoch 5 Batch 1600 Loss 0.6589\n","Epoch 5 Batch 1700 Loss 0.6636\n","Epoch 5 Batch 1800 Loss 0.6436\n","Epoch 5 Batch 1900 Loss 0.7145\n","Epoch 5 Batch 2000 Loss 0.6890\n","Epoch 5 Batch 2100 Loss 0.7075\n","Epoch 5 Batch 2200 Loss 0.7089\n","Epoch 5 Batch 2300 Loss 0.6954\n","Epoch 5 Batch 2400 Loss 0.7072\n","Epoch 5 Batch 2500 Loss 0.6682\n","Epoch 5 Batch 2600 Loss 0.6578\n","Epoch 5 Batch 2700 Loss 0.6498\n","Epoch 5 Batch 2800 Loss 0.5869\n","Epoch 5 Batch 2900 Loss 0.6776\n","Epoch 5 Batch 3000 Loss 0.6344\n","Epoch 5 Batch 3100 Loss 0.6368\n","Epoch 5 Batch 3200 Loss 0.7320\n","Epoch 5 Batch 3300 Loss 0.6521\n","Epoch 5 Batch 3400 Loss 0.7253\n","Epoch 5 Batch 3500 Loss 0.6309\n","Epoch 5 Batch 3600 Loss 0.7082\n","Epoch 5 Batch 3700 Loss 0.6656\n","Epoch 5 Batch 3800 Loss 0.6726\n","Epoch 5 Batch 3900 Loss 0.6530\n","Epoch 5 Batch 4000 Loss 0.6941\n","Epoch 5 Batch 4100 Loss 0.5904\n","Epoch 5 Loss 0.673745\n","Time taken for 1 epoch 579.20 sec\n","\n","Epoch 6 Batch 0 Loss 0.7167\n","Epoch 6 Batch 100 Loss 0.6329\n","Epoch 6 Batch 200 Loss 0.6574\n","Epoch 6 Batch 300 Loss 0.6198\n","Epoch 6 Batch 400 Loss 0.6327\n","Epoch 6 Batch 500 Loss 0.5993\n","Epoch 6 Batch 600 Loss 0.6185\n","Epoch 6 Batch 700 Loss 0.6622\n","Epoch 6 Batch 800 Loss 0.6703\n","Epoch 6 Batch 900 Loss 0.6019\n","Epoch 6 Batch 1000 Loss 0.6585\n","Epoch 6 Batch 1100 Loss 0.6580\n","Epoch 6 Batch 1200 Loss 0.6581\n","Epoch 6 Batch 1300 Loss 0.6569\n","Epoch 6 Batch 1400 Loss 0.6351\n","Epoch 6 Batch 1500 Loss 0.7084\n","Epoch 6 Batch 1600 Loss 0.7010\n","Epoch 6 Batch 1700 Loss 0.6128\n","Epoch 6 Batch 1800 Loss 0.6244\n","Epoch 6 Batch 1900 Loss 0.6535\n","Epoch 6 Batch 2000 Loss 0.6239\n","Epoch 6 Batch 2100 Loss 0.6204\n","Epoch 6 Batch 2200 Loss 0.6824\n","Epoch 6 Batch 2300 Loss 0.5656\n","Epoch 6 Batch 2400 Loss 0.6741\n","Epoch 6 Batch 2500 Loss 0.6538\n","Epoch 6 Batch 2600 Loss 0.6296\n","Epoch 6 Batch 2700 Loss 0.6361\n","Epoch 6 Batch 2800 Loss 0.6806\n","Epoch 6 Batch 2900 Loss 0.6386\n","Epoch 6 Batch 3000 Loss 0.7307\n","Epoch 6 Batch 3100 Loss 0.7016\n","Epoch 6 Batch 3200 Loss 0.7050\n","Epoch 6 Batch 3300 Loss 0.6518\n","Epoch 6 Batch 3400 Loss 0.6123\n","Epoch 6 Batch 3500 Loss 0.6804\n","Epoch 6 Batch 3600 Loss 0.5393\n","Epoch 6 Batch 3700 Loss 0.5802\n","Epoch 6 Batch 3800 Loss 0.6289\n","Epoch 6 Batch 3900 Loss 0.6829\n","Epoch 6 Batch 4000 Loss 0.6781\n","Epoch 6 Batch 4100 Loss 0.6945\n","Epoch 6 Loss 0.651306\n","Time taken for 1 epoch 579.43 sec\n","\n","Epoch 7 Batch 0 Loss 0.7048\n","Epoch 7 Batch 100 Loss 0.6261\n","Epoch 7 Batch 200 Loss 0.6219\n","Epoch 7 Batch 300 Loss 0.6669\n","Epoch 7 Batch 400 Loss 0.6822\n","Epoch 7 Batch 500 Loss 0.6150\n","Epoch 7 Batch 600 Loss 0.6349\n","Epoch 7 Batch 700 Loss 0.6321\n","Epoch 7 Batch 800 Loss 0.6456\n","Epoch 7 Batch 900 Loss 0.6684\n","Epoch 7 Batch 1000 Loss 0.6073\n","Epoch 7 Batch 1100 Loss 0.6493\n","Epoch 7 Batch 1200 Loss 0.5638\n","Epoch 7 Batch 1300 Loss 0.5704\n","Epoch 7 Batch 1400 Loss 0.6177\n","Epoch 7 Batch 1500 Loss 0.6421\n","Epoch 7 Batch 1600 Loss 0.6285\n","Epoch 7 Batch 1700 Loss 0.6011\n","Epoch 7 Batch 1800 Loss 0.6158\n","Epoch 7 Batch 1900 Loss 0.6196\n","Epoch 7 Batch 2000 Loss 0.5615\n","Epoch 7 Batch 2100 Loss 0.5811\n","Epoch 7 Batch 2200 Loss 0.5999\n","Epoch 7 Batch 2300 Loss 0.6174\n","Epoch 7 Batch 2400 Loss 0.6224\n","Epoch 7 Batch 2500 Loss 0.6225\n","Epoch 7 Batch 2600 Loss 0.6177\n","Epoch 7 Batch 2700 Loss 0.5607\n","Epoch 7 Batch 2800 Loss 0.5966\n","Epoch 7 Batch 2900 Loss 0.6013\n","Epoch 7 Batch 3000 Loss 0.6469\n","Epoch 7 Batch 3100 Loss 0.6136\n","Epoch 7 Batch 3200 Loss 0.6298\n","Epoch 7 Batch 3300 Loss 0.6079\n","Epoch 7 Batch 3400 Loss 0.7005\n","Epoch 7 Batch 3500 Loss 0.5478\n","Epoch 7 Batch 3600 Loss 0.5839\n","Epoch 7 Batch 3700 Loss 0.6115\n","Epoch 7 Batch 3800 Loss 0.6415\n","Epoch 7 Batch 3900 Loss 0.6726\n","Epoch 7 Batch 4000 Loss 0.5906\n","Epoch 7 Batch 4100 Loss 0.5695\n","Epoch 7 Loss 0.630318\n","Time taken for 1 epoch 578.91 sec\n","\n","Epoch 8 Batch 0 Loss 0.5877\n","Epoch 8 Batch 100 Loss 0.6015\n","Epoch 8 Batch 200 Loss 0.6209\n","Epoch 8 Batch 300 Loss 0.6707\n","Epoch 8 Batch 400 Loss 0.6839\n","Epoch 8 Batch 500 Loss 0.6415\n","Epoch 8 Batch 600 Loss 0.6701\n","Epoch 8 Batch 700 Loss 0.6042\n","Epoch 8 Batch 800 Loss 0.5988\n","Epoch 8 Batch 900 Loss 0.6096\n","Epoch 8 Batch 1000 Loss 0.6866\n","Epoch 8 Batch 1100 Loss 0.6483\n","Epoch 8 Batch 1200 Loss 0.6204\n","Epoch 8 Batch 1300 Loss 0.6216\n","Epoch 8 Batch 1400 Loss 0.5756\n","Epoch 8 Batch 1500 Loss 0.6944\n","Epoch 8 Batch 1600 Loss 0.6482\n","Epoch 8 Batch 1700 Loss 0.5876\n","Epoch 8 Batch 1800 Loss 0.5816\n","Epoch 8 Batch 1900 Loss 0.5649\n","Epoch 8 Batch 2000 Loss 0.6277\n","Epoch 8 Batch 2100 Loss 0.6125\n","Epoch 8 Batch 2200 Loss 0.5701\n","Epoch 8 Batch 2300 Loss 0.6436\n","Epoch 8 Batch 2400 Loss 0.6475\n","Epoch 8 Batch 2500 Loss 0.5721\n","Epoch 8 Batch 2600 Loss 0.6128\n","Epoch 8 Batch 2700 Loss 0.6200\n","Epoch 8 Batch 2800 Loss 0.5618\n","Epoch 8 Batch 2900 Loss 0.5790\n","Epoch 8 Batch 3000 Loss 0.6387\n","Epoch 8 Batch 3100 Loss 0.5886\n","Epoch 8 Batch 3200 Loss 0.6086\n","Epoch 8 Batch 3300 Loss 0.6485\n","Epoch 8 Batch 3400 Loss 0.6794\n","Epoch 8 Batch 3500 Loss 0.5531\n","Epoch 8 Batch 3600 Loss 0.6123\n","Epoch 8 Batch 3700 Loss 0.5851\n","Epoch 8 Batch 3800 Loss 0.6551\n","Epoch 8 Batch 3900 Loss 0.6062\n","Epoch 8 Batch 4000 Loss 0.6167\n","Epoch 8 Batch 4100 Loss 0.5971\n","Epoch 8 Loss 0.610972\n","Time taken for 1 epoch 579.15 sec\n","\n","Epoch 9 Batch 0 Loss 0.5373\n","Epoch 9 Batch 100 Loss 0.6679\n","Epoch 9 Batch 200 Loss 0.6629\n","Epoch 9 Batch 300 Loss 0.6181\n","Epoch 9 Batch 400 Loss 0.6101\n","Epoch 9 Batch 500 Loss 0.6274\n","Epoch 9 Batch 600 Loss 0.6240\n","Epoch 9 Batch 700 Loss 0.6341\n","Epoch 9 Batch 800 Loss 0.5880\n","Epoch 9 Batch 900 Loss 0.5848\n","Epoch 9 Batch 1000 Loss 0.5231\n","Epoch 9 Batch 1100 Loss 0.6219\n","Epoch 9 Batch 1200 Loss 0.6688\n","Epoch 9 Batch 1300 Loss 0.6317\n","Epoch 9 Batch 1400 Loss 0.5060\n","Epoch 9 Batch 1500 Loss 0.6612\n","Epoch 9 Batch 1600 Loss 0.5784\n","Epoch 9 Batch 1700 Loss 0.5911\n","Epoch 9 Batch 1800 Loss 0.6230\n","Epoch 9 Batch 1900 Loss 0.6518\n","Epoch 9 Batch 2000 Loss 0.5957\n","Epoch 9 Batch 2100 Loss 0.6236\n","Epoch 9 Batch 2200 Loss 0.5595\n","Epoch 9 Batch 2300 Loss 0.5624\n","Epoch 9 Batch 2400 Loss 0.5422\n","Epoch 9 Batch 2500 Loss 0.5854\n","Epoch 9 Batch 2600 Loss 0.6060\n","Epoch 9 Batch 2700 Loss 0.5596\n","Epoch 9 Batch 2800 Loss 0.5631\n","Epoch 9 Batch 2900 Loss 0.5626\n","Epoch 9 Batch 3000 Loss 0.6271\n","Epoch 9 Batch 3100 Loss 0.5567\n","Epoch 9 Batch 3200 Loss 0.5439\n","Epoch 9 Batch 3300 Loss 0.6525\n","Epoch 9 Batch 3400 Loss 0.6140\n","Epoch 9 Batch 3500 Loss 0.6020\n","Epoch 9 Batch 3600 Loss 0.5368\n","Epoch 9 Batch 3700 Loss 0.6643\n","Epoch 9 Batch 3800 Loss 0.5356\n","Epoch 9 Batch 3900 Loss 0.5283\n","Epoch 9 Batch 4000 Loss 0.6017\n","Epoch 9 Batch 4100 Loss 0.5713\n","Epoch 9 Loss 0.593099\n","Time taken for 1 epoch 579.47 sec\n","\n","Epoch 10 Batch 0 Loss 0.5637\n","Epoch 10 Batch 100 Loss 0.5635\n","Epoch 10 Batch 200 Loss 0.6014\n","Epoch 10 Batch 300 Loss 0.5795\n","Epoch 10 Batch 400 Loss 0.5786\n","Epoch 10 Batch 500 Loss 0.5545\n","Epoch 10 Batch 600 Loss 0.5538\n","Epoch 10 Batch 700 Loss 0.5410\n","Epoch 10 Batch 800 Loss 0.5985\n","Epoch 10 Batch 900 Loss 0.6347\n","Epoch 10 Batch 1000 Loss 0.5312\n","Epoch 10 Batch 1100 Loss 0.5407\n","Epoch 10 Batch 1200 Loss 0.6519\n","Epoch 10 Batch 1300 Loss 0.5344\n","Epoch 10 Batch 1400 Loss 0.5322\n","Epoch 10 Batch 1500 Loss 0.6024\n","Epoch 10 Batch 1600 Loss 0.5927\n","Epoch 10 Batch 1700 Loss 0.5818\n","Epoch 10 Batch 1800 Loss 0.5478\n","Epoch 10 Batch 1900 Loss 0.6204\n","Epoch 10 Batch 2000 Loss 0.5886\n","Epoch 10 Batch 2100 Loss 0.5363\n","Epoch 10 Batch 2200 Loss 0.5674\n","Epoch 10 Batch 2300 Loss 0.5425\n","Epoch 10 Batch 2400 Loss 0.5696\n","Epoch 10 Batch 2500 Loss 0.6047\n","Epoch 10 Batch 2600 Loss 0.5730\n","Epoch 10 Batch 2700 Loss 0.5579\n","Epoch 10 Batch 2800 Loss 0.5927\n","Epoch 10 Batch 2900 Loss 0.5565\n","Epoch 10 Batch 3000 Loss 0.5338\n","Epoch 10 Batch 3100 Loss 0.5668\n","Epoch 10 Batch 3200 Loss 0.5306\n","Epoch 10 Batch 3300 Loss 0.5606\n","Epoch 10 Batch 3400 Loss 0.5495\n","Epoch 10 Batch 3500 Loss 0.6089\n","Epoch 10 Batch 3600 Loss 0.6109\n","Epoch 10 Batch 3700 Loss 0.5327\n","Epoch 10 Batch 3800 Loss 0.5033\n","Epoch 10 Batch 3900 Loss 0.6006\n","Epoch 10 Batch 4000 Loss 0.5741\n","Epoch 10 Batch 4100 Loss 0.5632\n","Epoch 10 Loss 0.576601\n","Time taken for 1 epoch 579.51 sec\n","\n","Epoch 11 Batch 0 Loss 0.5596\n","Epoch 11 Batch 100 Loss 0.6178\n","Epoch 11 Batch 200 Loss 0.5406\n","Epoch 11 Batch 300 Loss 0.5679\n","Epoch 11 Batch 400 Loss 0.5750\n","Epoch 11 Batch 500 Loss 0.5077\n","Epoch 11 Batch 600 Loss 0.5677\n","Epoch 11 Batch 700 Loss 0.6321\n","Epoch 11 Batch 800 Loss 0.6534\n","Epoch 11 Batch 900 Loss 0.5565\n","Epoch 11 Batch 1000 Loss 0.5806\n","Epoch 11 Batch 1100 Loss 0.5835\n","Epoch 11 Batch 1200 Loss 0.5594\n","Epoch 11 Batch 1300 Loss 0.5209\n","Epoch 11 Batch 1400 Loss 0.5665\n","Epoch 11 Batch 1500 Loss 0.5689\n","Epoch 11 Batch 1600 Loss 0.5459\n","Epoch 11 Batch 1700 Loss 0.5398\n","Epoch 11 Batch 1800 Loss 0.5091\n","Epoch 11 Batch 1900 Loss 0.5692\n","Epoch 11 Batch 2000 Loss 0.5392\n","Epoch 11 Batch 2100 Loss 0.5345\n","Epoch 11 Batch 2200 Loss 0.5495\n","Epoch 11 Batch 2300 Loss 0.5131\n","Epoch 11 Batch 2400 Loss 0.4880\n","Epoch 11 Batch 2500 Loss 0.5495\n","Epoch 11 Batch 2600 Loss 0.5727\n","Epoch 11 Batch 2700 Loss 0.5225\n","Epoch 11 Batch 2800 Loss 0.5248\n","Epoch 11 Batch 2900 Loss 0.5087\n","Epoch 11 Batch 3000 Loss 0.6249\n","Epoch 11 Batch 3100 Loss 0.6026\n","Epoch 11 Batch 3200 Loss 0.6054\n","Epoch 11 Batch 3300 Loss 0.5815\n","Epoch 11 Batch 3400 Loss 0.5806\n","Epoch 11 Batch 3500 Loss 0.5674\n","Epoch 11 Batch 3600 Loss 0.4981\n","Epoch 11 Batch 3700 Loss 0.5237\n","Epoch 11 Batch 3800 Loss 0.5371\n","Epoch 11 Batch 3900 Loss 0.5329\n","Epoch 11 Batch 4000 Loss 0.5908\n","Epoch 11 Batch 4100 Loss 0.6427\n","Epoch 11 Loss 0.562006\n","Time taken for 1 epoch 579.72 sec\n","\n","Epoch 12 Batch 0 Loss 0.5507\n","Epoch 12 Batch 100 Loss 0.5512\n","Epoch 12 Batch 200 Loss 0.5723\n","Epoch 12 Batch 300 Loss 0.5835\n","Epoch 12 Batch 400 Loss 0.5212\n","Epoch 12 Batch 500 Loss 0.4896\n","Epoch 12 Batch 600 Loss 0.5341\n","Epoch 12 Batch 700 Loss 0.5005\n","Epoch 12 Batch 800 Loss 0.5992\n","Epoch 12 Batch 900 Loss 0.5095\n","Epoch 12 Batch 1000 Loss 0.5339\n","Epoch 12 Batch 1100 Loss 0.5252\n","Epoch 12 Batch 1200 Loss 0.5066\n","Epoch 12 Batch 1300 Loss 0.5880\n","Epoch 12 Batch 1400 Loss 0.5165\n","Epoch 12 Batch 1500 Loss 0.5392\n","Epoch 12 Batch 1600 Loss 0.5583\n","Epoch 12 Batch 1700 Loss 0.4882\n","Epoch 12 Batch 1800 Loss 0.5257\n","Epoch 12 Batch 1900 Loss 0.5588\n","Epoch 12 Batch 2000 Loss 0.5177\n","Epoch 12 Batch 2100 Loss 0.5549\n","Epoch 12 Batch 2200 Loss 0.5364\n","Epoch 12 Batch 2300 Loss 0.5707\n","Epoch 12 Batch 2400 Loss 0.5998\n","Epoch 12 Batch 2500 Loss 0.5658\n","Epoch 12 Batch 2600 Loss 0.5029\n","Epoch 12 Batch 2700 Loss 0.5194\n","Epoch 12 Batch 2800 Loss 0.5307\n","Epoch 12 Batch 2900 Loss 0.5663\n","Epoch 12 Batch 3000 Loss 0.5822\n","Epoch 12 Batch 3100 Loss 0.5003\n","Epoch 12 Batch 3200 Loss 0.5780\n","Epoch 12 Batch 3300 Loss 0.5824\n","Epoch 12 Batch 3400 Loss 0.5333\n","Epoch 12 Batch 3500 Loss 0.5964\n","Epoch 12 Batch 3600 Loss 0.5271\n","Epoch 12 Batch 3700 Loss 0.5280\n","Epoch 12 Batch 3800 Loss 0.5418\n","Epoch 12 Batch 3900 Loss 0.5371\n","Epoch 12 Batch 4000 Loss 0.5482\n","Epoch 12 Batch 4100 Loss 0.5300\n","Epoch 12 Loss 0.548750\n","Time taken for 1 epoch 578.64 sec\n","\n","Epoch 13 Batch 0 Loss 0.5417\n","Epoch 13 Batch 100 Loss 0.5336\n","Epoch 13 Batch 200 Loss 0.5451\n","Epoch 13 Batch 300 Loss 0.5839\n","Epoch 13 Batch 400 Loss 0.5400\n","Epoch 13 Batch 500 Loss 0.5184\n","Epoch 13 Batch 600 Loss 0.5733\n","Epoch 13 Batch 700 Loss 0.5263\n","Epoch 13 Batch 800 Loss 0.5286\n","Epoch 13 Batch 900 Loss 0.6687\n","Epoch 13 Batch 1000 Loss 0.4963\n","Epoch 13 Batch 1100 Loss 0.4835\n","Epoch 13 Batch 1200 Loss 0.5517\n","Epoch 13 Batch 1300 Loss 0.5148\n","Epoch 13 Batch 1400 Loss 0.5735\n","Epoch 13 Batch 1500 Loss 0.5387\n","Epoch 13 Batch 1600 Loss 0.5585\n","Epoch 13 Batch 1700 Loss 0.5563\n","Epoch 13 Batch 1800 Loss 0.5366\n","Epoch 13 Batch 1900 Loss 0.5887\n","Epoch 13 Batch 2000 Loss 0.5052\n","Epoch 13 Batch 2100 Loss 0.5460\n","Epoch 13 Batch 2200 Loss 0.5324\n","Epoch 13 Batch 2300 Loss 0.5772\n","Epoch 13 Batch 2400 Loss 0.5413\n","Epoch 13 Batch 2500 Loss 0.5588\n","Epoch 13 Batch 2600 Loss 0.4894\n","Epoch 13 Batch 2700 Loss 0.4946\n","Epoch 13 Batch 2800 Loss 0.6071\n","Epoch 13 Batch 2900 Loss 0.5447\n","Epoch 13 Batch 3000 Loss 0.5575\n","Epoch 13 Batch 3100 Loss 0.5079\n","Epoch 13 Batch 3200 Loss 0.5702\n","Epoch 13 Batch 3300 Loss 0.5453\n","Epoch 13 Batch 3400 Loss 0.5252\n","Epoch 13 Batch 3500 Loss 0.5559\n","Epoch 13 Batch 3600 Loss 0.5846\n","Epoch 13 Batch 3700 Loss 0.5179\n","Epoch 13 Batch 3800 Loss 0.5580\n","Epoch 13 Batch 3900 Loss 0.4959\n","Epoch 13 Batch 4000 Loss 0.5054\n","Epoch 13 Batch 4100 Loss 0.5391\n","Epoch 13 Loss 0.536779\n","Time taken for 1 epoch 579.02 sec\n","\n","Epoch 14 Batch 0 Loss 0.5504\n","Epoch 14 Batch 100 Loss 0.4991\n","Epoch 14 Batch 200 Loss 0.5787\n","Epoch 14 Batch 300 Loss 0.5175\n","Epoch 14 Batch 400 Loss 0.5197\n","Epoch 14 Batch 500 Loss 0.5036\n","Epoch 14 Batch 600 Loss 0.5439\n","Epoch 14 Batch 700 Loss 0.5012\n","Epoch 14 Batch 800 Loss 0.4971\n","Epoch 14 Batch 900 Loss 0.5553\n","Epoch 14 Batch 1000 Loss 0.5414\n","Epoch 14 Batch 1100 Loss 0.5608\n","Epoch 14 Batch 1200 Loss 0.5107\n","Epoch 14 Batch 1300 Loss 0.5563\n","Epoch 14 Batch 1400 Loss 0.5137\n","Epoch 14 Batch 1500 Loss 0.5621\n","Epoch 14 Batch 1600 Loss 0.5123\n","Epoch 14 Batch 1700 Loss 0.5145\n","Epoch 14 Batch 1800 Loss 0.5550\n","Epoch 14 Batch 1900 Loss 0.4758\n","Epoch 14 Batch 2000 Loss 0.5426\n","Epoch 14 Batch 2100 Loss 0.5360\n","Epoch 14 Batch 2200 Loss 0.5365\n","Epoch 14 Batch 2300 Loss 0.4895\n","Epoch 14 Batch 2400 Loss 0.5026\n","Epoch 14 Batch 2500 Loss 0.5414\n","Epoch 14 Batch 2600 Loss 0.5533\n","Epoch 14 Batch 2700 Loss 0.5576\n","Epoch 14 Batch 2800 Loss 0.4988\n","Epoch 14 Batch 2900 Loss 0.5054\n","Epoch 14 Batch 3000 Loss 0.5501\n","Epoch 14 Batch 3100 Loss 0.5425\n","Epoch 14 Batch 3200 Loss 0.5081\n","Epoch 14 Batch 3300 Loss 0.4832\n","Epoch 14 Batch 3400 Loss 0.5216\n","Epoch 14 Batch 3500 Loss 0.5224\n","Epoch 14 Batch 3600 Loss 0.5342\n","Epoch 14 Batch 3700 Loss 0.5055\n","Epoch 14 Batch 3800 Loss 0.4866\n","Epoch 14 Batch 3900 Loss 0.4926\n","Epoch 14 Batch 4000 Loss 0.5937\n","Epoch 14 Batch 4100 Loss 0.5356\n","Epoch 14 Loss 0.526299\n","Time taken for 1 epoch 578.88 sec\n","\n","Epoch 15 Batch 0 Loss 0.5462\n","Epoch 15 Batch 100 Loss 0.5202\n","Epoch 15 Batch 200 Loss 0.5149\n","Epoch 15 Batch 300 Loss 0.5306\n","Epoch 15 Batch 400 Loss 0.4566\n","Epoch 15 Batch 500 Loss 0.4793\n","Epoch 15 Batch 600 Loss 0.5054\n","Epoch 15 Batch 700 Loss 0.5351\n","Epoch 15 Batch 800 Loss 0.4898\n","Epoch 15 Batch 900 Loss 0.5333\n","Epoch 15 Batch 1000 Loss 0.5290\n","Epoch 15 Batch 1100 Loss 0.5147\n","Epoch 15 Batch 1200 Loss 0.5194\n","Epoch 15 Batch 1300 Loss 0.4895\n","Epoch 15 Batch 1400 Loss 0.4793\n","Epoch 15 Batch 1500 Loss 0.5438\n","Epoch 15 Batch 1600 Loss 0.4849\n","Epoch 15 Batch 1700 Loss 0.5131\n","Epoch 15 Batch 1800 Loss 0.5028\n","Epoch 15 Batch 1900 Loss 0.5571\n","Epoch 15 Batch 2000 Loss 0.5798\n","Epoch 15 Batch 2100 Loss 0.4905\n","Epoch 15 Batch 2200 Loss 0.5460\n","Epoch 15 Batch 2300 Loss 0.5651\n","Epoch 15 Batch 2400 Loss 0.4978\n","Epoch 15 Batch 2500 Loss 0.4991\n","Epoch 15 Batch 2600 Loss 0.5087\n","Epoch 15 Batch 2700 Loss 0.4653\n","Epoch 15 Batch 2800 Loss 0.5474\n","Epoch 15 Batch 2900 Loss 0.5208\n","Epoch 15 Batch 3000 Loss 0.5454\n","Epoch 15 Batch 3100 Loss 0.5121\n","Epoch 15 Batch 3200 Loss 0.4799\n","Epoch 15 Batch 3300 Loss 0.5137\n","Epoch 15 Batch 3400 Loss 0.5113\n","Epoch 15 Batch 3500 Loss 0.5118\n","Epoch 15 Batch 3600 Loss 0.4908\n","Epoch 15 Batch 3700 Loss 0.4718\n","Epoch 15 Batch 3800 Loss 0.5572\n","Epoch 15 Batch 3900 Loss 0.5187\n","Epoch 15 Batch 4000 Loss 0.5283\n","Epoch 15 Batch 4100 Loss 0.5098\n","Epoch 15 Loss 0.516754\n","Time taken for 1 epoch 579.14 sec\n","\n","Epoch 16 Batch 0 Loss 0.5602\n","Epoch 16 Batch 100 Loss 0.5243\n","Epoch 16 Batch 200 Loss 0.5338\n","Epoch 16 Batch 300 Loss 0.5301\n","Epoch 16 Batch 400 Loss 0.5298\n","Epoch 16 Batch 500 Loss 0.4574\n","Epoch 16 Batch 600 Loss 0.5141\n","Epoch 16 Batch 700 Loss 0.4791\n","Epoch 16 Batch 800 Loss 0.5091\n","Epoch 16 Batch 900 Loss 0.4896\n","Epoch 16 Batch 1000 Loss 0.5387\n","Epoch 16 Batch 1100 Loss 0.5203\n","Epoch 16 Batch 1200 Loss 0.5069\n","Epoch 16 Batch 1300 Loss 0.5158\n","Epoch 16 Batch 1400 Loss 0.4475\n","Epoch 16 Batch 1500 Loss 0.4537\n","Epoch 16 Batch 1600 Loss 0.4853\n","Epoch 16 Batch 1700 Loss 0.4480\n","Epoch 16 Batch 1800 Loss 0.4520\n","Epoch 16 Batch 1900 Loss 0.4396\n","Epoch 16 Batch 2000 Loss 0.5004\n","Epoch 16 Batch 2100 Loss 0.5023\n","Epoch 16 Batch 2200 Loss 0.5644\n","Epoch 16 Batch 2300 Loss 0.4673\n","Epoch 16 Batch 2400 Loss 0.4918\n","Epoch 16 Batch 2500 Loss 0.4549\n","Epoch 16 Batch 2600 Loss 0.4520\n","Epoch 16 Batch 2700 Loss 0.5086\n","Epoch 16 Batch 2800 Loss 0.4984\n","Epoch 16 Batch 2900 Loss 0.4826\n","Epoch 16 Batch 3000 Loss 0.5376\n","Epoch 16 Batch 3100 Loss 0.4715\n","Epoch 16 Batch 3200 Loss 0.5404\n","Epoch 16 Batch 3300 Loss 0.5140\n","Epoch 16 Batch 3400 Loss 0.5588\n","Epoch 16 Batch 3500 Loss 0.4825\n","Epoch 16 Batch 3600 Loss 0.4645\n","Epoch 16 Batch 3700 Loss 0.5139\n","Epoch 16 Batch 3800 Loss 0.4797\n","Epoch 16 Batch 3900 Loss 0.4530\n","Epoch 16 Batch 4000 Loss 0.5173\n","Epoch 16 Batch 4100 Loss 0.5043\n","Epoch 16 Loss 0.508830\n","Time taken for 1 epoch 622.79 sec\n","\n","Epoch 17 Batch 0 Loss 0.4879\n","Epoch 17 Batch 100 Loss 0.4816\n","Epoch 17 Batch 200 Loss 0.4927\n","Epoch 17 Batch 300 Loss 0.5193\n","Epoch 17 Batch 400 Loss 0.4687\n","Epoch 17 Batch 500 Loss 0.4985\n","Epoch 17 Batch 600 Loss 0.4571\n","Epoch 17 Batch 700 Loss 0.4855\n","Epoch 17 Batch 800 Loss 0.4727\n","Epoch 17 Batch 900 Loss 0.5028\n","Epoch 17 Batch 1000 Loss 0.4807\n","Epoch 17 Batch 1100 Loss 0.4907\n","Epoch 17 Batch 1200 Loss 0.4957\n","Epoch 17 Batch 1300 Loss 0.5455\n","Epoch 17 Batch 1400 Loss 0.5098\n","Epoch 17 Batch 1500 Loss 0.5407\n","Epoch 17 Batch 1600 Loss 0.5248\n","Epoch 17 Batch 1700 Loss 0.5328\n","Epoch 17 Batch 1800 Loss 0.4857\n","Epoch 17 Batch 1900 Loss 0.4934\n","Epoch 17 Batch 2000 Loss 0.5191\n","Epoch 17 Batch 2100 Loss 0.5270\n","Epoch 17 Batch 2200 Loss 0.4624\n","Epoch 17 Batch 2300 Loss 0.5311\n","Epoch 17 Batch 2400 Loss 0.4521\n","Epoch 17 Batch 2500 Loss 0.4831\n","Epoch 17 Batch 2600 Loss 0.4995\n","Epoch 17 Batch 2700 Loss 0.4917\n","Epoch 17 Batch 2800 Loss 0.5254\n","Epoch 17 Batch 2900 Loss 0.4978\n","Epoch 17 Batch 3000 Loss 0.5172\n","Epoch 17 Batch 3100 Loss 0.4588\n","Epoch 17 Batch 3200 Loss 0.5144\n","Epoch 17 Batch 3300 Loss 0.5654\n","Epoch 17 Batch 3400 Loss 0.4317\n","Epoch 17 Batch 3500 Loss 0.4918\n","Epoch 17 Batch 3600 Loss 0.5086\n","Epoch 17 Batch 3700 Loss 0.4426\n","Epoch 17 Batch 3800 Loss 0.4565\n","Epoch 17 Batch 3900 Loss 0.4973\n","Epoch 17 Batch 4000 Loss 0.4644\n","Epoch 17 Batch 4100 Loss 0.5413\n","Epoch 17 Loss 0.500390\n","Time taken for 1 epoch 579.43 sec\n","\n","Epoch 18 Batch 0 Loss 0.5007\n","Epoch 18 Batch 100 Loss 0.4937\n","Epoch 18 Batch 200 Loss 0.4968\n","Epoch 18 Batch 300 Loss 0.5276\n","Epoch 18 Batch 400 Loss 0.4870\n","Epoch 18 Batch 500 Loss 0.4948\n","Epoch 18 Batch 600 Loss 0.5186\n","Epoch 18 Batch 700 Loss 0.4855\n","Epoch 18 Batch 800 Loss 0.4722\n","Epoch 18 Batch 900 Loss 0.4958\n","Epoch 18 Batch 1000 Loss 0.5230\n","Epoch 18 Batch 1100 Loss 0.4853\n","Epoch 18 Batch 1200 Loss 0.4747\n","Epoch 18 Batch 1300 Loss 0.4853\n","Epoch 18 Batch 1400 Loss 0.5026\n","Epoch 18 Batch 1500 Loss 0.5279\n","Epoch 18 Batch 1600 Loss 0.5466\n","Epoch 18 Batch 1700 Loss 0.5684\n","Epoch 18 Batch 1800 Loss 0.4730\n","Epoch 18 Batch 1900 Loss 0.4948\n","Epoch 18 Batch 2000 Loss 0.4700\n","Epoch 18 Batch 2100 Loss 0.4596\n","Epoch 18 Batch 2200 Loss 0.4606\n","Epoch 18 Batch 2300 Loss 0.4636\n","Epoch 18 Batch 2400 Loss 0.4723\n","Epoch 18 Batch 2500 Loss 0.5183\n","Epoch 18 Batch 2600 Loss 0.4895\n","Epoch 18 Batch 2700 Loss 0.5240\n","Epoch 18 Batch 2800 Loss 0.4589\n","Epoch 18 Batch 2900 Loss 0.4811\n","Epoch 18 Batch 3000 Loss 0.5419\n","Epoch 18 Batch 3100 Loss 0.5412\n","Epoch 18 Batch 3200 Loss 0.4820\n","Epoch 18 Batch 3300 Loss 0.5048\n","Epoch 18 Batch 3400 Loss 0.5411\n","Epoch 18 Batch 3500 Loss 0.4572\n","Epoch 18 Batch 3600 Loss 0.5076\n","Epoch 18 Batch 3700 Loss 0.4999\n","Epoch 18 Batch 3800 Loss 0.4783\n","Epoch 18 Batch 3900 Loss 0.5014\n","Epoch 18 Batch 4000 Loss 0.5308\n","Epoch 18 Batch 4100 Loss 0.4807\n","Epoch 18 Loss 0.493450\n","Time taken for 1 epoch 579.45 sec\n","\n","Epoch 19 Batch 0 Loss 0.4904\n","Epoch 19 Batch 100 Loss 0.4794\n","Epoch 19 Batch 200 Loss 0.4742\n","Epoch 19 Batch 300 Loss 0.5166\n","Epoch 19 Batch 400 Loss 0.4797\n","Epoch 19 Batch 500 Loss 0.4877\n","Epoch 19 Batch 600 Loss 0.4635\n","Epoch 19 Batch 700 Loss 0.4810\n","Epoch 19 Batch 800 Loss 0.4644\n","Epoch 19 Batch 900 Loss 0.5620\n","Epoch 19 Batch 1000 Loss 0.4744\n","Epoch 19 Batch 1100 Loss 0.5279\n","Epoch 19 Batch 1200 Loss 0.5172\n","Epoch 19 Batch 1300 Loss 0.5249\n","Epoch 19 Batch 1400 Loss 0.5049\n","Epoch 19 Batch 1500 Loss 0.5318\n","Epoch 19 Batch 1600 Loss 0.4560\n","Epoch 19 Batch 1700 Loss 0.4533\n","Epoch 19 Batch 1800 Loss 0.4590\n","Epoch 19 Batch 1900 Loss 0.5245\n","Epoch 19 Batch 2000 Loss 0.4793\n","Epoch 19 Batch 2100 Loss 0.4615\n","Epoch 19 Batch 2200 Loss 0.4887\n","Epoch 19 Batch 2300 Loss 0.4621\n","Epoch 19 Batch 2400 Loss 0.4429\n","Epoch 19 Batch 2500 Loss 0.4573\n","Epoch 19 Batch 2600 Loss 0.4711\n","Epoch 19 Batch 2700 Loss 0.4958\n","Epoch 19 Batch 2800 Loss 0.4516\n","Epoch 19 Batch 2900 Loss 0.4974\n","Epoch 19 Batch 3000 Loss 0.4705\n","Epoch 19 Batch 3100 Loss 0.4446\n","Epoch 19 Batch 3200 Loss 0.4722\n","Epoch 19 Batch 3300 Loss 0.5243\n","Epoch 19 Batch 3400 Loss 0.4623\n","Epoch 19 Batch 3500 Loss 0.4282\n","Epoch 19 Batch 3600 Loss 0.4539\n","Epoch 19 Batch 3700 Loss 0.5179\n","Epoch 19 Batch 3800 Loss 0.5112\n","Epoch 19 Batch 3900 Loss 0.5254\n","Epoch 19 Batch 4000 Loss 0.4942\n","Epoch 19 Batch 4100 Loss 0.5250\n","Epoch 19 Loss 0.486790\n","Time taken for 1 epoch 579.32 sec\n","\n","Epoch 20 Batch 0 Loss 0.4427\n","Epoch 20 Batch 100 Loss 0.4246\n","Epoch 20 Batch 200 Loss 0.4967\n","Epoch 20 Batch 300 Loss 0.5018\n","Epoch 20 Batch 400 Loss 0.4691\n","Epoch 20 Batch 500 Loss 0.4652\n","Epoch 20 Batch 600 Loss 0.5059\n","Epoch 20 Batch 700 Loss 0.4876\n","Epoch 20 Batch 800 Loss 0.4606\n","Epoch 20 Batch 900 Loss 0.5225\n","Epoch 20 Batch 1000 Loss 0.4633\n","Epoch 20 Batch 1100 Loss 0.4482\n","Epoch 20 Batch 1200 Loss 0.5220\n","Epoch 20 Batch 1300 Loss 0.4421\n","Epoch 20 Batch 1400 Loss 0.4639\n","Epoch 20 Batch 1500 Loss 0.4640\n","Epoch 20 Batch 1600 Loss 0.5268\n","Epoch 20 Batch 1700 Loss 0.4575\n","Epoch 20 Batch 1800 Loss 0.4765\n","Epoch 20 Batch 1900 Loss 0.4368\n","Epoch 20 Batch 2000 Loss 0.4265\n","Epoch 20 Batch 2100 Loss 0.5147\n","Epoch 20 Batch 2200 Loss 0.4694\n","Epoch 20 Batch 2300 Loss 0.4828\n","Epoch 20 Batch 2400 Loss 0.4582\n","Epoch 20 Batch 2500 Loss 0.4604\n","Epoch 20 Batch 2600 Loss 0.4160\n","Epoch 20 Batch 2700 Loss 0.4756\n","Epoch 20 Batch 2800 Loss 0.4766\n","Epoch 20 Batch 2900 Loss 0.4853\n","Epoch 20 Batch 3000 Loss 0.4436\n","Epoch 20 Batch 3100 Loss 0.4346\n","Epoch 20 Batch 3200 Loss 0.4341\n","Epoch 20 Batch 3300 Loss 0.5265\n","Epoch 20 Batch 3400 Loss 0.4888\n","Epoch 20 Batch 3500 Loss 0.4517\n","Epoch 20 Batch 3600 Loss 0.5022\n","Epoch 20 Batch 3700 Loss 0.4753\n","Epoch 20 Batch 3800 Loss 0.5022\n","Epoch 20 Batch 3900 Loss 0.4771\n","Epoch 20 Batch 4000 Loss 0.5099\n","Epoch 20 Batch 4100 Loss 0.4561\n","Epoch 20 Loss 0.480867\n","Time taken for 1 epoch 579.12 sec\n","\n","Epoch 21 Batch 0 Loss 0.5169\n","Epoch 21 Batch 100 Loss 0.5179\n","Epoch 21 Batch 200 Loss 0.4929\n","Epoch 21 Batch 300 Loss 0.4828\n","Epoch 21 Batch 400 Loss 0.4674\n","Epoch 21 Batch 500 Loss 0.4551\n","Epoch 21 Batch 600 Loss 0.4452\n","Epoch 21 Batch 700 Loss 0.4621\n","Epoch 21 Batch 800 Loss 0.5024\n","Epoch 21 Batch 900 Loss 0.4750\n","Epoch 21 Batch 1000 Loss 0.4192\n","Epoch 21 Batch 1100 Loss 0.4301\n","Epoch 21 Batch 1200 Loss 0.5184\n","Epoch 21 Batch 1300 Loss 0.4534\n","Epoch 21 Batch 1400 Loss 0.4069\n","Epoch 21 Batch 1500 Loss 0.4504\n","Epoch 21 Batch 1600 Loss 0.4827\n","Epoch 21 Batch 1700 Loss 0.4926\n","Epoch 21 Batch 1800 Loss 0.4806\n","Epoch 21 Batch 1900 Loss 0.4881\n","Epoch 21 Batch 2000 Loss 0.4945\n","Epoch 21 Batch 2100 Loss 0.4549\n","Epoch 21 Batch 2200 Loss 0.4735\n","Epoch 21 Batch 2300 Loss 0.4878\n","Epoch 21 Batch 2400 Loss 0.4488\n","Epoch 21 Batch 2500 Loss 0.4645\n","Epoch 21 Batch 2600 Loss 0.4465\n","Epoch 21 Batch 2700 Loss 0.4784\n","Epoch 21 Batch 2800 Loss 0.4699\n","Epoch 21 Batch 2900 Loss 0.5516\n","Epoch 21 Batch 3000 Loss 0.4384\n","Epoch 21 Batch 3100 Loss 0.4215\n","Epoch 21 Batch 3200 Loss 0.4416\n","Epoch 21 Batch 3300 Loss 0.4907\n","Epoch 21 Batch 3400 Loss 0.4935\n","Epoch 21 Batch 3500 Loss 0.5324\n","Epoch 21 Batch 3600 Loss 0.4447\n","Epoch 21 Batch 3700 Loss 0.4635\n","Epoch 21 Batch 3800 Loss 0.4998\n","Epoch 21 Batch 3900 Loss 0.4543\n","Epoch 21 Batch 4000 Loss 0.4863\n","Epoch 21 Batch 4100 Loss 0.4634\n","Epoch 21 Loss 0.475713\n","Time taken for 1 epoch 579.36 sec\n","\n","Epoch 22 Batch 0 Loss 0.4528\n","Epoch 22 Batch 100 Loss 0.4399\n","Epoch 22 Batch 200 Loss 0.4898\n","Epoch 22 Batch 300 Loss 0.4762\n","Epoch 22 Batch 400 Loss 0.4798\n","Epoch 22 Batch 500 Loss 0.4823\n","Epoch 22 Batch 600 Loss 0.4742\n","Epoch 22 Batch 700 Loss 0.4986\n","Epoch 22 Batch 800 Loss 0.4728\n","Epoch 22 Batch 900 Loss 0.5016\n","Epoch 22 Batch 1000 Loss 0.4619\n","Epoch 22 Batch 1100 Loss 0.4836\n","Epoch 22 Batch 1200 Loss 0.4650\n","Epoch 22 Batch 1300 Loss 0.4492\n","Epoch 22 Batch 1400 Loss 0.5005\n","Epoch 22 Batch 1500 Loss 0.4272\n","Epoch 22 Batch 1600 Loss 0.4875\n","Epoch 22 Batch 1700 Loss 0.4549\n","Epoch 22 Batch 1800 Loss 0.4659\n","Epoch 22 Batch 1900 Loss 0.4891\n","Epoch 22 Batch 2000 Loss 0.4550\n","Epoch 22 Batch 2100 Loss 0.4859\n","Epoch 22 Batch 2200 Loss 0.4297\n","Epoch 22 Batch 2300 Loss 0.4738\n","Epoch 22 Batch 2400 Loss 0.4907\n","Epoch 22 Batch 2500 Loss 0.4653\n","Epoch 22 Batch 2600 Loss 0.4593\n","Epoch 22 Batch 2700 Loss 0.4612\n","Epoch 22 Batch 2800 Loss 0.4705\n","Epoch 22 Batch 2900 Loss 0.4676\n","Epoch 22 Batch 3000 Loss 0.4700\n","Epoch 22 Batch 3100 Loss 0.4726\n","Epoch 22 Batch 3200 Loss 0.4002\n","Epoch 22 Batch 3300 Loss 0.3998\n","Epoch 22 Batch 3400 Loss 0.4523\n","Epoch 22 Batch 3500 Loss 0.4871\n","Epoch 22 Batch 3600 Loss 0.4880\n","Epoch 22 Batch 3700 Loss 0.4407\n","Epoch 22 Batch 3800 Loss 0.4780\n","Epoch 22 Batch 3900 Loss 0.4714\n","Epoch 22 Batch 4000 Loss 0.4892\n","Epoch 22 Batch 4100 Loss 0.5146\n","Epoch 22 Loss 0.470395\n","Time taken for 1 epoch 579.83 sec\n","\n","Epoch 23 Batch 0 Loss 0.4443\n","Epoch 23 Batch 100 Loss 0.4538\n","Epoch 23 Batch 200 Loss 0.4989\n","Epoch 23 Batch 300 Loss 0.4355\n","Epoch 23 Batch 400 Loss 0.4418\n","Epoch 23 Batch 500 Loss 0.4288\n","Epoch 23 Batch 600 Loss 0.4318\n","Epoch 23 Batch 700 Loss 0.4405\n","Epoch 23 Batch 800 Loss 0.5102\n","Epoch 23 Batch 900 Loss 0.4907\n","Epoch 23 Batch 1000 Loss 0.4531\n","Epoch 23 Batch 1100 Loss 0.4453\n","Epoch 23 Batch 1200 Loss 0.4261\n","Epoch 23 Batch 1300 Loss 0.4386\n","Epoch 23 Batch 1400 Loss 0.4497\n","Epoch 23 Batch 1500 Loss 0.4422\n","Epoch 23 Batch 1600 Loss 0.4725\n","Epoch 23 Batch 1700 Loss 0.4872\n","Epoch 23 Batch 1800 Loss 0.4645\n","Epoch 23 Batch 1900 Loss 0.4583\n","Epoch 23 Batch 2000 Loss 0.4617\n","Epoch 23 Batch 2100 Loss 0.4769\n","Epoch 23 Batch 2200 Loss 0.4720\n","Epoch 23 Batch 2300 Loss 0.4326\n","Epoch 23 Batch 2400 Loss 0.5065\n","Epoch 23 Batch 2500 Loss 0.4541\n","Epoch 23 Batch 2600 Loss 0.4817\n","Epoch 23 Batch 2700 Loss 0.4892\n","Epoch 23 Batch 2800 Loss 0.4672\n","Epoch 23 Batch 2900 Loss 0.4344\n","Epoch 23 Batch 3000 Loss 0.4859\n","Epoch 23 Batch 3100 Loss 0.4709\n","Epoch 23 Batch 3200 Loss 0.4194\n","Epoch 23 Batch 3300 Loss 0.4778\n","Epoch 23 Batch 3400 Loss 0.4876\n","Epoch 23 Batch 3500 Loss 0.4370\n","Epoch 23 Batch 3600 Loss 0.4868\n","Epoch 23 Batch 3700 Loss 0.4649\n","Epoch 23 Batch 3800 Loss 0.4695\n","Epoch 23 Batch 3900 Loss 0.4588\n","Epoch 23 Batch 4000 Loss 0.4772\n","Epoch 23 Batch 4100 Loss 0.4616\n","Epoch 23 Loss 0.466175\n","Time taken for 1 epoch 579.98 sec\n","\n","Epoch 24 Batch 0 Loss 0.4852\n","Epoch 24 Batch 100 Loss 0.4716\n","Epoch 24 Batch 200 Loss 0.4871\n","Epoch 24 Batch 300 Loss 0.4472\n","Epoch 24 Batch 400 Loss 0.4633\n","Epoch 24 Batch 500 Loss 0.4189\n","Epoch 24 Batch 600 Loss 0.4461\n","Epoch 24 Batch 700 Loss 0.4577\n","Epoch 24 Batch 800 Loss 0.4337\n","Epoch 24 Batch 900 Loss 0.5388\n","Epoch 24 Batch 1000 Loss 0.5021\n","Epoch 24 Batch 1100 Loss 0.4732\n","Epoch 24 Batch 1200 Loss 0.4636\n","Epoch 24 Batch 1300 Loss 0.5396\n","Epoch 24 Batch 1400 Loss 0.4747\n","Epoch 24 Batch 1500 Loss 0.4507\n","Epoch 24 Batch 1600 Loss 0.4647\n","Epoch 24 Batch 1700 Loss 0.4370\n","Epoch 24 Batch 1800 Loss 0.4826\n","Epoch 24 Batch 1900 Loss 0.4461\n","Epoch 24 Batch 2000 Loss 0.4581\n","Epoch 24 Batch 2100 Loss 0.4733\n","Epoch 24 Batch 2200 Loss 0.4840\n","Epoch 24 Batch 2300 Loss 0.5048\n","Epoch 24 Batch 2400 Loss 0.4531\n","Epoch 24 Batch 2500 Loss 0.4349\n","Epoch 24 Batch 2600 Loss 0.4507\n","Epoch 24 Batch 2700 Loss 0.4464\n","Epoch 24 Batch 2800 Loss 0.4890\n","Epoch 24 Batch 2900 Loss 0.4732\n","Epoch 24 Batch 3000 Loss 0.4434\n","Epoch 24 Batch 3100 Loss 0.4184\n","Epoch 24 Batch 3200 Loss 0.4439\n","Epoch 24 Batch 3300 Loss 0.4705\n","Epoch 24 Batch 3400 Loss 0.4527\n","Epoch 24 Batch 3500 Loss 0.4274\n","Epoch 24 Batch 3600 Loss 0.4596\n","Epoch 24 Batch 3700 Loss 0.4083\n","Epoch 24 Batch 3800 Loss 0.4719\n","Epoch 24 Batch 3900 Loss 0.4852\n","Epoch 24 Batch 4000 Loss 0.4615\n","Epoch 24 Batch 4100 Loss 0.4661\n","Epoch 24 Loss 0.462076\n","Time taken for 1 epoch 580.00 sec\n","\n","Epoch 25 Batch 0 Loss 0.4700\n","Epoch 25 Batch 100 Loss 0.4962\n","Epoch 25 Batch 200 Loss 0.4123\n","Epoch 25 Batch 300 Loss 0.5462\n","Epoch 25 Batch 400 Loss 0.4316\n","Epoch 25 Batch 500 Loss 0.4707\n","Epoch 25 Batch 600 Loss 0.4652\n","Epoch 25 Batch 700 Loss 0.4697\n","Epoch 25 Batch 800 Loss 0.4236\n","Epoch 25 Batch 900 Loss 0.4745\n","Epoch 25 Batch 1000 Loss 0.4672\n","Epoch 25 Batch 1100 Loss 0.4991\n","Epoch 25 Batch 1200 Loss 0.4195\n","Epoch 25 Batch 1300 Loss 0.4720\n","Epoch 25 Batch 1400 Loss 0.4282\n","Epoch 25 Batch 1500 Loss 0.4598\n","Epoch 25 Batch 1600 Loss 0.4704\n","Epoch 25 Batch 1700 Loss 0.4563\n","Epoch 25 Batch 1800 Loss 0.4575\n","Epoch 25 Batch 1900 Loss 0.4766\n","Epoch 25 Batch 2000 Loss 0.4547\n","Epoch 25 Batch 2100 Loss 0.4599\n","Epoch 25 Batch 2200 Loss 0.4583\n","Epoch 25 Batch 2300 Loss 0.4348\n","Epoch 25 Batch 2400 Loss 0.4551\n","Epoch 25 Batch 2500 Loss 0.3976\n","Epoch 25 Batch 2600 Loss 0.4629\n","Epoch 25 Batch 2700 Loss 0.4794\n","Epoch 25 Batch 2800 Loss 0.4495\n","Epoch 25 Batch 2900 Loss 0.4113\n","Epoch 25 Batch 3000 Loss 0.4728\n","Epoch 25 Batch 3100 Loss 0.4259\n","Epoch 25 Batch 3200 Loss 0.4682\n","Epoch 25 Batch 3300 Loss 0.4571\n","Epoch 25 Batch 3400 Loss 0.4397\n","Epoch 25 Batch 3500 Loss 0.4302\n","Epoch 25 Batch 3600 Loss 0.5315\n","Epoch 25 Batch 3700 Loss 0.4751\n","Epoch 25 Batch 3800 Loss 0.4588\n","Epoch 25 Batch 3900 Loss 0.4292\n","Epoch 25 Batch 4000 Loss 0.4658\n","Epoch 25 Batch 4100 Loss 0.4288\n","Epoch 25 Loss 0.457694\n","Time taken for 1 epoch 579.11 sec\n","\n","Epoch 26 Batch 0 Loss 0.4450\n","Epoch 26 Batch 100 Loss 0.4453\n","Epoch 26 Batch 200 Loss 0.4495\n","Epoch 26 Batch 300 Loss 0.4637\n","Epoch 26 Batch 400 Loss 0.4371\n","Epoch 26 Batch 500 Loss 0.4566\n","Epoch 26 Batch 600 Loss 0.4449\n","Epoch 26 Batch 700 Loss 0.3585\n","Epoch 26 Batch 800 Loss 0.4527\n","Epoch 26 Batch 900 Loss 0.4873\n","Epoch 26 Batch 1000 Loss 0.4221\n","Epoch 26 Batch 1100 Loss 0.4101\n","Epoch 26 Batch 1200 Loss 0.4406\n","Epoch 26 Batch 1300 Loss 0.4427\n","Epoch 26 Batch 1400 Loss 0.4453\n","Epoch 26 Batch 1500 Loss 0.4951\n","Epoch 26 Batch 1600 Loss 0.4583\n","Epoch 26 Batch 1700 Loss 0.4002\n","Epoch 26 Batch 1800 Loss 0.4312\n","Epoch 26 Batch 1900 Loss 0.4380\n","Epoch 26 Batch 2000 Loss 0.4322\n","Epoch 26 Batch 2100 Loss 0.4283\n","Epoch 26 Batch 2200 Loss 0.5258\n","Epoch 26 Batch 2300 Loss 0.4744\n","Epoch 26 Batch 2400 Loss 0.4627\n","Epoch 26 Batch 2500 Loss 0.4505\n","Epoch 26 Batch 2600 Loss 0.4779\n","Epoch 26 Batch 2700 Loss 0.4769\n","Epoch 26 Batch 2800 Loss 0.4960\n","Epoch 26 Batch 2900 Loss 0.4473\n","Epoch 26 Batch 3000 Loss 0.4253\n","Epoch 26 Batch 3100 Loss 0.4243\n","Epoch 26 Batch 3200 Loss 0.4515\n","Epoch 26 Batch 3300 Loss 0.4597\n","Epoch 26 Batch 3400 Loss 0.4298\n","Epoch 26 Batch 3500 Loss 0.4080\n","Epoch 26 Batch 3600 Loss 0.4722\n","Epoch 26 Batch 3700 Loss 0.4666\n","Epoch 26 Batch 3800 Loss 0.4678\n","Epoch 26 Batch 3900 Loss 0.4238\n","Epoch 26 Batch 4000 Loss 0.4767\n","Epoch 26 Batch 4100 Loss 0.4656\n","Epoch 26 Loss 0.454661\n","Time taken for 1 epoch 578.96 sec\n","\n","Epoch 27 Batch 0 Loss 0.4788\n","Epoch 27 Batch 100 Loss 0.3961\n","Epoch 27 Batch 200 Loss 0.4554\n","Epoch 27 Batch 300 Loss 0.4755\n","Epoch 27 Batch 400 Loss 0.4451\n","Epoch 27 Batch 500 Loss 0.3981\n","Epoch 27 Batch 600 Loss 0.4517\n","Epoch 27 Batch 700 Loss 0.4540\n","Epoch 27 Batch 800 Loss 0.4576\n","Epoch 27 Batch 900 Loss 0.4579\n","Epoch 27 Batch 1000 Loss 0.4436\n","Epoch 27 Batch 1100 Loss 0.4702\n","Epoch 27 Batch 1200 Loss 0.4717\n","Epoch 27 Batch 1300 Loss 0.4623\n","Epoch 27 Batch 1400 Loss 0.4399\n","Epoch 27 Batch 1500 Loss 0.4669\n","Epoch 27 Batch 1600 Loss 0.4746\n","Epoch 27 Batch 1700 Loss 0.4513\n","Epoch 27 Batch 1800 Loss 0.4762\n","Epoch 27 Batch 1900 Loss 0.4483\n","Epoch 27 Batch 2000 Loss 0.4661\n","Epoch 27 Batch 2100 Loss 0.4404\n","Epoch 27 Batch 2200 Loss 0.4869\n","Epoch 27 Batch 2300 Loss 0.4338\n","Epoch 27 Batch 2400 Loss 0.4967\n","Epoch 27 Batch 2500 Loss 0.4895\n","Epoch 27 Batch 2600 Loss 0.4817\n","Epoch 27 Batch 2700 Loss 0.4367\n","Epoch 27 Batch 2800 Loss 0.4468\n","Epoch 27 Batch 2900 Loss 0.4301\n","Epoch 27 Batch 3000 Loss 0.4614\n","Epoch 27 Batch 3100 Loss 0.4170\n","Epoch 27 Batch 3200 Loss 0.4142\n","Epoch 27 Batch 3300 Loss 0.4272\n","Epoch 27 Batch 3400 Loss 0.4520\n","Epoch 27 Batch 3500 Loss 0.4673\n","Epoch 27 Batch 3600 Loss 0.4335\n","Epoch 27 Batch 3700 Loss 0.4403\n","Epoch 27 Batch 3800 Loss 0.4900\n","Epoch 27 Batch 3900 Loss 0.4853\n","Epoch 27 Batch 4000 Loss 0.5213\n","Epoch 27 Batch 4100 Loss 0.4303\n","Epoch 27 Loss 0.451034\n","Time taken for 1 epoch 578.64 sec\n","\n","Epoch 28 Batch 0 Loss 0.4538\n","Epoch 28 Batch 100 Loss 0.4335\n","Epoch 28 Batch 200 Loss 0.4278\n","Epoch 28 Batch 300 Loss 0.4080\n","Epoch 28 Batch 400 Loss 0.4456\n","Epoch 28 Batch 500 Loss 0.4402\n","Epoch 28 Batch 600 Loss 0.4297\n","Epoch 28 Batch 700 Loss 0.4976\n","Epoch 28 Batch 800 Loss 0.4685\n","Epoch 28 Batch 900 Loss 0.4590\n","Epoch 28 Batch 1000 Loss 0.4415\n","Epoch 28 Batch 1100 Loss 0.4310\n","Epoch 28 Batch 1200 Loss 0.4582\n","Epoch 28 Batch 1300 Loss 0.4380\n","Epoch 28 Batch 1400 Loss 0.5093\n","Epoch 28 Batch 1500 Loss 0.4313\n","Epoch 28 Batch 1600 Loss 0.5066\n","Epoch 28 Batch 1700 Loss 0.5043\n","Epoch 28 Batch 1800 Loss 0.4248\n","Epoch 28 Batch 1900 Loss 0.4186\n","Epoch 28 Batch 2000 Loss 0.4642\n","Epoch 28 Batch 2100 Loss 0.4278\n","Epoch 28 Batch 2200 Loss 0.4496\n","Epoch 28 Batch 2300 Loss 0.3762\n","Epoch 28 Batch 2400 Loss 0.4332\n","Epoch 28 Batch 2500 Loss 0.4402\n","Epoch 28 Batch 2600 Loss 0.4986\n","Epoch 28 Batch 2700 Loss 0.5192\n","Epoch 28 Batch 2800 Loss 0.4216\n","Epoch 28 Batch 2900 Loss 0.4382\n","Epoch 28 Batch 3000 Loss 0.4854\n","Epoch 28 Batch 3100 Loss 0.4714\n","Epoch 28 Batch 3200 Loss 0.4209\n","Epoch 28 Batch 3300 Loss 0.4427\n","Epoch 28 Batch 3400 Loss 0.4579\n","Epoch 28 Batch 3500 Loss 0.4119\n","Epoch 28 Batch 3600 Loss 0.4420\n","Epoch 28 Batch 3700 Loss 0.3945\n","Epoch 28 Batch 3800 Loss 0.4591\n","Epoch 28 Batch 3900 Loss 0.4073\n","Epoch 28 Batch 4000 Loss 0.4635\n","Epoch 28 Batch 4100 Loss 0.4378\n","Epoch 28 Loss 0.447994\n","Time taken for 1 epoch 578.57 sec\n","\n","Epoch 29 Batch 0 Loss 0.4116\n","Epoch 29 Batch 100 Loss 0.4523\n","Epoch 29 Batch 200 Loss 0.4982\n","Epoch 29 Batch 300 Loss 0.4753\n","Epoch 29 Batch 400 Loss 0.4060\n","Epoch 29 Batch 500 Loss 0.3929\n","Epoch 29 Batch 600 Loss 0.4473\n","Epoch 29 Batch 700 Loss 0.4435\n","Epoch 29 Batch 800 Loss 0.4490\n","Epoch 29 Batch 900 Loss 0.4906\n","Epoch 29 Batch 1000 Loss 0.4403\n","Epoch 29 Batch 1100 Loss 0.4311\n","Epoch 29 Batch 1200 Loss 0.4289\n","Epoch 29 Batch 1300 Loss 0.4885\n","Epoch 29 Batch 1400 Loss 0.4340\n","Epoch 29 Batch 1500 Loss 0.4524\n","Epoch 29 Batch 1600 Loss 0.4110\n","Epoch 29 Batch 1700 Loss 0.3943\n","Epoch 29 Batch 1800 Loss 0.4824\n","Epoch 29 Batch 1900 Loss 0.4400\n","Epoch 29 Batch 2000 Loss 0.4254\n","Epoch 29 Batch 2100 Loss 0.3973\n","Epoch 29 Batch 2200 Loss 0.4290\n","Epoch 29 Batch 2300 Loss 0.4353\n","Epoch 29 Batch 2400 Loss 0.4532\n","Epoch 29 Batch 2500 Loss 0.4369\n","Epoch 29 Batch 2600 Loss 0.4507\n","Epoch 29 Batch 2700 Loss 0.4366\n","Epoch 29 Batch 2800 Loss 0.4854\n","Epoch 29 Batch 2900 Loss 0.4373\n","Epoch 29 Batch 3000 Loss 0.3980\n","Epoch 29 Batch 3100 Loss 0.4110\n","Epoch 29 Batch 3200 Loss 0.4395\n","Epoch 29 Batch 3300 Loss 0.4160\n","Epoch 29 Batch 3400 Loss 0.4721\n","Epoch 29 Batch 3500 Loss 0.4150\n","Epoch 29 Batch 3600 Loss 0.4210\n","Epoch 29 Batch 3700 Loss 0.4406\n","Epoch 29 Batch 3800 Loss 0.4277\n","Epoch 29 Batch 3900 Loss 0.4298\n","Epoch 29 Batch 4000 Loss 0.4450\n","Epoch 29 Batch 4100 Loss 0.4501\n","Epoch 29 Loss 0.444993\n","Time taken for 1 epoch 580.06 sec\n","\n","Epoch 30 Batch 0 Loss 0.3874\n","Epoch 30 Batch 100 Loss 0.4265\n","Epoch 30 Batch 200 Loss 0.4283\n","Epoch 30 Batch 300 Loss 0.4406\n","Epoch 30 Batch 400 Loss 0.4168\n","Epoch 30 Batch 500 Loss 0.4592\n","Epoch 30 Batch 600 Loss 0.4472\n","Epoch 30 Batch 700 Loss 0.4480\n","Epoch 30 Batch 800 Loss 0.4087\n","Epoch 30 Batch 900 Loss 0.4774\n","Epoch 30 Batch 1000 Loss 0.4242\n","Epoch 30 Batch 1100 Loss 0.4358\n","Epoch 30 Batch 1200 Loss 0.4131\n","Epoch 30 Batch 1300 Loss 0.5478\n","Epoch 30 Batch 1400 Loss 0.4314\n","Epoch 30 Batch 1500 Loss 0.4738\n","Epoch 30 Batch 1600 Loss 0.4790\n","Epoch 30 Batch 1700 Loss 0.4504\n","Epoch 30 Batch 1800 Loss 0.4408\n","Epoch 30 Batch 1900 Loss 0.4348\n","Epoch 30 Batch 2000 Loss 0.5254\n","Epoch 30 Batch 2100 Loss 0.5055\n","Epoch 30 Batch 2200 Loss 0.4594\n","Epoch 30 Batch 2300 Loss 0.3964\n","Epoch 30 Batch 2400 Loss 0.4321\n","Epoch 30 Batch 2500 Loss 0.4707\n","Epoch 30 Batch 2600 Loss 0.4239\n","Epoch 30 Batch 2700 Loss 0.4496\n","Epoch 30 Batch 2800 Loss 0.4727\n","Epoch 30 Batch 2900 Loss 0.4264\n","Epoch 30 Batch 3000 Loss 0.4430\n","Epoch 30 Batch 3100 Loss 0.4048\n","Epoch 30 Batch 3200 Loss 0.4630\n","Epoch 30 Batch 3300 Loss 0.4022\n","Epoch 30 Batch 3400 Loss 0.4186\n","Epoch 30 Batch 3500 Loss 0.4257\n","Epoch 30 Batch 3600 Loss 0.4275\n","Epoch 30 Batch 3700 Loss 0.4516\n","Epoch 30 Batch 3800 Loss 0.4632\n","Epoch 30 Batch 3900 Loss 0.4395\n","Epoch 30 Batch 4000 Loss 0.4524\n","Epoch 30 Batch 4100 Loss 0.4754\n","Epoch 30 Loss 0.442320\n","Time taken for 1 epoch 579.49 sec\n","\n","Epoch 31 Batch 0 Loss 0.4616\n","Epoch 31 Batch 100 Loss 0.4128\n","Epoch 31 Batch 200 Loss 0.4228\n","Epoch 31 Batch 300 Loss 0.4799\n","Epoch 31 Batch 400 Loss 0.4010\n","Epoch 31 Batch 500 Loss 0.4594\n","Epoch 31 Batch 600 Loss 0.4488\n","Epoch 31 Batch 700 Loss 0.4330\n","Epoch 31 Batch 800 Loss 0.4451\n","Epoch 31 Batch 900 Loss 0.4333\n","Epoch 31 Batch 1000 Loss 0.4373\n","Epoch 31 Batch 1100 Loss 0.3909\n","Epoch 31 Batch 1200 Loss 0.4526\n","Epoch 31 Batch 1300 Loss 0.4156\n","Epoch 31 Batch 1400 Loss 0.3970\n","Epoch 31 Batch 1500 Loss 0.4485\n","Epoch 31 Batch 1600 Loss 0.3737\n","Epoch 31 Batch 1700 Loss 0.4691\n","Epoch 31 Batch 1800 Loss 0.4445\n","Epoch 31 Batch 1900 Loss 0.4633\n","Epoch 31 Batch 2000 Loss 0.4628\n","Epoch 31 Batch 2100 Loss 0.4077\n","Epoch 31 Batch 2200 Loss 0.4114\n","Epoch 31 Batch 2300 Loss 0.4229\n","Epoch 31 Batch 2400 Loss 0.3928\n","Epoch 31 Batch 2500 Loss 0.4723\n","Epoch 31 Batch 2600 Loss 0.4119\n","Epoch 31 Batch 2700 Loss 0.4398\n","Epoch 31 Batch 2800 Loss 0.4429\n","Epoch 31 Batch 2900 Loss 0.4242\n","Epoch 31 Batch 3000 Loss 0.4428\n","Epoch 31 Batch 3100 Loss 0.4259\n","Epoch 31 Batch 3200 Loss 0.4763\n","Epoch 31 Batch 3300 Loss 0.4238\n","Epoch 31 Batch 3400 Loss 0.4723\n","Epoch 31 Batch 3500 Loss 0.4260\n","Epoch 31 Batch 3600 Loss 0.4530\n","Epoch 31 Batch 3700 Loss 0.4279\n","Epoch 31 Batch 3800 Loss 0.4033\n","Epoch 31 Batch 3900 Loss 0.3841\n","Epoch 31 Batch 4000 Loss 0.4519\n","Epoch 31 Batch 4100 Loss 0.4781\n","Epoch 31 Loss 0.440331\n","Time taken for 1 epoch 579.67 sec\n","\n","Epoch 32 Batch 0 Loss 0.4975\n","Epoch 32 Batch 100 Loss 0.3886\n","Epoch 32 Batch 200 Loss 0.4412\n","Epoch 32 Batch 300 Loss 0.4655\n","Epoch 32 Batch 400 Loss 0.4551\n","Epoch 32 Batch 500 Loss 0.4111\n","Epoch 32 Batch 600 Loss 0.4203\n","Epoch 32 Batch 700 Loss 0.4593\n","Epoch 32 Batch 800 Loss 0.4400\n","Epoch 32 Batch 900 Loss 0.4095\n","Epoch 32 Batch 1000 Loss 0.4186\n","Epoch 32 Batch 1100 Loss 0.4408\n","Epoch 32 Batch 1200 Loss 0.4587\n","Epoch 32 Batch 1300 Loss 0.4216\n","Epoch 32 Batch 1400 Loss 0.4170\n","Epoch 32 Batch 1500 Loss 0.4250\n","Epoch 32 Batch 1600 Loss 0.4739\n","Epoch 32 Batch 1700 Loss 0.4065\n","Epoch 32 Batch 1800 Loss 0.4270\n","Epoch 32 Batch 1900 Loss 0.4265\n","Epoch 32 Batch 2000 Loss 0.4352\n","Epoch 32 Batch 2100 Loss 0.4313\n","Epoch 32 Batch 2200 Loss 0.4240\n","Epoch 32 Batch 2300 Loss 0.3846\n","Epoch 32 Batch 2400 Loss 0.4193\n","Epoch 32 Batch 2500 Loss 0.4574\n","Epoch 32 Batch 2600 Loss 0.4063\n","Epoch 32 Batch 2700 Loss 0.4281\n","Epoch 32 Batch 2800 Loss 0.4722\n","Epoch 32 Batch 2900 Loss 0.4436\n","Epoch 32 Batch 3000 Loss 0.4400\n","Epoch 32 Batch 3100 Loss 0.4388\n","Epoch 32 Batch 3200 Loss 0.4918\n","Epoch 32 Batch 3300 Loss 0.4289\n","Epoch 32 Batch 3400 Loss 0.4347\n","Epoch 32 Batch 3500 Loss 0.4802\n","Epoch 32 Batch 3600 Loss 0.4325\n","Epoch 32 Batch 3700 Loss 0.3840\n","Epoch 32 Batch 3800 Loss 0.4436\n","Epoch 32 Batch 3900 Loss 0.4059\n","Epoch 32 Batch 4000 Loss 0.4446\n","Epoch 32 Batch 4100 Loss 0.4434\n","Epoch 32 Loss 0.437245\n","Time taken for 1 epoch 579.49 sec\n","\n","Epoch 33 Batch 0 Loss 0.3986\n","Epoch 33 Batch 100 Loss 0.4650\n","Epoch 33 Batch 200 Loss 0.4313\n","Epoch 33 Batch 300 Loss 0.4425\n","Epoch 33 Batch 400 Loss 0.4330\n","Epoch 33 Batch 500 Loss 0.4511\n","Epoch 33 Batch 600 Loss 0.4348\n","Epoch 33 Batch 700 Loss 0.4561\n","Epoch 33 Batch 800 Loss 0.4902\n","Epoch 33 Batch 900 Loss 0.4668\n","Epoch 33 Batch 1000 Loss 0.3832\n","Epoch 33 Batch 1100 Loss 0.4343\n","Epoch 33 Batch 1200 Loss 0.4566\n","Epoch 33 Batch 1300 Loss 0.4493\n","Epoch 33 Batch 1400 Loss 0.4121\n","Epoch 33 Batch 1500 Loss 0.4519\n","Epoch 33 Batch 1600 Loss 0.4333\n","Epoch 33 Batch 1700 Loss 0.4176\n","Epoch 33 Batch 1800 Loss 0.4153\n","Epoch 33 Batch 1900 Loss 0.4262\n","Epoch 33 Batch 2000 Loss 0.4610\n","Epoch 33 Batch 2100 Loss 0.4538\n","Epoch 33 Batch 2200 Loss 0.4178\n","Epoch 33 Batch 2300 Loss 0.4778\n","Epoch 33 Batch 2400 Loss 0.4247\n","Epoch 33 Batch 2500 Loss 0.4432\n","Epoch 33 Batch 2600 Loss 0.4514\n","Epoch 33 Batch 2700 Loss 0.4508\n","Epoch 33 Batch 2800 Loss 0.4293\n","Epoch 33 Batch 2900 Loss 0.4510\n","Epoch 33 Batch 3000 Loss 0.4594\n","Epoch 33 Batch 3100 Loss 0.4311\n","Epoch 33 Batch 3200 Loss 0.4687\n","Epoch 33 Batch 3300 Loss 0.4390\n","Epoch 33 Batch 3400 Loss 0.4278\n","Epoch 33 Batch 3500 Loss 0.4335\n","Epoch 33 Batch 3600 Loss 0.4127\n","Epoch 33 Batch 3700 Loss 0.4376\n","Epoch 33 Batch 3800 Loss 0.4423\n","Epoch 33 Batch 3900 Loss 0.3995\n","Epoch 33 Batch 4000 Loss 0.4518\n","Epoch 33 Batch 4100 Loss 0.4509\n","Epoch 33 Loss 0.435300\n","Time taken for 1 epoch 579.57 sec\n","\n","Epoch 34 Batch 0 Loss 0.4357\n","Epoch 34 Batch 100 Loss 0.4471\n","Epoch 34 Batch 200 Loss 0.4445\n","Epoch 34 Batch 300 Loss 0.4471\n","Epoch 34 Batch 400 Loss 0.4099\n","Epoch 34 Batch 500 Loss 0.4040\n","Epoch 34 Batch 600 Loss 0.4151\n","Epoch 34 Batch 700 Loss 0.4126\n","Epoch 34 Batch 800 Loss 0.4187\n","Epoch 34 Batch 900 Loss 0.4577\n","Epoch 34 Batch 1000 Loss 0.4271\n","Epoch 34 Batch 1100 Loss 0.4611\n","Epoch 34 Batch 1200 Loss 0.4733\n","Epoch 34 Batch 1300 Loss 0.4118\n","Epoch 34 Batch 1400 Loss 0.3946\n","Epoch 34 Batch 1500 Loss 0.4044\n","Epoch 34 Batch 1600 Loss 0.3959\n","Epoch 34 Batch 1700 Loss 0.4316\n","Epoch 34 Batch 1800 Loss 0.4027\n","Epoch 34 Batch 1900 Loss 0.4260\n","Epoch 34 Batch 2000 Loss 0.4433\n","Epoch 34 Batch 2100 Loss 0.4522\n","Epoch 34 Batch 2200 Loss 0.4492\n","Epoch 34 Batch 2300 Loss 0.4099\n","Epoch 34 Batch 2400 Loss 0.4449\n","Epoch 34 Batch 2500 Loss 0.3936\n","Epoch 34 Batch 2600 Loss 0.4444\n","Epoch 34 Batch 2700 Loss 0.4684\n","Epoch 34 Batch 2800 Loss 0.4876\n","Epoch 34 Batch 2900 Loss 0.4511\n","Epoch 34 Batch 3000 Loss 0.4434\n","Epoch 34 Batch 3100 Loss 0.4236\n","Epoch 34 Batch 3200 Loss 0.4266\n","Epoch 34 Batch 3300 Loss 0.4754\n","Epoch 34 Batch 3400 Loss 0.4290\n","Epoch 34 Batch 3500 Loss 0.4120\n","Epoch 34 Batch 3600 Loss 0.4275\n","Epoch 34 Batch 3700 Loss 0.4405\n","Epoch 34 Batch 3800 Loss 0.4158\n","Epoch 34 Batch 3900 Loss 0.4065\n","Epoch 34 Batch 4000 Loss 0.4377\n","Epoch 34 Batch 4100 Loss 0.4466\n","Epoch 34 Loss 0.432695\n","Time taken for 1 epoch 578.85 sec\n","\n","Epoch 35 Batch 0 Loss 0.3582\n","Epoch 35 Batch 100 Loss 0.3932\n","Epoch 35 Batch 200 Loss 0.4339\n","Epoch 35 Batch 300 Loss 0.4319\n","Epoch 35 Batch 400 Loss 0.4307\n","Epoch 35 Batch 500 Loss 0.3633\n","Epoch 35 Batch 600 Loss 0.4306\n","Epoch 35 Batch 700 Loss 0.4513\n","Epoch 35 Batch 800 Loss 0.4726\n","Epoch 35 Batch 900 Loss 0.4048\n","Epoch 35 Batch 1000 Loss 0.3417\n","Epoch 35 Batch 1100 Loss 0.4485\n","Epoch 35 Batch 1200 Loss 0.4213\n","Epoch 35 Batch 1300 Loss 0.4246\n","Epoch 35 Batch 1400 Loss 0.4316\n","Epoch 35 Batch 1500 Loss 0.4488\n","Epoch 35 Batch 1600 Loss 0.4194\n","Epoch 35 Batch 1700 Loss 0.4147\n","Epoch 35 Batch 1800 Loss 0.4405\n","Epoch 35 Batch 1900 Loss 0.4074\n","Epoch 35 Batch 2000 Loss 0.4235\n","Epoch 35 Batch 2100 Loss 0.4336\n","Epoch 35 Batch 2200 Loss 0.3731\n","Epoch 35 Batch 2300 Loss 0.4150\n","Epoch 35 Batch 2400 Loss 0.4358\n","Epoch 35 Batch 2500 Loss 0.4638\n","Epoch 35 Batch 2600 Loss 0.4350\n","Epoch 35 Batch 2700 Loss 0.4245\n","Epoch 35 Batch 2800 Loss 0.4864\n","Epoch 35 Batch 2900 Loss 0.4126\n","Epoch 35 Batch 3000 Loss 0.4889\n","Epoch 35 Batch 3100 Loss 0.4522\n","Epoch 35 Batch 3200 Loss 0.4107\n","Epoch 35 Batch 3300 Loss 0.4306\n","Epoch 35 Batch 3400 Loss 0.3862\n","Epoch 35 Batch 3500 Loss 0.4087\n","Epoch 35 Batch 3600 Loss 0.4097\n","Epoch 35 Batch 3700 Loss 0.4638\n","Epoch 35 Batch 3800 Loss 0.4595\n","Epoch 35 Batch 3900 Loss 0.4969\n","Epoch 35 Batch 4000 Loss 0.4633\n","Epoch 35 Batch 4100 Loss 0.4450\n","Epoch 35 Loss 0.431358\n","Time taken for 1 epoch 578.92 sec\n","\n","Epoch 36 Batch 0 Loss 0.4928\n","Epoch 36 Batch 100 Loss 0.4030\n","Epoch 36 Batch 200 Loss 0.4644\n","Epoch 36 Batch 300 Loss 0.4124\n","Epoch 36 Batch 400 Loss 0.3871\n","Epoch 36 Batch 500 Loss 0.4262\n","Epoch 36 Batch 600 Loss 0.4082\n","Epoch 36 Batch 700 Loss 0.4187\n","Epoch 36 Batch 800 Loss 0.4744\n","Epoch 36 Batch 900 Loss 0.4782\n","Epoch 36 Batch 1000 Loss 0.4247\n","Epoch 36 Batch 1100 Loss 0.3987\n","Epoch 36 Batch 1200 Loss 0.4023\n","Epoch 36 Batch 1300 Loss 0.4425\n","Epoch 36 Batch 1400 Loss 0.4017\n","Epoch 36 Batch 1500 Loss 0.4402\n","Epoch 36 Batch 1600 Loss 0.4118\n","Epoch 36 Batch 1700 Loss 0.4138\n","Epoch 36 Batch 1800 Loss 0.4343\n","Epoch 36 Batch 1900 Loss 0.4276\n","Epoch 36 Batch 2000 Loss 0.5032\n","Epoch 36 Batch 2100 Loss 0.4376\n","Epoch 36 Batch 2200 Loss 0.4217\n","Epoch 36 Batch 2300 Loss 0.4293\n","Epoch 36 Batch 2400 Loss 0.4187\n","Epoch 36 Batch 2500 Loss 0.4405\n","Epoch 36 Batch 2600 Loss 0.4593\n","Epoch 36 Batch 2700 Loss 0.3845\n","Epoch 36 Batch 2800 Loss 0.4535\n","Epoch 36 Batch 2900 Loss 0.4395\n","Epoch 36 Batch 3000 Loss 0.4189\n","Epoch 36 Batch 3100 Loss 0.3910\n","Epoch 36 Batch 3200 Loss 0.3970\n","Epoch 36 Batch 3300 Loss 0.4461\n","Epoch 36 Batch 3400 Loss 0.4136\n","Epoch 36 Batch 3500 Loss 0.3999\n","Epoch 36 Batch 3600 Loss 0.4576\n","Epoch 36 Batch 3700 Loss 0.4268\n","Epoch 36 Batch 3800 Loss 0.4946\n","Epoch 36 Batch 3900 Loss 0.4651\n","Epoch 36 Batch 4000 Loss 0.4862\n","Epoch 36 Batch 4100 Loss 0.3990\n","Epoch 36 Loss 0.428833\n","Time taken for 1 epoch 579.59 sec\n","\n","Epoch 37 Batch 0 Loss 0.4497\n","Epoch 37 Batch 100 Loss 0.3999\n","Epoch 37 Batch 200 Loss 0.4693\n","Epoch 37 Batch 300 Loss 0.4201\n","Epoch 37 Batch 400 Loss 0.4268\n","Epoch 37 Batch 500 Loss 0.4219\n","Epoch 37 Batch 600 Loss 0.3951\n","Epoch 37 Batch 700 Loss 0.3506\n","Epoch 37 Batch 800 Loss 0.3933\n","Epoch 37 Batch 900 Loss 0.3958\n","Epoch 37 Batch 1000 Loss 0.3798\n","Epoch 37 Batch 1100 Loss 0.4694\n","Epoch 37 Batch 1200 Loss 0.4376\n","Epoch 37 Batch 1300 Loss 0.4435\n","Epoch 37 Batch 1400 Loss 0.4331\n","Epoch 37 Batch 1500 Loss 0.4178\n","Epoch 37 Batch 1600 Loss 0.4161\n","Epoch 37 Batch 1700 Loss 0.4171\n","Epoch 37 Batch 1800 Loss 0.3895\n","Epoch 37 Batch 1900 Loss 0.4254\n","Epoch 37 Batch 2000 Loss 0.4275\n","Epoch 37 Batch 2100 Loss 0.4671\n","Epoch 37 Batch 2200 Loss 0.4814\n","Epoch 37 Batch 2300 Loss 0.3630\n","Epoch 37 Batch 2400 Loss 0.4041\n","Epoch 37 Batch 2500 Loss 0.4161\n","Epoch 37 Batch 2600 Loss 0.4236\n","Epoch 37 Batch 2700 Loss 0.4110\n","Epoch 37 Batch 2800 Loss 0.4113\n","Epoch 37 Batch 2900 Loss 0.4000\n","Epoch 37 Batch 3000 Loss 0.4338\n","Epoch 37 Batch 3100 Loss 0.4309\n","Epoch 37 Batch 3200 Loss 0.4311\n","Epoch 37 Batch 3300 Loss 0.4459\n","Epoch 37 Batch 3400 Loss 0.3704\n","Epoch 37 Batch 3500 Loss 0.4280\n","Epoch 37 Batch 3600 Loss 0.4407\n","Epoch 37 Batch 3700 Loss 0.4313\n","Epoch 37 Batch 3800 Loss 0.4273\n","Epoch 37 Batch 3900 Loss 0.4073\n","Epoch 37 Batch 4000 Loss 0.4669\n","Epoch 37 Batch 4100 Loss 0.4074\n","Epoch 37 Loss 0.426450\n","Time taken for 1 epoch 578.88 sec\n","\n","Epoch 38 Batch 0 Loss 0.4425\n","Epoch 38 Batch 100 Loss 0.4085\n","Epoch 38 Batch 200 Loss 0.4419\n","Epoch 38 Batch 300 Loss 0.4167\n","Epoch 38 Batch 400 Loss 0.4270\n","Epoch 38 Batch 500 Loss 0.4378\n","Epoch 38 Batch 600 Loss 0.4658\n","Epoch 38 Batch 700 Loss 0.4529\n","Epoch 38 Batch 800 Loss 0.4354\n","Epoch 38 Batch 900 Loss 0.4153\n","Epoch 38 Batch 1000 Loss 0.4157\n","Epoch 38 Batch 1100 Loss 0.4298\n","Epoch 38 Batch 1200 Loss 0.4809\n","Epoch 38 Batch 1300 Loss 0.4639\n","Epoch 38 Batch 1400 Loss 0.4316\n","Epoch 38 Batch 1500 Loss 0.4195\n","Epoch 38 Batch 1600 Loss 0.4195\n","Epoch 38 Batch 1700 Loss 0.4285\n","Epoch 38 Batch 1800 Loss 0.4114\n","Epoch 38 Batch 1900 Loss 0.4097\n","Epoch 38 Batch 2000 Loss 0.4524\n","Epoch 38 Batch 2100 Loss 0.4449\n","Epoch 38 Batch 2200 Loss 0.3557\n","Epoch 38 Batch 2300 Loss 0.4486\n","Epoch 38 Batch 2400 Loss 0.4241\n","Epoch 38 Batch 2500 Loss 0.4222\n","Epoch 38 Batch 2600 Loss 0.4231\n","Epoch 38 Batch 2700 Loss 0.4068\n","Epoch 38 Batch 2800 Loss 0.4464\n","Epoch 38 Batch 2900 Loss 0.4398\n","Epoch 38 Batch 3000 Loss 0.4702\n","Epoch 38 Batch 3100 Loss 0.4158\n","Epoch 38 Batch 3200 Loss 0.3519\n","Epoch 38 Batch 3300 Loss 0.4254\n","Epoch 38 Batch 3400 Loss 0.4214\n","Epoch 38 Batch 3500 Loss 0.4205\n","Epoch 38 Batch 3600 Loss 0.4010\n","Epoch 38 Batch 3700 Loss 0.4096\n","Epoch 38 Batch 3800 Loss 0.3918\n","Epoch 38 Batch 3900 Loss 0.4302\n","Epoch 38 Batch 4000 Loss 0.4380\n","Epoch 38 Batch 4100 Loss 0.4762\n","Epoch 38 Loss 0.424995\n","Time taken for 1 epoch 579.90 sec\n","\n","Epoch 39 Batch 0 Loss 0.4694\n","Epoch 39 Batch 100 Loss 0.4557\n","Epoch 39 Batch 200 Loss 0.4280\n","Epoch 39 Batch 300 Loss 0.4305\n","Epoch 39 Batch 400 Loss 0.4109\n","Epoch 39 Batch 500 Loss 0.4304\n","Epoch 39 Batch 600 Loss 0.4556\n","Epoch 39 Batch 700 Loss 0.3893\n","Epoch 39 Batch 800 Loss 0.4429\n","Epoch 39 Batch 900 Loss 0.3990\n","Epoch 39 Batch 1000 Loss 0.4425\n","Epoch 39 Batch 1100 Loss 0.4269\n","Epoch 39 Batch 1200 Loss 0.4065\n","Epoch 39 Batch 1300 Loss 0.4157\n","Epoch 39 Batch 1400 Loss 0.4107\n","Epoch 39 Batch 1500 Loss 0.4195\n","Epoch 39 Batch 1600 Loss 0.4666\n","Epoch 39 Batch 1700 Loss 0.4236\n","Epoch 39 Batch 1800 Loss 0.4275\n","Epoch 39 Batch 1900 Loss 0.4407\n","Epoch 39 Batch 2000 Loss 0.4363\n","Epoch 39 Batch 2100 Loss 0.4164\n","Epoch 39 Batch 2200 Loss 0.4070\n","Epoch 39 Batch 2300 Loss 0.4120\n","Epoch 39 Batch 2400 Loss 0.4045\n","Epoch 39 Batch 2500 Loss 0.3931\n","Epoch 39 Batch 2600 Loss 0.4374\n","Epoch 39 Batch 2700 Loss 0.4153\n","Epoch 39 Batch 2800 Loss 0.4582\n","Epoch 39 Batch 2900 Loss 0.4077\n","Epoch 39 Batch 3000 Loss 0.4653\n","Epoch 39 Batch 3100 Loss 0.3952\n","Epoch 39 Batch 3200 Loss 0.4755\n","Epoch 39 Batch 3300 Loss 0.4591\n","Epoch 39 Batch 3400 Loss 0.4679\n","Epoch 39 Batch 3500 Loss 0.4380\n","Epoch 39 Batch 3600 Loss 0.3849\n","Epoch 39 Batch 3700 Loss 0.4254\n","Epoch 39 Batch 3800 Loss 0.4262\n","Epoch 39 Batch 3900 Loss 0.4240\n","Epoch 39 Batch 4000 Loss 0.4692\n","Epoch 39 Batch 4100 Loss 0.4499\n","Epoch 39 Loss 0.422956\n","Time taken for 1 epoch 578.95 sec\n","\n","Epoch 40 Batch 0 Loss 0.4429\n","Epoch 40 Batch 100 Loss 0.4384\n","Epoch 40 Batch 200 Loss 0.4051\n","Epoch 40 Batch 300 Loss 0.4552\n","Epoch 40 Batch 400 Loss 0.3949\n","Epoch 40 Batch 500 Loss 0.4284\n","Epoch 40 Batch 600 Loss 0.3978\n","Epoch 40 Batch 700 Loss 0.4346\n","Epoch 40 Batch 800 Loss 0.4464\n","Epoch 40 Batch 900 Loss 0.4218\n","Epoch 40 Batch 1000 Loss 0.4388\n","Epoch 40 Batch 1100 Loss 0.4106\n","Epoch 40 Batch 1200 Loss 0.4650\n","Epoch 40 Batch 1300 Loss 0.4065\n","Epoch 40 Batch 1400 Loss 0.4488\n","Epoch 40 Batch 1500 Loss 0.4563\n","Epoch 40 Batch 1600 Loss 0.4279\n","Epoch 40 Batch 1700 Loss 0.4257\n","Epoch 40 Batch 1800 Loss 0.3927\n","Epoch 40 Batch 1900 Loss 0.3876\n","Epoch 40 Batch 2000 Loss 0.4363\n","Epoch 40 Batch 2100 Loss 0.3935\n","Epoch 40 Batch 2200 Loss 0.4229\n","Epoch 40 Batch 2300 Loss 0.4195\n","Epoch 40 Batch 2400 Loss 0.4680\n","Epoch 40 Batch 2500 Loss 0.4218\n","Epoch 40 Batch 2600 Loss 0.4161\n","Epoch 40 Batch 2700 Loss 0.4210\n","Epoch 40 Batch 2800 Loss 0.4120\n","Epoch 40 Batch 2900 Loss 0.4281\n","Epoch 40 Batch 3000 Loss 0.4275\n","Epoch 40 Batch 3100 Loss 0.3983\n","Epoch 40 Batch 3200 Loss 0.4264\n","Epoch 40 Batch 3300 Loss 0.4451\n","Epoch 40 Batch 3400 Loss 0.4144\n","Epoch 40 Batch 3500 Loss 0.4681\n","Epoch 40 Batch 3600 Loss 0.4616\n","Epoch 40 Batch 3700 Loss 0.3975\n","Epoch 40 Batch 3800 Loss 0.4476\n","Epoch 40 Batch 3900 Loss 0.4025\n","Epoch 40 Batch 4000 Loss 0.4666\n","Epoch 40 Batch 4100 Loss 0.4145\n","Epoch 40 Loss 0.423118\n","Time taken for 1 epoch 578.92 sec\n","\n","Epoch 41 Batch 0 Loss 0.4220\n","Epoch 41 Batch 100 Loss 0.4219\n","Epoch 41 Batch 200 Loss 0.4031\n","Epoch 41 Batch 300 Loss 0.4241\n","Epoch 41 Batch 400 Loss 0.4461\n","Epoch 41 Batch 500 Loss 0.3952\n","Epoch 41 Batch 600 Loss 0.4122\n","Epoch 41 Batch 700 Loss 0.4392\n","Epoch 41 Batch 800 Loss 0.3908\n","Epoch 41 Batch 900 Loss 0.4343\n","Epoch 41 Batch 1000 Loss 0.3981\n","Epoch 41 Batch 1100 Loss 0.3943\n","Epoch 41 Batch 1200 Loss 0.4095\n","Epoch 41 Batch 1300 Loss 0.4212\n","Epoch 41 Batch 1400 Loss 0.4123\n","Epoch 41 Batch 1500 Loss 0.3549\n","Epoch 41 Batch 1600 Loss 0.4326\n","Epoch 41 Batch 1700 Loss 0.4118\n","Epoch 41 Batch 1800 Loss 0.4078\n","Epoch 41 Batch 1900 Loss 0.4360\n","Epoch 41 Batch 2000 Loss 0.4707\n","Epoch 41 Batch 2100 Loss 0.4586\n","Epoch 41 Batch 2200 Loss 0.4083\n","Epoch 41 Batch 2300 Loss 0.4375\n","Epoch 41 Batch 2400 Loss 0.4097\n","Epoch 41 Batch 2500 Loss 0.3831\n","Epoch 41 Batch 2600 Loss 0.4314\n","Epoch 41 Batch 2700 Loss 0.3971\n","Epoch 41 Batch 2800 Loss 0.4365\n","Epoch 41 Batch 2900 Loss 0.4024\n","Epoch 41 Batch 3000 Loss 0.4328\n","Epoch 41 Batch 3100 Loss 0.4115\n","Epoch 41 Batch 3200 Loss 0.4069\n","Epoch 41 Batch 3300 Loss 0.4233\n","Epoch 41 Batch 3400 Loss 0.4204\n","Epoch 41 Batch 3500 Loss 0.3957\n","Epoch 41 Batch 3600 Loss 0.4016\n","Epoch 41 Batch 3700 Loss 0.3733\n","Epoch 41 Batch 3800 Loss 0.3681\n","Epoch 41 Batch 3900 Loss 0.4488\n","Epoch 41 Batch 4000 Loss 0.4072\n","Epoch 41 Batch 4100 Loss 0.4138\n","Epoch 41 Loss 0.420756\n","Time taken for 1 epoch 579.99 sec\n","\n","Epoch 42 Batch 0 Loss 0.4359\n","Epoch 42 Batch 100 Loss 0.4048\n","Epoch 42 Batch 200 Loss 0.4703\n","Epoch 42 Batch 300 Loss 0.4397\n","Epoch 42 Batch 400 Loss 0.4030\n","Epoch 42 Batch 500 Loss 0.4548\n","Epoch 42 Batch 600 Loss 0.4359\n","Epoch 42 Batch 700 Loss 0.4049\n","Epoch 42 Batch 800 Loss 0.3816\n","Epoch 42 Batch 900 Loss 0.4176\n","Epoch 42 Batch 1000 Loss 0.3939\n","Epoch 42 Batch 1100 Loss 0.4228\n","Epoch 42 Batch 1200 Loss 0.4362\n","Epoch 42 Batch 1300 Loss 0.4262\n","Epoch 42 Batch 1400 Loss 0.3826\n","Epoch 42 Batch 1500 Loss 0.4148\n","Epoch 42 Batch 1600 Loss 0.4000\n","Epoch 42 Batch 1700 Loss 0.3978\n","Epoch 42 Batch 1800 Loss 0.4463\n","Epoch 42 Batch 1900 Loss 0.3985\n","Epoch 42 Batch 2000 Loss 0.4060\n","Epoch 42 Batch 2100 Loss 0.4209\n","Epoch 42 Batch 2200 Loss 0.3994\n","Epoch 42 Batch 2300 Loss 0.4177\n","Epoch 42 Batch 2400 Loss 0.4142\n","Epoch 42 Batch 2500 Loss 0.4125\n","Epoch 42 Batch 2600 Loss 0.4237\n","Epoch 42 Batch 2700 Loss 0.3930\n","Epoch 42 Batch 2800 Loss 0.4674\n","Epoch 42 Batch 2900 Loss 0.4282\n","Epoch 42 Batch 3000 Loss 0.4033\n","Epoch 42 Batch 3100 Loss 0.3953\n","Epoch 42 Batch 3200 Loss 0.3785\n","Epoch 42 Batch 3300 Loss 0.4388\n","Epoch 42 Batch 3400 Loss 0.4153\n","Epoch 42 Batch 3500 Loss 0.4390\n","Epoch 42 Batch 3600 Loss 0.4376\n","Epoch 42 Batch 3700 Loss 0.4406\n","Epoch 42 Batch 3800 Loss 0.3995\n","Epoch 42 Batch 3900 Loss 0.4579\n","Epoch 42 Batch 4000 Loss 0.4364\n","Epoch 42 Batch 4100 Loss 0.4727\n","Epoch 42 Loss 0.418697\n","Time taken for 1 epoch 578.86 sec\n","\n","Epoch 43 Batch 0 Loss 0.4062\n","Epoch 43 Batch 100 Loss 0.4226\n","Epoch 43 Batch 200 Loss 0.3755\n","Epoch 43 Batch 300 Loss 0.4493\n","Epoch 43 Batch 400 Loss 0.4175\n","Epoch 43 Batch 500 Loss 0.4130\n","Epoch 43 Batch 600 Loss 0.3649\n","Epoch 43 Batch 700 Loss 0.4129\n","Epoch 43 Batch 800 Loss 0.4139\n","Epoch 43 Batch 900 Loss 0.3931\n","Epoch 43 Batch 1000 Loss 0.3565\n","Epoch 43 Batch 1100 Loss 0.3891\n","Epoch 43 Batch 1200 Loss 0.4140\n","Epoch 43 Batch 1300 Loss 0.4287\n","Epoch 43 Batch 1400 Loss 0.3870\n","Epoch 43 Batch 1500 Loss 0.4138\n","Epoch 43 Batch 1600 Loss 0.4180\n","Epoch 43 Batch 1700 Loss 0.3936\n","Epoch 43 Batch 1800 Loss 0.4117\n","Epoch 43 Batch 1900 Loss 0.4204\n","Epoch 43 Batch 2000 Loss 0.4290\n","Epoch 43 Batch 2100 Loss 0.4839\n","Epoch 43 Batch 2200 Loss 0.3937\n","Epoch 43 Batch 2300 Loss 0.4094\n","Epoch 43 Batch 2400 Loss 0.4813\n","Epoch 43 Batch 2500 Loss 0.4084\n","Epoch 43 Batch 2600 Loss 0.4487\n","Epoch 43 Batch 2700 Loss 0.4256\n","Epoch 43 Batch 2800 Loss 0.4318\n","Epoch 43 Batch 2900 Loss 0.3891\n","Epoch 43 Batch 3000 Loss 0.4443\n","Epoch 43 Batch 3100 Loss 0.4373\n","Epoch 43 Batch 3200 Loss 0.3897\n","Epoch 43 Batch 3300 Loss 0.3974\n","Epoch 43 Batch 3400 Loss 0.4398\n","Epoch 43 Batch 3500 Loss 0.4134\n","Epoch 43 Batch 3600 Loss 0.4550\n","Epoch 43 Batch 3700 Loss 0.4325\n","Epoch 43 Batch 3800 Loss 0.4211\n","Epoch 43 Batch 3900 Loss 0.4018\n","Epoch 43 Batch 4000 Loss 0.4022\n","Epoch 43 Batch 4100 Loss 0.4313\n","Epoch 43 Loss 0.417640\n","Time taken for 1 epoch 580.47 sec\n","\n","Epoch 44 Batch 0 Loss 0.4155\n","Epoch 44 Batch 100 Loss 0.3882\n","Epoch 44 Batch 200 Loss 0.4365\n","Epoch 44 Batch 300 Loss 0.4069\n","Epoch 44 Batch 400 Loss 0.4079\n","Epoch 44 Batch 500 Loss 0.3898\n","Epoch 44 Batch 600 Loss 0.4197\n","Epoch 44 Batch 700 Loss 0.4288\n","Epoch 44 Batch 800 Loss 0.4096\n","Epoch 44 Batch 900 Loss 0.4814\n","Epoch 44 Batch 1000 Loss 0.4276\n","Epoch 44 Batch 1100 Loss 0.3733\n","Epoch 44 Batch 1200 Loss 0.4074\n","Epoch 44 Batch 1300 Loss 0.4395\n","Epoch 44 Batch 1400 Loss 0.3724\n","Epoch 44 Batch 1500 Loss 0.4633\n","Epoch 44 Batch 1600 Loss 0.3970\n","Epoch 44 Batch 1700 Loss 0.4311\n","Epoch 44 Batch 1800 Loss 0.3989\n","Epoch 44 Batch 1900 Loss 0.4212\n","Epoch 44 Batch 2000 Loss 0.4098\n","Epoch 44 Batch 2100 Loss 0.4148\n","Epoch 44 Batch 2200 Loss 0.4444\n","Epoch 44 Batch 2300 Loss 0.4028\n","Epoch 44 Batch 2400 Loss 0.4304\n","Epoch 44 Batch 2500 Loss 0.3538\n","Epoch 44 Batch 2600 Loss 0.3969\n","Epoch 44 Batch 2700 Loss 0.4802\n","Epoch 44 Batch 2800 Loss 0.3708\n","Epoch 44 Batch 2900 Loss 0.4543\n","Epoch 44 Batch 3000 Loss 0.3934\n","Epoch 44 Batch 3100 Loss 0.3566\n","Epoch 44 Batch 3200 Loss 0.4127\n","Epoch 44 Batch 3300 Loss 0.3988\n","Epoch 44 Batch 3400 Loss 0.4082\n","Epoch 44 Batch 3500 Loss 0.3883\n","Epoch 44 Batch 3600 Loss 0.3900\n","Epoch 44 Batch 3700 Loss 0.4094\n","Epoch 44 Batch 3800 Loss 0.3821\n","Epoch 44 Batch 3900 Loss 0.4053\n","Epoch 44 Batch 4000 Loss 0.4036\n","Epoch 44 Batch 4100 Loss 0.4298\n","Epoch 44 Loss 0.415582\n","Time taken for 1 epoch 579.51 sec\n","\n","Epoch 45 Batch 0 Loss 0.4333\n","Epoch 45 Batch 100 Loss 0.4156\n","Epoch 45 Batch 200 Loss 0.5023\n","Epoch 45 Batch 300 Loss 0.4446\n","Epoch 45 Batch 400 Loss 0.3858\n","Epoch 45 Batch 500 Loss 0.3664\n","Epoch 45 Batch 600 Loss 0.4083\n","Epoch 45 Batch 700 Loss 0.4444\n","Epoch 45 Batch 800 Loss 0.4073\n","Epoch 45 Batch 900 Loss 0.4210\n","Epoch 45 Batch 1000 Loss 0.4985\n","Epoch 45 Batch 1100 Loss 0.3935\n","Epoch 45 Batch 1200 Loss 0.3731\n","Epoch 45 Batch 1300 Loss 0.4242\n","Epoch 45 Batch 1400 Loss 0.4112\n","Epoch 45 Batch 1500 Loss 0.4037\n","Epoch 45 Batch 1600 Loss 0.4508\n","Epoch 45 Batch 1700 Loss 0.4295\n","Epoch 45 Batch 1800 Loss 0.4632\n","Epoch 45 Batch 1900 Loss 0.4137\n","Epoch 45 Batch 2000 Loss 0.3839\n","Epoch 45 Batch 2100 Loss 0.3891\n","Epoch 45 Batch 2200 Loss 0.3968\n","Epoch 45 Batch 2300 Loss 0.4580\n","Epoch 45 Batch 2400 Loss 0.4428\n","Epoch 45 Batch 2500 Loss 0.3956\n","Epoch 45 Batch 2600 Loss 0.4119\n","Epoch 45 Batch 2700 Loss 0.4157\n","Epoch 45 Batch 2800 Loss 0.4019\n","Epoch 45 Batch 2900 Loss 0.4309\n","Epoch 45 Batch 3000 Loss 0.4716\n","Epoch 45 Batch 3100 Loss 0.4362\n","Epoch 45 Batch 3200 Loss 0.3659\n","Epoch 45 Batch 3300 Loss 0.4226\n","Epoch 45 Batch 3400 Loss 0.4180\n","Epoch 45 Batch 3500 Loss 0.4133\n","Epoch 45 Batch 3600 Loss 0.4444\n","Epoch 45 Batch 3700 Loss 0.4064\n","Epoch 45 Batch 3800 Loss 0.3644\n","Epoch 45 Batch 3900 Loss 0.4520\n","Epoch 45 Batch 4000 Loss 0.4235\n","Epoch 45 Batch 4100 Loss 0.4014\n","Epoch 45 Loss 0.414936\n","Time taken for 1 epoch 579.21 sec\n","\n","Epoch 46 Batch 0 Loss 0.4034\n","Epoch 46 Batch 100 Loss 0.4299\n","Epoch 46 Batch 200 Loss 0.4300\n","Epoch 46 Batch 300 Loss 0.4274\n","Epoch 46 Batch 400 Loss 0.4313\n","Epoch 46 Batch 500 Loss 0.4080\n","Epoch 46 Batch 600 Loss 0.3925\n","Epoch 46 Batch 700 Loss 0.3650\n","Epoch 46 Batch 800 Loss 0.3945\n","Epoch 46 Batch 900 Loss 0.4023\n","Epoch 46 Batch 1000 Loss 0.4180\n","Epoch 46 Batch 1100 Loss 0.4224\n","Epoch 46 Batch 1200 Loss 0.4335\n","Epoch 46 Batch 1300 Loss 0.4528\n","Epoch 46 Batch 1400 Loss 0.3542\n","Epoch 46 Batch 1500 Loss 0.4339\n","Epoch 46 Batch 1600 Loss 0.4129\n","Epoch 46 Batch 1700 Loss 0.4524\n","Epoch 46 Batch 1800 Loss 0.3797\n","Epoch 46 Batch 1900 Loss 0.4276\n","Epoch 46 Batch 2000 Loss 0.3886\n","Epoch 46 Batch 2100 Loss 0.4212\n","Epoch 46 Batch 2200 Loss 0.4057\n","Epoch 46 Batch 2300 Loss 0.4370\n","Epoch 46 Batch 2400 Loss 0.4143\n","Epoch 46 Batch 2500 Loss 0.4229\n","Epoch 46 Batch 2600 Loss 0.4048\n","Epoch 46 Batch 2700 Loss 0.4439\n","Epoch 46 Batch 2800 Loss 0.3826\n","Epoch 46 Batch 2900 Loss 0.4045\n","Epoch 46 Batch 3000 Loss 0.4240\n","Epoch 46 Batch 3100 Loss 0.4191\n","Epoch 46 Batch 3200 Loss 0.3650\n","Epoch 46 Batch 3300 Loss 0.3967\n","Epoch 46 Batch 3400 Loss 0.4098\n","Epoch 46 Batch 3500 Loss 0.4723\n","Epoch 46 Batch 3600 Loss 0.3894\n","Epoch 46 Batch 3700 Loss 0.3943\n","Epoch 46 Batch 3800 Loss 0.4189\n","Epoch 46 Batch 3900 Loss 0.3554\n","Epoch 46 Batch 4000 Loss 0.4146\n","Epoch 46 Batch 4100 Loss 0.4037\n","Epoch 46 Loss 0.413263\n","Time taken for 1 epoch 579.57 sec\n","\n","Epoch 47 Batch 0 Loss 0.3837\n","Epoch 47 Batch 100 Loss 0.4176\n","Epoch 47 Batch 200 Loss 0.4411\n","Epoch 47 Batch 300 Loss 0.4281\n","Epoch 47 Batch 400 Loss 0.3888\n","Epoch 47 Batch 500 Loss 0.4174\n","Epoch 47 Batch 600 Loss 0.4147\n","Epoch 47 Batch 700 Loss 0.4529\n","Epoch 47 Batch 800 Loss 0.4288\n","Epoch 47 Batch 900 Loss 0.4166\n","Epoch 47 Batch 1000 Loss 0.3972\n","Epoch 47 Batch 1100 Loss 0.4337\n","Epoch 47 Batch 1200 Loss 0.3997\n","Epoch 47 Batch 1300 Loss 0.4483\n","Epoch 47 Batch 1400 Loss 0.3567\n","Epoch 47 Batch 1500 Loss 0.3983\n","Epoch 47 Batch 1600 Loss 0.4144\n","Epoch 47 Batch 1700 Loss 0.4110\n","Epoch 47 Batch 1800 Loss 0.3416\n","Epoch 47 Batch 1900 Loss 0.3933\n","Epoch 47 Batch 2000 Loss 0.3704\n","Epoch 47 Batch 2100 Loss 0.4806\n","Epoch 47 Batch 2200 Loss 0.4084\n","Epoch 47 Batch 2300 Loss 0.4167\n","Epoch 47 Batch 2400 Loss 0.4687\n","Epoch 47 Batch 2500 Loss 0.3684\n","Epoch 47 Batch 2600 Loss 0.4214\n","Epoch 47 Batch 2700 Loss 0.4198\n","Epoch 47 Batch 2800 Loss 0.3823\n","Epoch 47 Batch 2900 Loss 0.3834\n","Epoch 47 Batch 3000 Loss 0.4213\n","Epoch 47 Batch 3100 Loss 0.3952\n","Epoch 47 Batch 3200 Loss 0.4094\n","Epoch 47 Batch 3300 Loss 0.3857\n","Epoch 47 Batch 3400 Loss 0.3825\n","Epoch 47 Batch 3500 Loss 0.4024\n","Epoch 47 Batch 3600 Loss 0.3863\n","Epoch 47 Batch 3700 Loss 0.4027\n","Epoch 47 Batch 3800 Loss 0.4339\n","Epoch 47 Batch 3900 Loss 0.4218\n","Epoch 47 Batch 4000 Loss 0.4071\n","Epoch 47 Batch 4100 Loss 0.3977\n","Epoch 47 Loss 0.412402\n","Time taken for 1 epoch 579.76 sec\n","\n","Epoch 48 Batch 0 Loss 0.3854\n","Epoch 48 Batch 100 Loss 0.4556\n","Epoch 48 Batch 200 Loss 0.4474\n","Epoch 48 Batch 300 Loss 0.4252\n","Epoch 48 Batch 400 Loss 0.4875\n","Epoch 48 Batch 500 Loss 0.3873\n","Epoch 48 Batch 600 Loss 0.4328\n","Epoch 48 Batch 700 Loss 0.3988\n","Epoch 48 Batch 800 Loss 0.4183\n","Epoch 48 Batch 900 Loss 0.4065\n","Epoch 48 Batch 1000 Loss 0.4217\n","Epoch 48 Batch 1100 Loss 0.4113\n","Epoch 48 Batch 1200 Loss 0.3996\n","Epoch 48 Batch 1300 Loss 0.3770\n","Epoch 48 Batch 1400 Loss 0.3743\n","Epoch 48 Batch 1500 Loss 0.4178\n","Epoch 48 Batch 1600 Loss 0.4132\n","Epoch 48 Batch 1700 Loss 0.3596\n","Epoch 48 Batch 1800 Loss 0.4184\n","Epoch 48 Batch 1900 Loss 0.4135\n","Epoch 48 Batch 2000 Loss 0.4096\n","Epoch 48 Batch 2100 Loss 0.4464\n","Epoch 48 Batch 2200 Loss 0.4155\n","Epoch 48 Batch 2300 Loss 0.3771\n","Epoch 48 Batch 2400 Loss 0.4198\n","Epoch 48 Batch 2500 Loss 0.4287\n","Epoch 48 Batch 2600 Loss 0.4194\n","Epoch 48 Batch 2700 Loss 0.3910\n","Epoch 48 Batch 2800 Loss 0.4386\n","Epoch 48 Batch 2900 Loss 0.3933\n","Epoch 48 Batch 3000 Loss 0.4272\n","Epoch 48 Batch 3100 Loss 0.4139\n","Epoch 48 Batch 3200 Loss 0.4279\n","Epoch 48 Batch 3300 Loss 0.4140\n","Epoch 48 Batch 3400 Loss 0.4082\n","Epoch 48 Batch 3500 Loss 0.3974\n","Epoch 48 Batch 3600 Loss 0.3709\n","Epoch 48 Batch 3700 Loss 0.3690\n","Epoch 48 Batch 3800 Loss 0.4421\n","Epoch 48 Batch 3900 Loss 0.3851\n","Epoch 48 Batch 4000 Loss 0.4744\n","Epoch 48 Batch 4100 Loss 0.4010\n","Epoch 48 Loss 0.411459\n","Time taken for 1 epoch 579.22 sec\n","\n","Epoch 49 Batch 0 Loss 0.3890\n","Epoch 49 Batch 100 Loss 0.3742\n","Epoch 49 Batch 200 Loss 0.3762\n","Epoch 49 Batch 300 Loss 0.4192\n","Epoch 49 Batch 400 Loss 0.4007\n","Epoch 49 Batch 500 Loss 0.3763\n","Epoch 49 Batch 600 Loss 0.3911\n","Epoch 49 Batch 700 Loss 0.4020\n","Epoch 49 Batch 800 Loss 0.4098\n","Epoch 49 Batch 900 Loss 0.4121\n","Epoch 49 Batch 1000 Loss 0.4106\n","Epoch 49 Batch 1100 Loss 0.3893\n","Epoch 49 Batch 1200 Loss 0.3966\n","Epoch 49 Batch 1300 Loss 0.4173\n","Epoch 49 Batch 1400 Loss 0.4148\n","Epoch 49 Batch 1500 Loss 0.4605\n","Epoch 49 Batch 1600 Loss 0.4209\n","Epoch 49 Batch 1700 Loss 0.3901\n","Epoch 49 Batch 1800 Loss 0.3750\n","Epoch 49 Batch 1900 Loss 0.4050\n","Epoch 49 Batch 2000 Loss 0.4745\n","Epoch 49 Batch 2100 Loss 0.4243\n","Epoch 49 Batch 2200 Loss 0.4069\n","Epoch 49 Batch 2300 Loss 0.3983\n","Epoch 49 Batch 2400 Loss 0.4148\n","Epoch 49 Batch 2500 Loss 0.4604\n","Epoch 49 Batch 2600 Loss 0.4081\n","Epoch 49 Batch 2700 Loss 0.3756\n","Epoch 49 Batch 2800 Loss 0.4105\n","Epoch 49 Batch 2900 Loss 0.4557\n","Epoch 49 Batch 3000 Loss 0.4494\n","Epoch 49 Batch 3100 Loss 0.3889\n","Epoch 49 Batch 3200 Loss 0.4147\n","Epoch 49 Batch 3300 Loss 0.3833\n","Epoch 49 Batch 3400 Loss 0.3758\n","Epoch 49 Batch 3500 Loss 0.3871\n","Epoch 49 Batch 3600 Loss 0.4003\n","Epoch 49 Batch 3700 Loss 0.4171\n","Epoch 49 Batch 3800 Loss 0.4907\n","Epoch 49 Batch 3900 Loss 0.3745\n","Epoch 49 Batch 4000 Loss 0.4244\n","Epoch 49 Batch 4100 Loss 0.4306\n","Epoch 49 Loss 0.409798\n","Time taken for 1 epoch 579.53 sec\n","\n","Epoch 50 Batch 0 Loss 0.3944\n","Epoch 50 Batch 100 Loss 0.4128\n","Epoch 50 Batch 200 Loss 0.4037\n","Epoch 50 Batch 300 Loss 0.3948\n","Epoch 50 Batch 400 Loss 0.3686\n","Epoch 50 Batch 500 Loss 0.4417\n","Epoch 50 Batch 600 Loss 0.3703\n","Epoch 50 Batch 700 Loss 0.4329\n","Epoch 50 Batch 800 Loss 0.4366\n","Epoch 50 Batch 900 Loss 0.4502\n","Epoch 50 Batch 1000 Loss 0.3982\n","Epoch 50 Batch 1100 Loss 0.4183\n","Epoch 50 Batch 1200 Loss 0.3935\n","Epoch 50 Batch 1300 Loss 0.4248\n","Epoch 50 Batch 1400 Loss 0.4077\n","Epoch 50 Batch 1500 Loss 0.3866\n","Epoch 50 Batch 1600 Loss 0.4312\n","Epoch 50 Batch 1700 Loss 0.4032\n","Epoch 50 Batch 1800 Loss 0.3725\n","Epoch 50 Batch 1900 Loss 0.4001\n","Epoch 50 Batch 2000 Loss 0.4415\n","Epoch 50 Batch 2100 Loss 0.3801\n","Epoch 50 Batch 2200 Loss 0.3585\n","Epoch 50 Batch 2300 Loss 0.3940\n","Epoch 50 Batch 2400 Loss 0.4575\n","Epoch 50 Batch 2500 Loss 0.4188\n","Epoch 50 Batch 2600 Loss 0.4081\n","Epoch 50 Batch 2700 Loss 0.4007\n","Epoch 50 Batch 2800 Loss 0.4554\n","Epoch 50 Batch 2900 Loss 0.4194\n","Epoch 50 Batch 3000 Loss 0.3912\n","Epoch 50 Batch 3100 Loss 0.4108\n","Epoch 50 Batch 3200 Loss 0.4094\n","Epoch 50 Batch 3300 Loss 0.4071\n","Epoch 50 Batch 3400 Loss 0.3635\n","Epoch 50 Batch 3500 Loss 0.4425\n","Epoch 50 Batch 3600 Loss 0.3963\n","Epoch 50 Batch 3700 Loss 0.4140\n","Epoch 50 Batch 3800 Loss 0.4618\n","Epoch 50 Batch 3900 Loss 0.3817\n","Epoch 50 Batch 4000 Loss 0.4029\n","Epoch 50 Batch 4100 Loss 0.3902\n","Epoch 50 Loss 0.409483\n","Time taken for 1 epoch 578.73 sec\n","\n","Epoch 51 Batch 0 Loss 0.3962\n","Epoch 51 Batch 100 Loss 0.3624\n","Epoch 51 Batch 200 Loss 0.4617\n","Epoch 51 Batch 300 Loss 0.3863\n","Epoch 51 Batch 400 Loss 0.4166\n","Epoch 51 Batch 500 Loss 0.3661\n","Epoch 51 Batch 600 Loss 0.4271\n","Epoch 51 Batch 700 Loss 0.4070\n","Epoch 51 Batch 800 Loss 0.4352\n","Epoch 51 Batch 900 Loss 0.4493\n","Epoch 51 Batch 1000 Loss 0.3891\n","Epoch 51 Batch 1100 Loss 0.3739\n","Epoch 51 Batch 1200 Loss 0.3754\n","Epoch 51 Batch 1300 Loss 0.4048\n","Epoch 51 Batch 1400 Loss 0.4065\n","Epoch 51 Batch 1500 Loss 0.4068\n","Epoch 51 Batch 1600 Loss 0.4608\n","Epoch 51 Batch 1700 Loss 0.3925\n","Epoch 51 Batch 1800 Loss 0.4038\n","Epoch 51 Batch 1900 Loss 0.4321\n","Epoch 51 Batch 2000 Loss 0.4422\n","Epoch 51 Batch 2100 Loss 0.4236\n","Epoch 51 Batch 2200 Loss 0.3685\n","Epoch 51 Batch 2300 Loss 0.4002\n","Epoch 51 Batch 2400 Loss 0.4248\n","Epoch 51 Batch 2500 Loss 0.3933\n","Epoch 51 Batch 2600 Loss 0.4064\n","Epoch 51 Batch 2700 Loss 0.3948\n","Epoch 51 Batch 2800 Loss 0.4216\n","Epoch 51 Batch 2900 Loss 0.4303\n","Epoch 51 Batch 3000 Loss 0.4061\n","Epoch 51 Batch 3100 Loss 0.4136\n","Epoch 51 Batch 3200 Loss 0.4087\n","Epoch 51 Batch 3300 Loss 0.3951\n","Epoch 51 Batch 3400 Loss 0.4608\n","Epoch 51 Batch 3500 Loss 0.4170\n","Epoch 51 Batch 3600 Loss 0.4420\n","Epoch 51 Batch 3700 Loss 0.4288\n","Epoch 51 Batch 3800 Loss 0.3969\n","Epoch 51 Batch 3900 Loss 0.3981\n","Epoch 51 Batch 4000 Loss 0.3970\n","Epoch 51 Batch 4100 Loss 0.4492\n","Epoch 51 Loss 0.409304\n","Time taken for 1 epoch 579.99 sec\n","\n","Epoch 52 Batch 0 Loss 0.4064\n","Epoch 52 Batch 100 Loss 0.3834\n","Epoch 52 Batch 200 Loss 0.4470\n","Epoch 52 Batch 300 Loss 0.4298\n","Epoch 52 Batch 400 Loss 0.4288\n","Epoch 52 Batch 500 Loss 0.4051\n","Epoch 52 Batch 600 Loss 0.4097\n","Epoch 52 Batch 700 Loss 0.4659\n","Epoch 52 Batch 800 Loss 0.4077\n","Epoch 52 Batch 900 Loss 0.4087\n","Epoch 52 Batch 1000 Loss 0.3919\n","Epoch 52 Batch 1100 Loss 0.4279\n","Epoch 52 Batch 1200 Loss 0.4202\n","Epoch 52 Batch 1300 Loss 0.4071\n","Epoch 52 Batch 1400 Loss 0.4072\n","Epoch 52 Batch 1500 Loss 0.4396\n","Epoch 52 Batch 1600 Loss 0.4567\n","Epoch 52 Batch 1700 Loss 0.3911\n","Epoch 52 Batch 1800 Loss 0.4104\n","Epoch 52 Batch 1900 Loss 0.3653\n","Epoch 52 Batch 2000 Loss 0.3949\n","Epoch 52 Batch 2100 Loss 0.4258\n","Epoch 52 Batch 2200 Loss 0.3789\n","Epoch 52 Batch 2300 Loss 0.4482\n","Epoch 52 Batch 2400 Loss 0.4024\n","Epoch 52 Batch 2500 Loss 0.3952\n","Epoch 52 Batch 2600 Loss 0.4170\n","Epoch 52 Batch 2700 Loss 0.3766\n","Epoch 52 Batch 2800 Loss 0.3637\n","Epoch 52 Batch 2900 Loss 0.3926\n","Epoch 52 Batch 3000 Loss 0.4290\n","Epoch 52 Batch 3100 Loss 0.3959\n","Epoch 52 Batch 3200 Loss 0.4136\n","Epoch 52 Batch 3300 Loss 0.4013\n","Epoch 52 Batch 3400 Loss 0.4150\n","Epoch 52 Batch 3500 Loss 0.3951\n","Epoch 52 Batch 3600 Loss 0.3967\n","Epoch 52 Batch 3700 Loss 0.4423\n","Epoch 52 Batch 3800 Loss 0.4254\n","Epoch 52 Batch 3900 Loss 0.3667\n","Epoch 52 Batch 4000 Loss 0.4405\n","Epoch 52 Batch 4100 Loss 0.3960\n","Epoch 52 Loss 0.407575\n","Time taken for 1 epoch 578.65 sec\n","\n","Epoch 53 Batch 0 Loss 0.3822\n","Epoch 53 Batch 100 Loss 0.4196\n","Epoch 53 Batch 200 Loss 0.4002\n","Epoch 53 Batch 300 Loss 0.3890\n","Epoch 53 Batch 400 Loss 0.4231\n","Epoch 53 Batch 500 Loss 0.4153\n","Epoch 53 Batch 600 Loss 0.3684\n","Epoch 53 Batch 700 Loss 0.3783\n","Epoch 53 Batch 800 Loss 0.4449\n","Epoch 53 Batch 900 Loss 0.4012\n","Epoch 53 Batch 1000 Loss 0.4475\n","Epoch 53 Batch 1100 Loss 0.4369\n","Epoch 53 Batch 1200 Loss 0.4185\n","Epoch 53 Batch 1300 Loss 0.4504\n","Epoch 53 Batch 1400 Loss 0.4026\n","Epoch 53 Batch 1500 Loss 0.3910\n","Epoch 53 Batch 1600 Loss 0.4355\n","Epoch 53 Batch 1700 Loss 0.3912\n","Epoch 53 Batch 1800 Loss 0.4057\n","Epoch 53 Batch 1900 Loss 0.4316\n","Epoch 53 Batch 2000 Loss 0.4013\n","Epoch 53 Batch 2100 Loss 0.4241\n","Epoch 53 Batch 2200 Loss 0.3973\n","Epoch 53 Batch 2300 Loss 0.4124\n","Epoch 53 Batch 2400 Loss 0.4058\n","Epoch 53 Batch 2500 Loss 0.3966\n","Epoch 53 Batch 2600 Loss 0.4097\n","Epoch 53 Batch 2700 Loss 0.3526\n","Epoch 53 Batch 2800 Loss 0.4214\n","Epoch 53 Batch 2900 Loss 0.4519\n","Epoch 53 Batch 3000 Loss 0.4164\n","Epoch 53 Batch 3100 Loss 0.4156\n","Epoch 53 Batch 3200 Loss 0.4196\n","Epoch 53 Batch 3300 Loss 0.4081\n","Epoch 53 Batch 3400 Loss 0.4102\n","Epoch 53 Batch 3500 Loss 0.3723\n","Epoch 53 Batch 3600 Loss 0.4583\n","Epoch 53 Batch 3700 Loss 0.3477\n","Epoch 53 Batch 3800 Loss 0.3771\n","Epoch 53 Batch 3900 Loss 0.3786\n","Epoch 53 Batch 4000 Loss 0.3863\n","Epoch 53 Batch 4100 Loss 0.4008\n","Epoch 53 Loss 0.406235\n","Time taken for 1 epoch 578.80 sec\n","\n","Epoch 54 Batch 0 Loss 0.4166\n","Epoch 54 Batch 100 Loss 0.4323\n","Epoch 54 Batch 200 Loss 0.4575\n","Epoch 54 Batch 300 Loss 0.4109\n","Epoch 54 Batch 400 Loss 0.4355\n","Epoch 54 Batch 500 Loss 0.4121\n","Epoch 54 Batch 600 Loss 0.4181\n","Epoch 54 Batch 700 Loss 0.3706\n","Epoch 54 Batch 800 Loss 0.4061\n","Epoch 54 Batch 900 Loss 0.4144\n","Epoch 54 Batch 1000 Loss 0.4068\n","Epoch 54 Batch 1100 Loss 0.4187\n","Epoch 54 Batch 1200 Loss 0.3876\n","Epoch 54 Batch 1300 Loss 0.4069\n","Epoch 54 Batch 1400 Loss 0.4232\n","Epoch 54 Batch 1500 Loss 0.3804\n","Epoch 54 Batch 1600 Loss 0.4539\n","Epoch 54 Batch 1700 Loss 0.3863\n","Epoch 54 Batch 1800 Loss 0.4329\n","Epoch 54 Batch 1900 Loss 0.4022\n","Epoch 54 Batch 2000 Loss 0.4196\n","Epoch 54 Batch 2100 Loss 0.3915\n","Epoch 54 Batch 2200 Loss 0.4000\n","Epoch 54 Batch 2300 Loss 0.4191\n","Epoch 54 Batch 2400 Loss 0.3942\n","Epoch 54 Batch 2500 Loss 0.3998\n","Epoch 54 Batch 2600 Loss 0.3552\n","Epoch 54 Batch 2700 Loss 0.4308\n","Epoch 54 Batch 2800 Loss 0.4170\n","Epoch 54 Batch 2900 Loss 0.3635\n","Epoch 54 Batch 3000 Loss 0.4053\n","Epoch 54 Batch 3100 Loss 0.3784\n","Epoch 54 Batch 3200 Loss 0.3801\n","Epoch 54 Batch 3300 Loss 0.4166\n","Epoch 54 Batch 3400 Loss 0.4133\n","Epoch 54 Batch 3500 Loss 0.4078\n","Epoch 54 Batch 3600 Loss 0.4151\n","Epoch 54 Batch 3700 Loss 0.4917\n","Epoch 54 Batch 3800 Loss 0.3793\n","Epoch 54 Batch 3900 Loss 0.3825\n","Epoch 54 Batch 4000 Loss 0.4076\n","Epoch 54 Batch 4100 Loss 0.4015\n","Epoch 54 Loss 0.404494\n","Time taken for 1 epoch 579.42 sec\n","\n","Epoch 55 Batch 0 Loss 0.3569\n","Epoch 55 Batch 100 Loss 0.3974\n","Epoch 55 Batch 200 Loss 0.4294\n","Epoch 55 Batch 300 Loss 0.4003\n","Epoch 55 Batch 400 Loss 0.4051\n","Epoch 55 Batch 500 Loss 0.3963\n","Epoch 55 Batch 600 Loss 0.3644\n","Epoch 55 Batch 700 Loss 0.3880\n","Epoch 55 Batch 800 Loss 0.3799\n","Epoch 55 Batch 900 Loss 0.4336\n","Epoch 55 Batch 1000 Loss 0.4657\n","Epoch 55 Batch 1100 Loss 0.3804\n","Epoch 55 Batch 1200 Loss 0.3914\n","Epoch 55 Batch 1300 Loss 0.4249\n","Epoch 55 Batch 1400 Loss 0.4225\n","Epoch 55 Batch 1500 Loss 0.4299\n","Epoch 55 Batch 1600 Loss 0.4353\n","Epoch 55 Batch 1700 Loss 0.3942\n","Epoch 55 Batch 1800 Loss 0.4272\n","Epoch 55 Batch 1900 Loss 0.3952\n","Epoch 55 Batch 2000 Loss 0.3702\n","Epoch 55 Batch 2100 Loss 0.4500\n","Epoch 55 Batch 2200 Loss 0.4002\n","Epoch 55 Batch 2300 Loss 0.3921\n","Epoch 55 Batch 2400 Loss 0.4041\n","Epoch 55 Batch 2500 Loss 0.3920\n","Epoch 55 Batch 2600 Loss 0.4074\n","Epoch 55 Batch 2700 Loss 0.3678\n","Epoch 55 Batch 2800 Loss 0.4189\n","Epoch 55 Batch 2900 Loss 0.3953\n","Epoch 55 Batch 3000 Loss 0.3737\n","Epoch 55 Batch 3100 Loss 0.3816\n","Epoch 55 Batch 3200 Loss 0.4186\n","Epoch 55 Batch 3300 Loss 0.4022\n","Epoch 55 Batch 3400 Loss 0.4108\n","Epoch 55 Batch 3500 Loss 0.4629\n","Epoch 55 Batch 3600 Loss 0.3624\n","Epoch 55 Batch 3700 Loss 0.3831\n","Epoch 55 Batch 3800 Loss 0.3886\n","Epoch 55 Batch 3900 Loss 0.3954\n","Epoch 55 Batch 4000 Loss 0.3850\n","Epoch 55 Batch 4100 Loss 0.3993\n","Epoch 55 Loss 0.405214\n","Time taken for 1 epoch 579.12 sec\n","\n","Epoch 56 Batch 0 Loss 0.3958\n","Epoch 56 Batch 100 Loss 0.4121\n","Epoch 56 Batch 200 Loss 0.3979\n","Epoch 56 Batch 300 Loss 0.4020\n","Epoch 56 Batch 400 Loss 0.3940\n","Epoch 56 Batch 500 Loss 0.4298\n","Epoch 56 Batch 600 Loss 0.4375\n","Epoch 56 Batch 700 Loss 0.3727\n","Epoch 56 Batch 800 Loss 0.4015\n","Epoch 56 Batch 900 Loss 0.4502\n","Epoch 56 Batch 1000 Loss 0.3798\n","Epoch 56 Batch 1100 Loss 0.4154\n","Epoch 56 Batch 1200 Loss 0.4163\n","Epoch 56 Batch 1300 Loss 0.4320\n","Epoch 56 Batch 1400 Loss 0.3487\n","Epoch 56 Batch 1500 Loss 0.4100\n","Epoch 56 Batch 1600 Loss 0.4772\n","Epoch 56 Batch 1700 Loss 0.3942\n","Epoch 56 Batch 1800 Loss 0.3676\n","Epoch 56 Batch 1900 Loss 0.4174\n","Epoch 56 Batch 2000 Loss 0.4074\n","Epoch 56 Batch 2100 Loss 0.4030\n","Epoch 56 Batch 2200 Loss 0.3569\n","Epoch 56 Batch 2300 Loss 0.4197\n","Epoch 56 Batch 2400 Loss 0.4264\n","Epoch 56 Batch 2500 Loss 0.3778\n","Epoch 56 Batch 2600 Loss 0.3713\n","Epoch 56 Batch 2700 Loss 0.4035\n","Epoch 56 Batch 2800 Loss 0.4371\n","Epoch 56 Batch 2900 Loss 0.4285\n","Epoch 56 Batch 3000 Loss 0.4050\n","Epoch 56 Batch 3100 Loss 0.3963\n","Epoch 56 Batch 3200 Loss 0.4245\n","Epoch 56 Batch 3300 Loss 0.4286\n","Epoch 56 Batch 3400 Loss 0.4139\n","Epoch 56 Batch 3500 Loss 0.4048\n","Epoch 56 Batch 3600 Loss 0.3687\n","Epoch 56 Batch 3700 Loss 0.3844\n","Epoch 56 Batch 3800 Loss 0.3764\n","Epoch 56 Batch 3900 Loss 0.4298\n","Epoch 56 Batch 4000 Loss 0.4273\n","Epoch 56 Batch 4100 Loss 0.4046\n","Epoch 56 Loss 0.403015\n","Time taken for 1 epoch 579.48 sec\n","\n","Epoch 57 Batch 0 Loss 0.4143\n","Epoch 57 Batch 100 Loss 0.4229\n","Epoch 57 Batch 200 Loss 0.4327\n","Epoch 57 Batch 300 Loss 0.4293\n","Epoch 57 Batch 400 Loss 0.4269\n","Epoch 57 Batch 500 Loss 0.3759\n","Epoch 57 Batch 600 Loss 0.3996\n","Epoch 57 Batch 700 Loss 0.3574\n","Epoch 57 Batch 800 Loss 0.3924\n","Epoch 57 Batch 900 Loss 0.4230\n","Epoch 57 Batch 1000 Loss 0.3567\n","Epoch 57 Batch 1100 Loss 0.3878\n","Epoch 57 Batch 1200 Loss 0.3817\n","Epoch 57 Batch 1300 Loss 0.4066\n","Epoch 57 Batch 1400 Loss 0.4159\n","Epoch 57 Batch 1500 Loss 0.3877\n","Epoch 57 Batch 1600 Loss 0.4371\n","Epoch 57 Batch 1700 Loss 0.4196\n","Epoch 57 Batch 1800 Loss 0.4035\n","Epoch 57 Batch 1900 Loss 0.3876\n","Epoch 57 Batch 2000 Loss 0.3959\n","Epoch 57 Batch 2100 Loss 0.3548\n","Epoch 57 Batch 2200 Loss 0.3673\n","Epoch 57 Batch 2300 Loss 0.4041\n","Epoch 57 Batch 2400 Loss 0.4235\n","Epoch 57 Batch 2500 Loss 0.4446\n","Epoch 57 Batch 2600 Loss 0.3758\n","Epoch 57 Batch 2700 Loss 0.4062\n","Epoch 57 Batch 2800 Loss 0.3727\n","Epoch 57 Batch 2900 Loss 0.3868\n","Epoch 57 Batch 3000 Loss 0.3988\n","Epoch 57 Batch 3100 Loss 0.3931\n","Epoch 57 Batch 3200 Loss 0.3794\n","Epoch 57 Batch 3300 Loss 0.4251\n","Epoch 57 Batch 3400 Loss 0.4083\n","Epoch 57 Batch 3500 Loss 0.4059\n","Epoch 57 Batch 3600 Loss 0.4095\n","Epoch 57 Batch 3700 Loss 0.4070\n","Epoch 57 Batch 3800 Loss 0.3956\n","Epoch 57 Batch 3900 Loss 0.3729\n","Epoch 57 Batch 4000 Loss 0.4332\n","Epoch 57 Batch 4100 Loss 0.4409\n","Epoch 57 Loss 0.402433\n","Time taken for 1 epoch 578.20 sec\n","\n","Epoch 58 Batch 0 Loss 0.3434\n","Epoch 58 Batch 100 Loss 0.4128\n","Epoch 58 Batch 200 Loss 0.4006\n","Epoch 58 Batch 300 Loss 0.4833\n","Epoch 58 Batch 400 Loss 0.4249\n","Epoch 58 Batch 500 Loss 0.3825\n","Epoch 58 Batch 600 Loss 0.4465\n","Epoch 58 Batch 700 Loss 0.3881\n","Epoch 58 Batch 800 Loss 0.4303\n","Epoch 58 Batch 900 Loss 0.4143\n","Epoch 58 Batch 1000 Loss 0.3377\n","Epoch 58 Batch 1100 Loss 0.4363\n","Epoch 58 Batch 1200 Loss 0.3996\n","Epoch 58 Batch 1300 Loss 0.3950\n","Epoch 58 Batch 1400 Loss 0.4119\n","Epoch 58 Batch 1500 Loss 0.3703\n","Epoch 58 Batch 1600 Loss 0.4225\n","Epoch 58 Batch 1700 Loss 0.4590\n","Epoch 58 Batch 1800 Loss 0.4279\n","Epoch 58 Batch 1900 Loss 0.4629\n","Epoch 58 Batch 2000 Loss 0.4603\n","Epoch 58 Batch 2100 Loss 0.4248\n","Epoch 58 Batch 2200 Loss 0.4046\n","Epoch 58 Batch 2300 Loss 0.3886\n","Epoch 58 Batch 2400 Loss 0.4106\n","Epoch 58 Batch 2500 Loss 0.3933\n","Epoch 58 Batch 2600 Loss 0.3871\n","Epoch 58 Batch 2700 Loss 0.3994\n","Epoch 58 Batch 2800 Loss 0.3863\n","Epoch 58 Batch 2900 Loss 0.3910\n","Epoch 58 Batch 3000 Loss 0.3836\n","Epoch 58 Batch 3100 Loss 0.3628\n","Epoch 58 Batch 3200 Loss 0.4016\n","Epoch 58 Batch 3300 Loss 0.3816\n","Epoch 58 Batch 3400 Loss 0.3897\n","Epoch 58 Batch 3500 Loss 0.4685\n","Epoch 58 Batch 3600 Loss 0.3616\n","Epoch 58 Batch 3700 Loss 0.3864\n","Epoch 58 Batch 3800 Loss 0.3668\n","Epoch 58 Batch 3900 Loss 0.3933\n","Epoch 58 Batch 4000 Loss 0.4113\n","Epoch 58 Batch 4100 Loss 0.3785\n","Epoch 58 Loss 0.401332\n","Time taken for 1 epoch 578.80 sec\n","\n","Epoch 59 Batch 0 Loss 0.3831\n","Epoch 59 Batch 100 Loss 0.4312\n","Epoch 59 Batch 200 Loss 0.3928\n","Epoch 59 Batch 300 Loss 0.4039\n","Epoch 59 Batch 400 Loss 0.4068\n","Epoch 59 Batch 500 Loss 0.3741\n","Epoch 59 Batch 600 Loss 0.4893\n","Epoch 59 Batch 700 Loss 0.3867\n","Epoch 59 Batch 800 Loss 0.4338\n","Epoch 59 Batch 900 Loss 0.3959\n","Epoch 59 Batch 1000 Loss 0.4125\n","Epoch 59 Batch 1100 Loss 0.3878\n","Epoch 59 Batch 1200 Loss 0.3759\n","Epoch 59 Batch 1300 Loss 0.4054\n","Epoch 59 Batch 1400 Loss 0.4201\n","Epoch 59 Batch 1500 Loss 0.4315\n","Epoch 59 Batch 1600 Loss 0.3974\n","Epoch 59 Batch 1700 Loss 0.4085\n","Epoch 59 Batch 1800 Loss 0.4126\n","Epoch 59 Batch 1900 Loss 0.3729\n","Epoch 59 Batch 2000 Loss 0.3521\n","Epoch 59 Batch 2100 Loss 0.4315\n","Epoch 59 Batch 2200 Loss 0.3820\n","Epoch 59 Batch 2300 Loss 0.3986\n","Epoch 59 Batch 2400 Loss 0.3821\n","Epoch 59 Batch 2500 Loss 0.3939\n","Epoch 59 Batch 2600 Loss 0.3598\n","Epoch 59 Batch 2700 Loss 0.3958\n","Epoch 59 Batch 2800 Loss 0.4304\n","Epoch 59 Batch 2900 Loss 0.4250\n","Epoch 59 Batch 3000 Loss 0.3934\n","Epoch 59 Batch 3100 Loss 0.3872\n","Epoch 59 Batch 3200 Loss 0.4392\n","Epoch 59 Batch 3300 Loss 0.3630\n","Epoch 59 Batch 3400 Loss 0.3893\n","Epoch 59 Batch 3500 Loss 0.4153\n","Epoch 59 Batch 3600 Loss 0.3836\n","Epoch 59 Batch 3700 Loss 0.3498\n","Epoch 59 Batch 3800 Loss 0.3725\n","Epoch 59 Batch 3900 Loss 0.3464\n","Epoch 59 Batch 4000 Loss 0.4552\n","Epoch 59 Batch 4100 Loss 0.3584\n","Epoch 59 Loss 0.400467\n","Time taken for 1 epoch 579.42 sec\n","\n","Epoch 60 Batch 0 Loss 0.3936\n","Epoch 60 Batch 100 Loss 0.3886\n","Epoch 60 Batch 200 Loss 0.4233\n","Epoch 60 Batch 300 Loss 0.4046\n","Epoch 60 Batch 400 Loss 0.3548\n","Epoch 60 Batch 500 Loss 0.4069\n","Epoch 60 Batch 600 Loss 0.3993\n","Epoch 60 Batch 700 Loss 0.4106\n","Epoch 60 Batch 800 Loss 0.4084\n","Epoch 60 Batch 900 Loss 0.4355\n","Epoch 60 Batch 1000 Loss 0.4399\n","Epoch 60 Batch 1100 Loss 0.3500\n","Epoch 60 Batch 1200 Loss 0.4601\n","Epoch 60 Batch 1300 Loss 0.3803\n","Epoch 60 Batch 1400 Loss 0.3651\n","Epoch 60 Batch 1500 Loss 0.4166\n","Epoch 60 Batch 1600 Loss 0.3918\n","Epoch 60 Batch 1700 Loss 0.3598\n","Epoch 60 Batch 1800 Loss 0.4189\n","Epoch 60 Batch 1900 Loss 0.4080\n","Epoch 60 Batch 2000 Loss 0.3727\n","Epoch 60 Batch 2100 Loss 0.4073\n","Epoch 60 Batch 2200 Loss 0.4041\n","Epoch 60 Batch 2300 Loss 0.3608\n","Epoch 60 Batch 2400 Loss 0.3736\n","Epoch 60 Batch 2500 Loss 0.3998\n","Epoch 60 Batch 2600 Loss 0.3991\n","Epoch 60 Batch 2700 Loss 0.3866\n","Epoch 60 Batch 2800 Loss 0.4443\n","Epoch 60 Batch 2900 Loss 0.3911\n","Epoch 60 Batch 3000 Loss 0.4229\n","Epoch 60 Batch 3100 Loss 0.4005\n","Epoch 60 Batch 3200 Loss 0.3633\n","Epoch 60 Batch 3300 Loss 0.3942\n","Epoch 60 Batch 3400 Loss 0.4214\n","Epoch 60 Batch 3500 Loss 0.3745\n","Epoch 60 Batch 3600 Loss 0.4146\n","Epoch 60 Batch 3700 Loss 0.3874\n","Epoch 60 Batch 3800 Loss 0.3768\n","Epoch 60 Batch 3900 Loss 0.4216\n","Epoch 60 Batch 4000 Loss 0.3899\n","Epoch 60 Batch 4100 Loss 0.3963\n","Epoch 60 Loss 0.399855\n","Time taken for 1 epoch 578.19 sec\n","\n","Epoch 61 Batch 0 Loss 0.3789\n","Epoch 61 Batch 100 Loss 0.3785\n","Epoch 61 Batch 200 Loss 0.3952\n","Epoch 61 Batch 300 Loss 0.4597\n","Epoch 61 Batch 400 Loss 0.3953\n","Epoch 61 Batch 500 Loss 0.3924\n","Epoch 61 Batch 600 Loss 0.4212\n","Epoch 61 Batch 700 Loss 0.3391\n","Epoch 61 Batch 800 Loss 0.3933\n","Epoch 61 Batch 900 Loss 0.4145\n","Epoch 61 Batch 1000 Loss 0.4171\n","Epoch 61 Batch 1100 Loss 0.3944\n","Epoch 61 Batch 1200 Loss 0.4449\n","Epoch 61 Batch 1300 Loss 0.4325\n","Epoch 61 Batch 1400 Loss 0.4139\n","Epoch 61 Batch 1500 Loss 0.4021\n","Epoch 61 Batch 1600 Loss 0.3918\n","Epoch 61 Batch 1700 Loss 0.4054\n","Epoch 61 Batch 1800 Loss 0.4203\n","Epoch 61 Batch 1900 Loss 0.4364\n","Epoch 61 Batch 2000 Loss 0.3940\n","Epoch 61 Batch 2100 Loss 0.4388\n","Epoch 61 Batch 2200 Loss 0.3995\n","Epoch 61 Batch 2300 Loss 0.3823\n","Epoch 61 Batch 2400 Loss 0.4016\n","Epoch 61 Batch 2500 Loss 0.3911\n","Epoch 61 Batch 2600 Loss 0.4476\n","Epoch 61 Batch 2700 Loss 0.3810\n","Epoch 61 Batch 2800 Loss 0.4194\n","Epoch 61 Batch 2900 Loss 0.4059\n","Epoch 61 Batch 3000 Loss 0.4228\n","Epoch 61 Batch 3100 Loss 0.4427\n","Epoch 61 Batch 3200 Loss 0.4115\n","Epoch 61 Batch 3300 Loss 0.3910\n","Epoch 61 Batch 3400 Loss 0.4137\n","Epoch 61 Batch 3500 Loss 0.4025\n","Epoch 61 Batch 3600 Loss 0.3950\n","Epoch 61 Batch 3700 Loss 0.4097\n","Epoch 61 Batch 3800 Loss 0.3695\n","Epoch 61 Batch 3900 Loss 0.3756\n","Epoch 61 Batch 4000 Loss 0.3928\n","Epoch 61 Batch 4100 Loss 0.4230\n","Epoch 61 Loss 0.400314\n","Time taken for 1 epoch 580.10 sec\n","\n","Epoch 62 Batch 0 Loss 0.3915\n","Epoch 62 Batch 100 Loss 0.3730\n","Epoch 62 Batch 200 Loss 0.4179\n","Epoch 62 Batch 300 Loss 0.3979\n","Epoch 62 Batch 400 Loss 0.3796\n","Epoch 62 Batch 500 Loss 0.3842\n","Epoch 62 Batch 600 Loss 0.4021\n","Epoch 62 Batch 700 Loss 0.4167\n","Epoch 62 Batch 800 Loss 0.4210\n","Epoch 62 Batch 900 Loss 0.3905\n","Epoch 62 Batch 1000 Loss 0.3904\n","Epoch 62 Batch 1100 Loss 0.4250\n","Epoch 62 Batch 1200 Loss 0.4341\n","Epoch 62 Batch 1300 Loss 0.3764\n","Epoch 62 Batch 1400 Loss 0.3781\n","Epoch 62 Batch 1500 Loss 0.4561\n","Epoch 62 Batch 1600 Loss 0.4380\n","Epoch 62 Batch 1700 Loss 0.4180\n","Epoch 62 Batch 1800 Loss 0.3693\n","Epoch 62 Batch 1900 Loss 0.4608\n","Epoch 62 Batch 2000 Loss 0.3905\n","Epoch 62 Batch 2100 Loss 0.4191\n","Epoch 62 Batch 2200 Loss 0.4026\n","Epoch 62 Batch 2300 Loss 0.3420\n","Epoch 62 Batch 2400 Loss 0.4059\n","Epoch 62 Batch 2500 Loss 0.3930\n","Epoch 62 Batch 2600 Loss 0.3903\n","Epoch 62 Batch 2700 Loss 0.3892\n","Epoch 62 Batch 2800 Loss 0.3746\n","Epoch 62 Batch 2900 Loss 0.4283\n","Epoch 62 Batch 3000 Loss 0.4349\n","Epoch 62 Batch 3100 Loss 0.4301\n","Epoch 62 Batch 3200 Loss 0.4057\n","Epoch 62 Batch 3300 Loss 0.4088\n","Epoch 62 Batch 3400 Loss 0.4096\n","Epoch 62 Batch 3500 Loss 0.3683\n","Epoch 62 Batch 3600 Loss 0.4040\n","Epoch 62 Batch 3700 Loss 0.3962\n","Epoch 62 Batch 3800 Loss 0.3567\n","Epoch 62 Batch 3900 Loss 0.4224\n","Epoch 62 Batch 4000 Loss 0.4060\n","Epoch 62 Batch 4100 Loss 0.4185\n","Epoch 62 Loss 0.398985\n","Time taken for 1 epoch 579.11 sec\n","\n","Epoch 63 Batch 0 Loss 0.3928\n","Epoch 63 Batch 100 Loss 0.3988\n","Epoch 63 Batch 200 Loss 0.4107\n","Epoch 63 Batch 300 Loss 0.3754\n","Epoch 63 Batch 400 Loss 0.4007\n","Epoch 63 Batch 500 Loss 0.3793\n","Epoch 63 Batch 600 Loss 0.3938\n","Epoch 63 Batch 700 Loss 0.4093\n","Epoch 63 Batch 800 Loss 0.4128\n","Epoch 63 Batch 900 Loss 0.3874\n","Epoch 63 Batch 1000 Loss 0.4038\n","Epoch 63 Batch 1100 Loss 0.3388\n","Epoch 63 Batch 1200 Loss 0.4207\n","Epoch 63 Batch 1300 Loss 0.3640\n","Epoch 63 Batch 1400 Loss 0.3963\n","Epoch 63 Batch 1500 Loss 0.4265\n","Epoch 63 Batch 1600 Loss 0.3708\n","Epoch 63 Batch 1700 Loss 0.3865\n","Epoch 63 Batch 1800 Loss 0.4271\n","Epoch 63 Batch 1900 Loss 0.4174\n","Epoch 63 Batch 2000 Loss 0.3633\n","Epoch 63 Batch 2100 Loss 0.4151\n","Epoch 63 Batch 2200 Loss 0.4230\n","Epoch 63 Batch 2300 Loss 0.3878\n","Epoch 63 Batch 2400 Loss 0.3973\n","Epoch 63 Batch 2500 Loss 0.3655\n","Epoch 63 Batch 2600 Loss 0.3763\n","Epoch 63 Batch 2700 Loss 0.3936\n","Epoch 63 Batch 2800 Loss 0.3831\n","Epoch 63 Batch 2900 Loss 0.4125\n","Epoch 63 Batch 3000 Loss 0.3707\n","Epoch 63 Batch 3100 Loss 0.3646\n","Epoch 63 Batch 3200 Loss 0.4206\n","Epoch 63 Batch 3300 Loss 0.3979\n","Epoch 63 Batch 3400 Loss 0.4226\n","Epoch 63 Batch 3500 Loss 0.3863\n","Epoch 63 Batch 3600 Loss 0.4265\n","Epoch 63 Batch 3700 Loss 0.3503\n","Epoch 63 Batch 3800 Loss 0.3708\n","Epoch 63 Batch 3900 Loss 0.3815\n","Epoch 63 Batch 4000 Loss 0.3922\n","Epoch 63 Batch 4100 Loss 0.4550\n","Epoch 63 Loss 0.399958\n","Time taken for 1 epoch 579.75 sec\n","\n","Epoch 64 Batch 0 Loss 0.3637\n","Epoch 64 Batch 100 Loss 0.3988\n","Epoch 64 Batch 200 Loss 0.3742\n","Epoch 64 Batch 300 Loss 0.3950\n","Epoch 64 Batch 400 Loss 0.3720\n","Epoch 64 Batch 500 Loss 0.3871\n","Epoch 64 Batch 600 Loss 0.4039\n","Epoch 64 Batch 700 Loss 0.3870\n","Epoch 64 Batch 800 Loss 0.3902\n","Epoch 64 Batch 900 Loss 0.3888\n","Epoch 64 Batch 1000 Loss 0.3808\n","Epoch 64 Batch 1100 Loss 0.3931\n","Epoch 64 Batch 1200 Loss 0.4035\n","Epoch 64 Batch 1300 Loss 0.4367\n","Epoch 64 Batch 1400 Loss 0.4229\n","Epoch 64 Batch 1500 Loss 0.4150\n","Epoch 64 Batch 1600 Loss 0.3722\n","Epoch 64 Batch 1700 Loss 0.3884\n","Epoch 64 Batch 1800 Loss 0.3786\n","Epoch 64 Batch 1900 Loss 0.3880\n","Epoch 64 Batch 2000 Loss 0.4094\n","Epoch 64 Batch 2100 Loss 0.3747\n","Epoch 64 Batch 2200 Loss 0.3837\n","Epoch 64 Batch 2300 Loss 0.4184\n","Epoch 64 Batch 2400 Loss 0.3755\n","Epoch 64 Batch 2500 Loss 0.3555\n","Epoch 64 Batch 2600 Loss 0.4096\n","Epoch 64 Batch 2700 Loss 0.3974\n","Epoch 64 Batch 2800 Loss 0.4151\n","Epoch 64 Batch 2900 Loss 0.4183\n","Epoch 64 Batch 3000 Loss 0.3918\n","Epoch 64 Batch 3100 Loss 0.3794\n","Epoch 64 Batch 3200 Loss 0.3829\n","Epoch 64 Batch 3300 Loss 0.4111\n","Epoch 64 Batch 3400 Loss 0.4241\n","Epoch 64 Batch 3500 Loss 0.3744\n","Epoch 64 Batch 3600 Loss 0.4000\n","Epoch 64 Batch 3700 Loss 0.3743\n","Epoch 64 Batch 3800 Loss 0.4033\n","Epoch 64 Batch 3900 Loss 0.4001\n","Epoch 64 Batch 4000 Loss 0.4069\n","Epoch 64 Batch 4100 Loss 0.3701\n","Epoch 64 Loss 0.396936\n","Time taken for 1 epoch 579.84 sec\n","\n","Epoch 65 Batch 0 Loss 0.3899\n","Epoch 65 Batch 100 Loss 0.3762\n","Epoch 65 Batch 200 Loss 0.4294\n","Epoch 65 Batch 300 Loss 0.3978\n","Epoch 65 Batch 400 Loss 0.3498\n","Epoch 65 Batch 500 Loss 0.3534\n","Epoch 65 Batch 600 Loss 0.4083\n","Epoch 65 Batch 700 Loss 0.3779\n","Epoch 65 Batch 800 Loss 0.3800\n","Epoch 65 Batch 900 Loss 0.4412\n","Epoch 65 Batch 1000 Loss 0.4004\n","Epoch 65 Batch 1100 Loss 0.3906\n","Epoch 65 Batch 1200 Loss 0.4108\n","Epoch 65 Batch 1300 Loss 0.3601\n","Epoch 65 Batch 1400 Loss 0.3755\n","Epoch 65 Batch 1500 Loss 0.3715\n","Epoch 65 Batch 1600 Loss 0.3904\n","Epoch 65 Batch 1700 Loss 0.3878\n","Epoch 65 Batch 1800 Loss 0.3926\n","Epoch 65 Batch 1900 Loss 0.4047\n","Epoch 65 Batch 2000 Loss 0.3876\n","Epoch 65 Batch 2100 Loss 0.4048\n","Epoch 65 Batch 2200 Loss 0.3482\n","Epoch 65 Batch 2300 Loss 0.4036\n","Epoch 65 Batch 2400 Loss 0.3930\n","Epoch 65 Batch 2500 Loss 0.3823\n","Epoch 65 Batch 2600 Loss 0.4248\n","Epoch 65 Batch 2700 Loss 0.3812\n","Epoch 65 Batch 2800 Loss 0.4168\n","Epoch 65 Batch 2900 Loss 0.4396\n","Epoch 65 Batch 3000 Loss 0.4174\n","Epoch 65 Batch 3100 Loss 0.3908\n","Epoch 65 Batch 3200 Loss 0.3691\n","Epoch 65 Batch 3300 Loss 0.4245\n","Epoch 65 Batch 3400 Loss 0.4324\n","Epoch 65 Batch 3500 Loss 0.3978\n","Epoch 65 Batch 3600 Loss 0.4059\n","Epoch 65 Batch 3700 Loss 0.3909\n","Epoch 65 Batch 3800 Loss 0.4196\n","Epoch 65 Batch 3900 Loss 0.3889\n","Epoch 65 Batch 4000 Loss 0.4260\n","Epoch 65 Batch 4100 Loss 0.3708\n","Epoch 65 Loss 0.397582\n","Time taken for 1 epoch 578.88 sec\n","\n","Epoch 66 Batch 0 Loss 0.3794\n","Epoch 66 Batch 100 Loss 0.3654\n","Epoch 66 Batch 200 Loss 0.3668\n","Epoch 66 Batch 300 Loss 0.3853\n","Epoch 66 Batch 400 Loss 0.3499\n","Epoch 66 Batch 500 Loss 0.3996\n","Epoch 66 Batch 600 Loss 0.3966\n","Epoch 66 Batch 700 Loss 0.4070\n","Epoch 66 Batch 800 Loss 0.3840\n","Epoch 66 Batch 900 Loss 0.3957\n","Epoch 66 Batch 1000 Loss 0.3552\n","Epoch 66 Batch 1100 Loss 0.4076\n","Epoch 66 Batch 1200 Loss 0.3941\n","Epoch 66 Batch 1300 Loss 0.4571\n","Epoch 66 Batch 1400 Loss 0.4360\n","Epoch 66 Batch 1500 Loss 0.3750\n","Epoch 66 Batch 1600 Loss 0.3967\n","Epoch 66 Batch 1700 Loss 0.3723\n","Epoch 66 Batch 1800 Loss 0.3546\n","Epoch 66 Batch 1900 Loss 0.3861\n","Epoch 66 Batch 2000 Loss 0.3733\n","Epoch 66 Batch 2100 Loss 0.4398\n","Epoch 66 Batch 2200 Loss 0.3984\n","Epoch 66 Batch 2300 Loss 0.4067\n","Epoch 66 Batch 2400 Loss 0.4095\n","Epoch 66 Batch 2500 Loss 0.3588\n","Epoch 66 Batch 2600 Loss 0.3928\n","Epoch 66 Batch 2700 Loss 0.4107\n","Epoch 66 Batch 2800 Loss 0.4161\n","Epoch 66 Batch 2900 Loss 0.3734\n","Epoch 66 Batch 3000 Loss 0.3987\n","Epoch 66 Batch 3100 Loss 0.3939\n","Epoch 66 Batch 3200 Loss 0.3485\n","Epoch 66 Batch 3300 Loss 0.3900\n","Epoch 66 Batch 3400 Loss 0.4160\n","Epoch 66 Batch 3500 Loss 0.3588\n","Epoch 66 Batch 3600 Loss 0.3819\n","Epoch 66 Batch 3700 Loss 0.3863\n","Epoch 66 Batch 3800 Loss 0.3483\n","Epoch 66 Batch 3900 Loss 0.3918\n","Epoch 66 Batch 4000 Loss 0.3844\n","Epoch 66 Batch 4100 Loss 0.3883\n","Epoch 66 Loss 0.395479\n","Time taken for 1 epoch 580.16 sec\n","\n","Epoch 67 Batch 0 Loss 0.4151\n","Epoch 67 Batch 100 Loss 0.3819\n","Epoch 67 Batch 200 Loss 0.3861\n","Epoch 67 Batch 300 Loss 0.3830\n","Epoch 67 Batch 400 Loss 0.4044\n","Epoch 67 Batch 500 Loss 0.3551\n","Epoch 67 Batch 600 Loss 0.3514\n","Epoch 67 Batch 700 Loss 0.3565\n","Epoch 67 Batch 800 Loss 0.4141\n","Epoch 67 Batch 900 Loss 0.4146\n","Epoch 67 Batch 1000 Loss 0.3928\n","Epoch 67 Batch 1100 Loss 0.4112\n","Epoch 67 Batch 1200 Loss 0.4188\n","Epoch 67 Batch 1300 Loss 0.3561\n","Epoch 67 Batch 1400 Loss 0.3922\n","Epoch 67 Batch 1500 Loss 0.3851\n","Epoch 67 Batch 1600 Loss 0.4492\n","Epoch 67 Batch 1700 Loss 0.3856\n","Epoch 67 Batch 1800 Loss 0.3829\n","Epoch 67 Batch 1900 Loss 0.4115\n","Epoch 67 Batch 2000 Loss 0.4050\n","Epoch 67 Batch 2100 Loss 0.4138\n","Epoch 67 Batch 2200 Loss 0.4161\n","Epoch 67 Batch 2300 Loss 0.3696\n","Epoch 67 Batch 2400 Loss 0.3723\n","Epoch 67 Batch 2500 Loss 0.3992\n","Epoch 67 Batch 2600 Loss 0.3844\n","Epoch 67 Batch 2700 Loss 0.4016\n","Epoch 67 Batch 2800 Loss 0.3954\n","Epoch 67 Batch 2900 Loss 0.3735\n","Epoch 67 Batch 3000 Loss 0.4086\n","Epoch 67 Batch 3100 Loss 0.3434\n","Epoch 67 Batch 3200 Loss 0.3741\n","Epoch 67 Batch 3300 Loss 0.3656\n","Epoch 67 Batch 3400 Loss 0.4012\n","Epoch 67 Batch 3500 Loss 0.3380\n","Epoch 67 Batch 3600 Loss 0.3950\n","Epoch 67 Batch 3700 Loss 0.3870\n","Epoch 67 Batch 3800 Loss 0.4204\n","Epoch 67 Batch 3900 Loss 0.3699\n","Epoch 67 Batch 4000 Loss 0.4054\n","Epoch 67 Batch 4100 Loss 0.3935\n","Epoch 67 Loss 0.396045\n","Time taken for 1 epoch 579.58 sec\n","\n","Epoch 68 Batch 0 Loss 0.4425\n","Epoch 68 Batch 100 Loss 0.4211\n","Epoch 68 Batch 200 Loss 0.4212\n","Epoch 68 Batch 300 Loss 0.3578\n","Epoch 68 Batch 400 Loss 0.3908\n","Epoch 68 Batch 500 Loss 0.3629\n","Epoch 68 Batch 600 Loss 0.3880\n","Epoch 68 Batch 700 Loss 0.4332\n","Epoch 68 Batch 800 Loss 0.4003\n","Epoch 68 Batch 900 Loss 0.4088\n","Epoch 68 Batch 1000 Loss 0.3881\n","Epoch 68 Batch 1100 Loss 0.3876\n","Epoch 68 Batch 1200 Loss 0.3897\n","Epoch 68 Batch 1300 Loss 0.4449\n","Epoch 68 Batch 1400 Loss 0.3400\n","Epoch 68 Batch 1500 Loss 0.3700\n","Epoch 68 Batch 1600 Loss 0.4648\n","Epoch 68 Batch 1700 Loss 0.4177\n","Epoch 68 Batch 1800 Loss 0.3830\n","Epoch 68 Batch 1900 Loss 0.4094\n","Epoch 68 Batch 2000 Loss 0.3494\n","Epoch 68 Batch 2100 Loss 0.4470\n","Epoch 68 Batch 2200 Loss 0.4006\n","Epoch 68 Batch 2300 Loss 0.3614\n","Epoch 68 Batch 2400 Loss 0.3392\n","Epoch 68 Batch 2500 Loss 0.4424\n","Epoch 68 Batch 2600 Loss 0.3842\n","Epoch 68 Batch 2700 Loss 0.3886\n","Epoch 68 Batch 2800 Loss 0.4472\n","Epoch 68 Batch 2900 Loss 0.3886\n","Epoch 68 Batch 3000 Loss 0.4032\n","Epoch 68 Batch 3100 Loss 0.3536\n","Epoch 68 Batch 3200 Loss 0.3996\n","Epoch 68 Batch 3300 Loss 0.3899\n","Epoch 68 Batch 3400 Loss 0.3936\n","Epoch 68 Batch 3500 Loss 0.4252\n","Epoch 68 Batch 3600 Loss 0.3964\n","Epoch 68 Batch 3700 Loss 0.4394\n","Epoch 68 Batch 3800 Loss 0.4275\n","Epoch 68 Batch 3900 Loss 0.3883\n","Epoch 68 Batch 4000 Loss 0.4114\n","Epoch 68 Batch 4100 Loss 0.4044\n","Epoch 68 Loss 0.395300\n","Time taken for 1 epoch 622.34 sec\n","\n","Epoch 69 Batch 0 Loss 0.3977\n","Epoch 69 Batch 100 Loss 0.3799\n","Epoch 69 Batch 200 Loss 0.4200\n","Epoch 69 Batch 300 Loss 0.4006\n","Epoch 69 Batch 400 Loss 0.3822\n","Epoch 69 Batch 500 Loss 0.3799\n","Epoch 69 Batch 600 Loss 0.4031\n","Epoch 69 Batch 700 Loss 0.3874\n","Epoch 69 Batch 800 Loss 0.3965\n","Epoch 69 Batch 900 Loss 0.4215\n","Epoch 69 Batch 1000 Loss 0.4301\n","Epoch 69 Batch 1100 Loss 0.4169\n","Epoch 69 Batch 1200 Loss 0.4315\n","Epoch 69 Batch 1300 Loss 0.3877\n","Epoch 69 Batch 1400 Loss 0.4181\n","Epoch 69 Batch 1500 Loss 0.4035\n","Epoch 69 Batch 1600 Loss 0.4023\n","Epoch 69 Batch 1700 Loss 0.4250\n","Epoch 69 Batch 1800 Loss 0.3715\n","Epoch 69 Batch 1900 Loss 0.3660\n","Epoch 69 Batch 2000 Loss 0.4051\n","Epoch 69 Batch 2100 Loss 0.3989\n","Epoch 69 Batch 2200 Loss 0.4132\n","Epoch 69 Batch 2300 Loss 0.3875\n","Epoch 69 Batch 2400 Loss 0.4497\n","Epoch 69 Batch 2500 Loss 0.4132\n","Epoch 69 Batch 2600 Loss 0.3858\n","Epoch 69 Batch 2700 Loss 0.3956\n","Epoch 69 Batch 2800 Loss 0.4422\n","Epoch 69 Batch 2900 Loss 0.4214\n","Epoch 69 Batch 3000 Loss 0.4060\n","Epoch 69 Batch 3100 Loss 0.3865\n","Epoch 69 Batch 3200 Loss 0.4147\n","Epoch 69 Batch 3300 Loss 0.3940\n","Epoch 69 Batch 3400 Loss 0.3942\n","Epoch 69 Batch 3500 Loss 0.3933\n","Epoch 69 Batch 3600 Loss 0.3637\n","Epoch 69 Batch 3700 Loss 0.4188\n","Epoch 69 Batch 3800 Loss 0.3815\n","Epoch 69 Batch 3900 Loss 0.3933\n","Epoch 69 Batch 4000 Loss 0.3914\n","Epoch 69 Batch 4100 Loss 0.4487\n","Epoch 69 Loss 0.393950\n","Time taken for 1 epoch 578.76 sec\n","\n","Epoch 70 Batch 0 Loss 0.3709\n","Epoch 70 Batch 100 Loss 0.4067\n","Epoch 70 Batch 200 Loss 0.3982\n","Epoch 70 Batch 300 Loss 0.3734\n","Epoch 70 Batch 400 Loss 0.4161\n","Epoch 70 Batch 500 Loss 0.3709\n","Epoch 70 Batch 600 Loss 0.3938\n","Epoch 70 Batch 700 Loss 0.4153\n","Epoch 70 Batch 800 Loss 0.3772\n","Epoch 70 Batch 900 Loss 0.3902\n","Epoch 70 Batch 1000 Loss 0.3647\n","Epoch 70 Batch 1100 Loss 0.4241\n","Epoch 70 Batch 1200 Loss 0.3929\n","Epoch 70 Batch 1300 Loss 0.3964\n","Epoch 70 Batch 1400 Loss 0.3983\n","Epoch 70 Batch 1500 Loss 0.4420\n","Epoch 70 Batch 1600 Loss 0.3909\n","Epoch 70 Batch 1700 Loss 0.4009\n","Epoch 70 Batch 1800 Loss 0.4049\n","Epoch 70 Batch 1900 Loss 0.3939\n","Epoch 70 Batch 2000 Loss 0.3904\n","Epoch 70 Batch 2100 Loss 0.3968\n","Epoch 70 Batch 2200 Loss 0.4113\n","Epoch 70 Batch 2300 Loss 0.3493\n","Epoch 70 Batch 2400 Loss 0.4127\n","Epoch 70 Batch 2500 Loss 0.4094\n","Epoch 70 Batch 2600 Loss 0.3583\n","Epoch 70 Batch 2700 Loss 0.4379\n","Epoch 70 Batch 2800 Loss 0.4189\n","Epoch 70 Batch 2900 Loss 0.4372\n","Epoch 70 Batch 3000 Loss 0.4000\n","Epoch 70 Batch 3100 Loss 0.3458\n","Epoch 70 Batch 3200 Loss 0.3745\n","Epoch 70 Batch 3300 Loss 0.4532\n","Epoch 70 Batch 3400 Loss 0.4100\n","Epoch 70 Batch 3500 Loss 0.3746\n","Epoch 70 Batch 3600 Loss 0.3961\n","Epoch 70 Batch 3700 Loss 0.3962\n","Epoch 70 Batch 3800 Loss 0.3796\n","Epoch 70 Batch 3900 Loss 0.4013\n","Epoch 70 Batch 4000 Loss 0.4007\n","Epoch 70 Batch 4100 Loss 0.4001\n","Epoch 70 Loss 0.393943\n","Time taken for 1 epoch 579.47 sec\n","\n","Epoch 71 Batch 0 Loss 0.4272\n","Epoch 71 Batch 100 Loss 0.4043\n","Epoch 71 Batch 200 Loss 0.3731\n","Epoch 71 Batch 300 Loss 0.4170\n","Epoch 71 Batch 400 Loss 0.3903\n","Epoch 71 Batch 500 Loss 0.4326\n","Epoch 71 Batch 600 Loss 0.4389\n","Epoch 71 Batch 700 Loss 0.3377\n","Epoch 71 Batch 800 Loss 0.3550\n","Epoch 71 Batch 900 Loss 0.4276\n","Epoch 71 Batch 1000 Loss 0.3806\n","Epoch 71 Batch 1100 Loss 0.4041\n","Epoch 71 Batch 1200 Loss 0.3915\n","Epoch 71 Batch 1300 Loss 0.4050\n","Epoch 71 Batch 1400 Loss 0.4181\n","Epoch 71 Batch 1500 Loss 0.3955\n","Epoch 71 Batch 1600 Loss 0.4160\n","Epoch 71 Batch 1700 Loss 0.3913\n","Epoch 71 Batch 1800 Loss 0.3444\n","Epoch 71 Batch 1900 Loss 0.3877\n","Epoch 71 Batch 2000 Loss 0.3972\n","Epoch 71 Batch 2100 Loss 0.4731\n","Epoch 71 Batch 2200 Loss 0.4054\n","Epoch 71 Batch 2300 Loss 0.4059\n","Epoch 71 Batch 2400 Loss 0.4331\n","Epoch 71 Batch 2500 Loss 0.3849\n","Epoch 71 Batch 2600 Loss 0.3926\n","Epoch 71 Batch 2700 Loss 0.3987\n","Epoch 71 Batch 2800 Loss 0.4008\n","Epoch 71 Batch 2900 Loss 0.3771\n","Epoch 71 Batch 3000 Loss 0.4003\n","Epoch 71 Batch 3100 Loss 0.3928\n","Epoch 71 Batch 3200 Loss 0.3724\n","Epoch 71 Batch 3300 Loss 0.4240\n","Epoch 71 Batch 3400 Loss 0.3935\n","Epoch 71 Batch 3500 Loss 0.4086\n","Epoch 71 Batch 3600 Loss 0.4259\n","Epoch 71 Batch 3700 Loss 0.3859\n","Epoch 71 Batch 3800 Loss 0.4040\n","Epoch 71 Batch 3900 Loss 0.3833\n","Epoch 71 Batch 4000 Loss 0.4153\n","Epoch 71 Batch 4100 Loss 0.3899\n","Epoch 71 Loss 0.393318\n","Time taken for 1 epoch 580.59 sec\n","\n","Epoch 72 Batch 0 Loss 0.3916\n","Epoch 72 Batch 100 Loss 0.3829\n","Epoch 72 Batch 200 Loss 0.4183\n","Epoch 72 Batch 300 Loss 0.4133\n","Epoch 72 Batch 400 Loss 0.3467\n","Epoch 72 Batch 500 Loss 0.4011\n","Epoch 72 Batch 600 Loss 0.4172\n","Epoch 72 Batch 700 Loss 0.3924\n","Epoch 72 Batch 800 Loss 0.3902\n","Epoch 72 Batch 900 Loss 0.4485\n","Epoch 72 Batch 1000 Loss 0.4141\n","Epoch 72 Batch 1100 Loss 0.4061\n","Epoch 72 Batch 1200 Loss 0.3910\n","Epoch 72 Batch 1300 Loss 0.4056\n","Epoch 72 Batch 1400 Loss 0.4258\n","Epoch 72 Batch 1500 Loss 0.3833\n","Epoch 72 Batch 1600 Loss 0.3756\n","Epoch 72 Batch 1700 Loss 0.3636\n","Epoch 72 Batch 1800 Loss 0.3670\n","Epoch 72 Batch 1900 Loss 0.3983\n","Epoch 72 Batch 2000 Loss 0.4218\n","Epoch 72 Batch 2100 Loss 0.4145\n","Epoch 72 Batch 2200 Loss 0.3571\n","Epoch 72 Batch 2300 Loss 0.3968\n","Epoch 72 Batch 2400 Loss 0.3810\n","Epoch 72 Batch 2500 Loss 0.4260\n","Epoch 72 Batch 2600 Loss 0.3970\n","Epoch 72 Batch 2700 Loss 0.3842\n","Epoch 72 Batch 2800 Loss 0.3941\n","Epoch 72 Batch 2900 Loss 0.4017\n","Epoch 72 Batch 3000 Loss 0.4163\n","Epoch 72 Batch 3100 Loss 0.4040\n","Epoch 72 Batch 3200 Loss 0.4161\n","Epoch 72 Batch 3300 Loss 0.4379\n","Epoch 72 Batch 3400 Loss 0.3942\n","Epoch 72 Batch 3500 Loss 0.4256\n","Epoch 72 Batch 3600 Loss 0.3858\n","Epoch 72 Batch 3700 Loss 0.4804\n","Epoch 72 Batch 3800 Loss 0.3272\n","Epoch 72 Batch 3900 Loss 0.3478\n","Epoch 72 Batch 4000 Loss 0.4065\n","Epoch 72 Batch 4100 Loss 0.3902\n","Epoch 72 Loss 0.392758\n","Time taken for 1 epoch 579.10 sec\n","\n","Epoch 73 Batch 0 Loss 0.3821\n","Epoch 73 Batch 100 Loss 0.3978\n","Epoch 73 Batch 200 Loss 0.3934\n","Epoch 73 Batch 300 Loss 0.4146\n","Epoch 73 Batch 400 Loss 0.4558\n","Epoch 73 Batch 500 Loss 0.3822\n","Epoch 73 Batch 600 Loss 0.3853\n","Epoch 73 Batch 700 Loss 0.4208\n","Epoch 73 Batch 800 Loss 0.4317\n","Epoch 73 Batch 900 Loss 0.3827\n","Epoch 73 Batch 1000 Loss 0.3471\n","Epoch 73 Batch 1100 Loss 0.4112\n","Epoch 73 Batch 1200 Loss 0.4074\n","Epoch 73 Batch 1300 Loss 0.4327\n","Epoch 73 Batch 1400 Loss 0.3741\n","Epoch 73 Batch 1500 Loss 0.3495\n","Epoch 73 Batch 1600 Loss 0.3905\n","Epoch 73 Batch 1700 Loss 0.3724\n","Epoch 73 Batch 1800 Loss 0.4091\n","Epoch 73 Batch 1900 Loss 0.4081\n","Epoch 73 Batch 2000 Loss 0.3867\n","Epoch 73 Batch 2100 Loss 0.3740\n","Epoch 73 Batch 2200 Loss 0.4013\n","Epoch 73 Batch 2300 Loss 0.3906\n","Epoch 73 Batch 2400 Loss 0.4024\n","Epoch 73 Batch 2500 Loss 0.3757\n","Epoch 73 Batch 2600 Loss 0.3819\n","Epoch 73 Batch 2700 Loss 0.3734\n","Epoch 73 Batch 2800 Loss 0.3771\n","Epoch 73 Batch 2900 Loss 0.3564\n","Epoch 73 Batch 3000 Loss 0.4191\n","Epoch 73 Batch 3100 Loss 0.3903\n","Epoch 73 Batch 3200 Loss 0.3904\n","Epoch 73 Batch 3300 Loss 0.3965\n","Epoch 73 Batch 3400 Loss 0.4294\n","Epoch 73 Batch 3500 Loss 0.3809\n","Epoch 73 Batch 3600 Loss 0.4080\n","Epoch 73 Batch 3700 Loss 0.4217\n","Epoch 73 Batch 3800 Loss 0.3946\n","Epoch 73 Batch 3900 Loss 0.3796\n","Epoch 73 Batch 4000 Loss 0.3883\n","Epoch 73 Batch 4100 Loss 0.4108\n","Epoch 73 Loss 0.392167\n","Time taken for 1 epoch 579.16 sec\n","\n","Epoch 74 Batch 0 Loss 0.3857\n","Epoch 74 Batch 100 Loss 0.4079\n","Epoch 74 Batch 200 Loss 0.3685\n","Epoch 74 Batch 300 Loss 0.3902\n","Epoch 74 Batch 400 Loss 0.3928\n","Epoch 74 Batch 500 Loss 0.3898\n","Epoch 74 Batch 600 Loss 0.4188\n","Epoch 74 Batch 700 Loss 0.4118\n","Epoch 74 Batch 800 Loss 0.4012\n","Epoch 74 Batch 900 Loss 0.3701\n","Epoch 74 Batch 1000 Loss 0.3872\n","Epoch 74 Batch 1100 Loss 0.4605\n","Epoch 74 Batch 1200 Loss 0.3919\n","Epoch 74 Batch 1300 Loss 0.3952\n","Epoch 74 Batch 1400 Loss 0.3733\n","Epoch 74 Batch 1500 Loss 0.3949\n","Epoch 74 Batch 1600 Loss 0.4059\n","Epoch 74 Batch 1700 Loss 0.4223\n","Epoch 74 Batch 1800 Loss 0.3682\n","Epoch 74 Batch 1900 Loss 0.3886\n","Epoch 74 Batch 2000 Loss 0.3670\n","Epoch 74 Batch 2100 Loss 0.3970\n","Epoch 74 Batch 2200 Loss 0.4289\n","Epoch 74 Batch 2300 Loss 0.4124\n","Epoch 74 Batch 2400 Loss 0.4034\n","Epoch 74 Batch 2500 Loss 0.3496\n","Epoch 74 Batch 2600 Loss 0.3721\n","Epoch 74 Batch 2700 Loss 0.3754\n","Epoch 74 Batch 2800 Loss 0.3869\n","Epoch 74 Batch 2900 Loss 0.3975\n","Epoch 74 Batch 3000 Loss 0.4229\n","Epoch 74 Batch 3100 Loss 0.3577\n","Epoch 74 Batch 3200 Loss 0.3807\n","Epoch 74 Batch 3300 Loss 0.3848\n","Epoch 74 Batch 3400 Loss 0.4016\n","Epoch 74 Batch 3500 Loss 0.4052\n","Epoch 74 Batch 3600 Loss 0.4209\n","Epoch 74 Batch 3700 Loss 0.4101\n","Epoch 74 Batch 3800 Loss 0.4047\n","Epoch 74 Batch 3900 Loss 0.3858\n","Epoch 74 Batch 4000 Loss 0.3886\n","Epoch 74 Batch 4100 Loss 0.4366\n","Epoch 74 Loss 0.392150\n","Time taken for 1 epoch 578.97 sec\n","\n","Epoch 75 Batch 0 Loss 0.3700\n","Epoch 75 Batch 100 Loss 0.3911\n","Epoch 75 Batch 200 Loss 0.3914\n","Epoch 75 Batch 300 Loss 0.3505\n","Epoch 75 Batch 400 Loss 0.4089\n","Epoch 75 Batch 500 Loss 0.3554\n","Epoch 75 Batch 600 Loss 0.3728\n","Epoch 75 Batch 700 Loss 0.4374\n","Epoch 75 Batch 800 Loss 0.4264\n","Epoch 75 Batch 900 Loss 0.4105\n","Epoch 75 Batch 1000 Loss 0.4228\n","Epoch 75 Batch 1100 Loss 0.3376\n","Epoch 75 Batch 1200 Loss 0.3913\n","Epoch 75 Batch 1300 Loss 0.4136\n","Epoch 75 Batch 1400 Loss 0.4049\n","Epoch 75 Batch 1500 Loss 0.4043\n","Epoch 75 Batch 1600 Loss 0.3876\n","Epoch 75 Batch 1700 Loss 0.3804\n","Epoch 75 Batch 1800 Loss 0.4085\n","Epoch 75 Batch 1900 Loss 0.3835\n","Epoch 75 Batch 2000 Loss 0.3985\n","Epoch 75 Batch 2100 Loss 0.3577\n","Epoch 75 Batch 2200 Loss 0.3620\n","Epoch 75 Batch 2300 Loss 0.3970\n","Epoch 75 Batch 2400 Loss 0.3803\n","Epoch 75 Batch 2500 Loss 0.3716\n","Epoch 75 Batch 2600 Loss 0.4065\n","Epoch 75 Batch 2700 Loss 0.3797\n","Epoch 75 Batch 2800 Loss 0.3977\n","Epoch 75 Batch 2900 Loss 0.3764\n","Epoch 75 Batch 3000 Loss 0.3920\n","Epoch 75 Batch 3100 Loss 0.4187\n","Epoch 75 Batch 3200 Loss 0.4040\n","Epoch 75 Batch 3300 Loss 0.3905\n","Epoch 75 Batch 3400 Loss 0.3804\n","Epoch 75 Batch 3500 Loss 0.3949\n","Epoch 75 Batch 3600 Loss 0.3922\n","Epoch 75 Batch 3700 Loss 0.4167\n","Epoch 75 Batch 3800 Loss 0.3569\n","Epoch 75 Batch 3900 Loss 0.3482\n","Epoch 75 Batch 4000 Loss 0.3662\n","Epoch 75 Batch 4100 Loss 0.3749\n","Epoch 75 Loss 0.392989\n","Time taken for 1 epoch 579.40 sec\n","\n","Epoch 76 Batch 0 Loss 0.3650\n","Epoch 76 Batch 100 Loss 0.3520\n","Epoch 76 Batch 200 Loss 0.4023\n","Epoch 76 Batch 300 Loss 0.4020\n","Epoch 76 Batch 400 Loss 0.3770\n","Epoch 76 Batch 500 Loss 0.4045\n","Epoch 76 Batch 600 Loss 0.4264\n","Epoch 76 Batch 700 Loss 0.3818\n","Epoch 76 Batch 800 Loss 0.3841\n","Epoch 76 Batch 900 Loss 0.4379\n","Epoch 76 Batch 1000 Loss 0.3624\n","Epoch 76 Batch 1100 Loss 0.3989\n","Epoch 76 Batch 1200 Loss 0.4619\n","Epoch 76 Batch 1300 Loss 0.3910\n","Epoch 76 Batch 1400 Loss 0.3494\n","Epoch 76 Batch 1500 Loss 0.4192\n","Epoch 76 Batch 1600 Loss 0.3655\n","Epoch 76 Batch 1700 Loss 0.3973\n","Epoch 76 Batch 1800 Loss 0.3945\n","Epoch 76 Batch 1900 Loss 0.4224\n","Epoch 76 Batch 2000 Loss 0.3886\n","Epoch 76 Batch 2100 Loss 0.3683\n","Epoch 76 Batch 2200 Loss 0.3711\n","Epoch 76 Batch 2300 Loss 0.3839\n","Epoch 76 Batch 2400 Loss 0.4127\n","Epoch 76 Batch 2500 Loss 0.3707\n","Epoch 76 Batch 2600 Loss 0.3953\n","Epoch 76 Batch 2700 Loss 0.4302\n","Epoch 76 Batch 2800 Loss 0.4037\n","Epoch 76 Batch 2900 Loss 0.4171\n","Epoch 76 Batch 3000 Loss 0.3490\n","Epoch 76 Batch 3100 Loss 0.3335\n","Epoch 76 Batch 3200 Loss 0.3484\n","Epoch 76 Batch 3300 Loss 0.3846\n","Epoch 76 Batch 3400 Loss 0.3876\n","Epoch 76 Batch 3500 Loss 0.3855\n","Epoch 76 Batch 3600 Loss 0.3510\n","Epoch 76 Batch 3700 Loss 0.3612\n","Epoch 76 Batch 3800 Loss 0.4022\n","Epoch 76 Batch 3900 Loss 0.3767\n","Epoch 76 Batch 4000 Loss 0.3446\n","Epoch 76 Batch 4100 Loss 0.3819\n","Epoch 76 Loss 0.393131\n","Time taken for 1 epoch 579.34 sec\n","\n","Epoch 77 Batch 0 Loss 0.4192\n","Epoch 77 Batch 100 Loss 0.3970\n","Epoch 77 Batch 200 Loss 0.3974\n","Epoch 77 Batch 300 Loss 0.4062\n","Epoch 77 Batch 400 Loss 0.4406\n","Epoch 77 Batch 500 Loss 0.3488\n","Epoch 77 Batch 600 Loss 0.4087\n","Epoch 77 Batch 700 Loss 0.3613\n","Epoch 77 Batch 800 Loss 0.3812\n","Epoch 77 Batch 900 Loss 0.3838\n","Epoch 77 Batch 1000 Loss 0.3525\n","Epoch 77 Batch 1100 Loss 0.3777\n","Epoch 77 Batch 1200 Loss 0.3653\n","Epoch 77 Batch 1300 Loss 0.3943\n","Epoch 77 Batch 1400 Loss 0.3589\n","Epoch 77 Batch 1500 Loss 0.3977\n","Epoch 77 Batch 1600 Loss 0.3931\n","Epoch 77 Batch 1700 Loss 0.3860\n","Epoch 77 Batch 1800 Loss 0.3933\n","Epoch 77 Batch 1900 Loss 0.3819\n","Epoch 77 Batch 2000 Loss 0.4126\n","Epoch 77 Batch 2100 Loss 0.3784\n","Epoch 77 Batch 2200 Loss 0.4142\n","Epoch 77 Batch 2300 Loss 0.3590\n","Epoch 77 Batch 2400 Loss 0.3413\n","Epoch 77 Batch 2500 Loss 0.3921\n","Epoch 77 Batch 2600 Loss 0.3484\n","Epoch 77 Batch 2700 Loss 0.3986\n","Epoch 77 Batch 2800 Loss 0.4027\n","Epoch 77 Batch 2900 Loss 0.4372\n","Epoch 77 Batch 3000 Loss 0.4400\n","Epoch 77 Batch 3100 Loss 0.3631\n","Epoch 77 Batch 3200 Loss 0.3780\n","Epoch 77 Batch 3300 Loss 0.3885\n","Epoch 77 Batch 3400 Loss 0.4050\n","Epoch 77 Batch 3500 Loss 0.4047\n","Epoch 77 Batch 3600 Loss 0.3497\n","Epoch 77 Batch 3700 Loss 0.3654\n","Epoch 77 Batch 3800 Loss 0.3879\n","Epoch 77 Batch 3900 Loss 0.3692\n","Epoch 77 Batch 4000 Loss 0.3962\n","Epoch 77 Batch 4100 Loss 0.3953\n","Epoch 77 Loss 0.389858\n","Time taken for 1 epoch 578.51 sec\n","\n","Epoch 78 Batch 0 Loss 0.3784\n","Epoch 78 Batch 100 Loss 0.4124\n","Epoch 78 Batch 200 Loss 0.4184\n","Epoch 78 Batch 300 Loss 0.4304\n","Epoch 78 Batch 400 Loss 0.3973\n","Epoch 78 Batch 500 Loss 0.3891\n","Epoch 78 Batch 600 Loss 0.3486\n","Epoch 78 Batch 700 Loss 0.3561\n","Epoch 78 Batch 800 Loss 0.3991\n","Epoch 78 Batch 900 Loss 0.3847\n","Epoch 78 Batch 1000 Loss 0.4117\n","Epoch 78 Batch 1100 Loss 0.4253\n","Epoch 78 Batch 1200 Loss 0.3747\n","Epoch 78 Batch 1300 Loss 0.4163\n","Epoch 78 Batch 1400 Loss 0.3857\n","Epoch 78 Batch 1500 Loss 0.3880\n","Epoch 78 Batch 1600 Loss 0.3984\n","Epoch 78 Batch 1700 Loss 0.3931\n","Epoch 78 Batch 1800 Loss 0.4102\n","Epoch 78 Batch 1900 Loss 0.4166\n","Epoch 78 Batch 2000 Loss 0.3767\n","Epoch 78 Batch 2100 Loss 0.3984\n","Epoch 78 Batch 2200 Loss 0.3851\n","Epoch 78 Batch 2300 Loss 0.3748\n","Epoch 78 Batch 2400 Loss 0.3637\n","Epoch 78 Batch 2500 Loss 0.3801\n","Epoch 78 Batch 2600 Loss 0.4138\n","Epoch 78 Batch 2700 Loss 0.3807\n","Epoch 78 Batch 2800 Loss 0.4167\n","Epoch 78 Batch 2900 Loss 0.3878\n","Epoch 78 Batch 3000 Loss 0.3964\n","Epoch 78 Batch 3100 Loss 0.3632\n","Epoch 78 Batch 3200 Loss 0.4134\n","Epoch 78 Batch 3300 Loss 0.3908\n","Epoch 78 Batch 3400 Loss 0.3610\n","Epoch 78 Batch 3500 Loss 0.3739\n","Epoch 78 Batch 3600 Loss 0.3746\n","Epoch 78 Batch 3700 Loss 0.4029\n","Epoch 78 Batch 3800 Loss 0.3817\n","Epoch 78 Batch 3900 Loss 0.4740\n","Epoch 78 Batch 4000 Loss 0.3916\n","Epoch 78 Batch 4100 Loss 0.4171\n","Epoch 78 Loss 0.390481\n","Time taken for 1 epoch 579.34 sec\n","\n","Epoch 79 Batch 0 Loss 0.3825\n","Epoch 79 Batch 100 Loss 0.3865\n","Epoch 79 Batch 200 Loss 0.4305\n","Epoch 79 Batch 300 Loss 0.3469\n","Epoch 79 Batch 400 Loss 0.3746\n","Epoch 79 Batch 500 Loss 0.3798\n","Epoch 79 Batch 600 Loss 0.4359\n","Epoch 79 Batch 700 Loss 0.3654\n","Epoch 79 Batch 800 Loss 0.3971\n","Epoch 79 Batch 900 Loss 0.3776\n","Epoch 79 Batch 1000 Loss 0.4015\n","Epoch 79 Batch 1100 Loss 0.3877\n","Epoch 79 Batch 1200 Loss 0.3698\n","Epoch 79 Batch 1300 Loss 0.4077\n","Epoch 79 Batch 1400 Loss 0.3775\n","Epoch 79 Batch 1500 Loss 0.3763\n","Epoch 79 Batch 1600 Loss 0.3681\n","Epoch 79 Batch 1700 Loss 0.4409\n","Epoch 79 Batch 1800 Loss 0.3929\n","Epoch 79 Batch 1900 Loss 0.4285\n","Epoch 79 Batch 2000 Loss 0.3805\n","Epoch 79 Batch 2100 Loss 0.3503\n","Epoch 79 Batch 2200 Loss 0.3984\n","Epoch 79 Batch 2300 Loss 0.3749\n","Epoch 79 Batch 2400 Loss 0.3925\n","Epoch 79 Batch 2500 Loss 0.3630\n","Epoch 79 Batch 2600 Loss 0.3991\n","Epoch 79 Batch 2700 Loss 0.3734\n","Epoch 79 Batch 2800 Loss 0.3803\n","Epoch 79 Batch 2900 Loss 0.4279\n","Epoch 79 Batch 3000 Loss 0.3911\n","Epoch 79 Batch 3100 Loss 0.3867\n","Epoch 79 Batch 3200 Loss 0.3940\n","Epoch 79 Batch 3300 Loss 0.3955\n","Epoch 79 Batch 3400 Loss 0.3731\n","Epoch 79 Batch 3500 Loss 0.3541\n","Epoch 79 Batch 3600 Loss 0.4133\n","Epoch 79 Batch 3700 Loss 0.3667\n","Epoch 79 Batch 3800 Loss 0.4240\n","Epoch 79 Batch 3900 Loss 0.3999\n","Epoch 79 Batch 4000 Loss 0.3979\n","Epoch 79 Batch 4100 Loss 0.3976\n","Epoch 79 Loss 0.389019\n","Time taken for 1 epoch 579.82 sec\n","\n","Epoch 80 Batch 0 Loss 0.3369\n","Epoch 80 Batch 100 Loss 0.3445\n","Epoch 80 Batch 200 Loss 0.3758\n","Epoch 80 Batch 300 Loss 0.3813\n","Epoch 80 Batch 400 Loss 0.3877\n","Epoch 80 Batch 500 Loss 0.3596\n","Epoch 80 Batch 600 Loss 0.3876\n","Epoch 80 Batch 700 Loss 0.3863\n","Epoch 80 Batch 800 Loss 0.4017\n","Epoch 80 Batch 900 Loss 0.4516\n","Epoch 80 Batch 1000 Loss 0.3933\n","Epoch 80 Batch 1100 Loss 0.3975\n","Epoch 80 Batch 1200 Loss 0.4245\n","Epoch 80 Batch 1300 Loss 0.4196\n","Epoch 80 Batch 1400 Loss 0.4459\n","Epoch 80 Batch 1500 Loss 0.3847\n","Epoch 80 Batch 1600 Loss 0.4181\n","Epoch 80 Batch 1700 Loss 0.3976\n","Epoch 80 Batch 1800 Loss 0.3725\n","Epoch 80 Batch 1900 Loss 0.3605\n","Epoch 80 Batch 2000 Loss 0.3919\n","Epoch 80 Batch 2100 Loss 0.4000\n","Epoch 80 Batch 2200 Loss 0.3992\n","Epoch 80 Batch 2300 Loss 0.3805\n","Epoch 80 Batch 2400 Loss 0.4193\n","Epoch 80 Batch 2500 Loss 0.3962\n","Epoch 80 Batch 2600 Loss 0.3979\n","Epoch 80 Batch 2700 Loss 0.3682\n","Epoch 80 Batch 2800 Loss 0.3945\n","Epoch 80 Batch 2900 Loss 0.3971\n","Epoch 80 Batch 3000 Loss 0.3978\n","Epoch 80 Batch 3100 Loss 0.3729\n","Epoch 80 Batch 3200 Loss 0.3798\n","Epoch 80 Batch 3300 Loss 0.3624\n","Epoch 80 Batch 3400 Loss 0.4333\n","Epoch 80 Batch 3500 Loss 0.3979\n","Epoch 80 Batch 3600 Loss 0.4207\n","Epoch 80 Batch 3700 Loss 0.3895\n","Epoch 80 Batch 3800 Loss 0.3795\n","Epoch 80 Batch 3900 Loss 0.3913\n","Epoch 80 Batch 4000 Loss 0.4005\n","Epoch 80 Batch 4100 Loss 0.3900\n","Epoch 80 Loss 0.390355\n","Time taken for 1 epoch 578.89 sec\n","\n","Epoch 81 Batch 0 Loss 0.3958\n","Epoch 81 Batch 100 Loss 0.4016\n","Epoch 81 Batch 200 Loss 0.3956\n","Epoch 81 Batch 300 Loss 0.3728\n","Epoch 81 Batch 400 Loss 0.4242\n","Epoch 81 Batch 500 Loss 0.3818\n","Epoch 81 Batch 600 Loss 0.4336\n","Epoch 81 Batch 700 Loss 0.3546\n","Epoch 81 Batch 800 Loss 0.3704\n","Epoch 81 Batch 900 Loss 0.3855\n","Epoch 81 Batch 1000 Loss 0.3812\n","Epoch 81 Batch 1100 Loss 0.4162\n","Epoch 81 Batch 1200 Loss 0.3930\n","Epoch 81 Batch 1300 Loss 0.4092\n","Epoch 81 Batch 1400 Loss 0.3611\n","Epoch 81 Batch 1500 Loss 0.3881\n","Epoch 81 Batch 1600 Loss 0.4399\n","Epoch 81 Batch 1700 Loss 0.3613\n","Epoch 81 Batch 1800 Loss 0.3744\n","Epoch 81 Batch 1900 Loss 0.3423\n","Epoch 81 Batch 2000 Loss 0.4317\n","Epoch 81 Batch 2100 Loss 0.3904\n","Epoch 81 Batch 2200 Loss 0.3633\n","Epoch 81 Batch 2300 Loss 0.3706\n","Epoch 81 Batch 2400 Loss 0.4033\n","Epoch 81 Batch 2500 Loss 0.4193\n","Epoch 81 Batch 2600 Loss 0.3688\n","Epoch 81 Batch 2700 Loss 0.4185\n","Epoch 81 Batch 2800 Loss 0.3890\n","Epoch 81 Batch 2900 Loss 0.3994\n","Epoch 81 Batch 3000 Loss 0.3954\n","Epoch 81 Batch 3100 Loss 0.4046\n","Epoch 81 Batch 3200 Loss 0.3759\n","Epoch 81 Batch 3300 Loss 0.3939\n","Epoch 81 Batch 3400 Loss 0.3830\n","Epoch 81 Batch 3500 Loss 0.3969\n","Epoch 81 Batch 3600 Loss 0.3743\n","Epoch 81 Batch 3700 Loss 0.4084\n","Epoch 81 Batch 3800 Loss 0.3332\n","Epoch 81 Batch 3900 Loss 0.4228\n","Epoch 81 Batch 4000 Loss 0.4328\n","Epoch 81 Batch 4100 Loss 0.4432\n","Epoch 81 Loss 0.388907\n","Time taken for 1 epoch 579.63 sec\n","\n","Epoch 82 Batch 0 Loss 0.4388\n","Epoch 82 Batch 100 Loss 0.3561\n","Epoch 82 Batch 200 Loss 0.4567\n","Epoch 82 Batch 300 Loss 0.4489\n","Epoch 82 Batch 400 Loss 0.4007\n","Epoch 82 Batch 500 Loss 0.3961\n","Epoch 82 Batch 600 Loss 0.3623\n","Epoch 82 Batch 700 Loss 0.3828\n","Epoch 82 Batch 800 Loss 0.3927\n","Epoch 82 Batch 900 Loss 0.3804\n","Epoch 82 Batch 1000 Loss 0.4453\n","Epoch 82 Batch 1100 Loss 0.4267\n","Epoch 82 Batch 1200 Loss 0.4153\n","Epoch 82 Batch 1300 Loss 0.4545\n","Epoch 82 Batch 1400 Loss 0.3453\n","Epoch 82 Batch 1500 Loss 0.3866\n","Epoch 82 Batch 1600 Loss 0.4219\n","Epoch 82 Batch 1700 Loss 0.3930\n","Epoch 82 Batch 1800 Loss 0.3841\n","Epoch 82 Batch 1900 Loss 0.3575\n","Epoch 82 Batch 2000 Loss 0.4097\n","Epoch 82 Batch 2100 Loss 0.4450\n","Epoch 82 Batch 2200 Loss 0.3871\n","Epoch 82 Batch 2300 Loss 0.3598\n","Epoch 82 Batch 2400 Loss 0.3786\n","Epoch 82 Batch 2500 Loss 0.4015\n","Epoch 82 Batch 2600 Loss 0.3803\n","Epoch 82 Batch 2700 Loss 0.4178\n","Epoch 82 Batch 2800 Loss 0.3478\n","Epoch 82 Batch 2900 Loss 0.3963\n","Epoch 82 Batch 3000 Loss 0.4283\n","Epoch 82 Batch 3100 Loss 0.3512\n","Epoch 82 Batch 3200 Loss 0.3339\n","Epoch 82 Batch 3300 Loss 0.4114\n","Epoch 82 Batch 3400 Loss 0.3918\n","Epoch 82 Batch 3500 Loss 0.4033\n","Epoch 82 Batch 3600 Loss 0.3901\n","Epoch 82 Batch 3700 Loss 0.3752\n","Epoch 82 Batch 3800 Loss 0.4104\n","Epoch 82 Batch 3900 Loss 0.4285\n","Epoch 82 Batch 4000 Loss 0.4211\n","Epoch 82 Batch 4100 Loss 0.4147\n","Epoch 82 Loss 0.389451\n","Time taken for 1 epoch 580.26 sec\n","\n","Epoch 83 Batch 0 Loss 0.3824\n","Epoch 83 Batch 100 Loss 0.3601\n","Epoch 83 Batch 200 Loss 0.3591\n","Epoch 83 Batch 300 Loss 0.4022\n","Epoch 83 Batch 400 Loss 0.3640\n","Epoch 83 Batch 500 Loss 0.3441\n","Epoch 83 Batch 600 Loss 0.4223\n","Epoch 83 Batch 700 Loss 0.4043\n","Epoch 83 Batch 800 Loss 0.3650\n","Epoch 83 Batch 900 Loss 0.4096\n","Epoch 83 Batch 1000 Loss 0.3711\n","Epoch 83 Batch 1100 Loss 0.3996\n","Epoch 83 Batch 1200 Loss 0.4204\n","Epoch 83 Batch 1300 Loss 0.3962\n","Epoch 83 Batch 1400 Loss 0.3951\n","Epoch 83 Batch 1500 Loss 0.3923\n","Epoch 83 Batch 1600 Loss 0.4110\n","Epoch 83 Batch 1700 Loss 0.4056\n","Epoch 83 Batch 1800 Loss 0.3668\n","Epoch 83 Batch 1900 Loss 0.3954\n","Epoch 83 Batch 2000 Loss 0.3644\n","Epoch 83 Batch 2100 Loss 0.3921\n","Epoch 83 Batch 2200 Loss 0.4017\n","Epoch 83 Batch 2300 Loss 0.4223\n","Epoch 83 Batch 2400 Loss 0.3751\n","Epoch 83 Batch 2500 Loss 0.3984\n","Epoch 83 Batch 2600 Loss 0.4360\n","Epoch 83 Batch 2700 Loss 0.4127\n","Epoch 83 Batch 2800 Loss 0.4165\n","Epoch 83 Batch 2900 Loss 0.3998\n","Epoch 83 Batch 3000 Loss 0.4238\n","Epoch 83 Batch 3100 Loss 0.3725\n","Epoch 83 Batch 3200 Loss 0.3884\n","Epoch 83 Batch 3300 Loss 0.4128\n","Epoch 83 Batch 3400 Loss 0.3844\n","Epoch 83 Batch 3500 Loss 0.3461\n","Epoch 83 Batch 3600 Loss 0.4058\n","Epoch 83 Batch 3700 Loss 0.3510\n","Epoch 83 Batch 3800 Loss 0.3842\n","Epoch 83 Batch 3900 Loss 0.3711\n","Epoch 83 Batch 4000 Loss 0.3992\n","Epoch 83 Batch 4100 Loss 0.3993\n","Epoch 83 Loss 0.387981\n","Time taken for 1 epoch 578.96 sec\n","\n","Epoch 84 Batch 0 Loss 0.3450\n","Epoch 84 Batch 100 Loss 0.3537\n","Epoch 84 Batch 200 Loss 0.3630\n","Epoch 84 Batch 300 Loss 0.3924\n","Epoch 84 Batch 400 Loss 0.4025\n","Epoch 84 Batch 500 Loss 0.4011\n","Epoch 84 Batch 600 Loss 0.3850\n","Epoch 84 Batch 700 Loss 0.4015\n","Epoch 84 Batch 800 Loss 0.3987\n","Epoch 84 Batch 900 Loss 0.3936\n","Epoch 84 Batch 1000 Loss 0.3699\n","Epoch 84 Batch 1100 Loss 0.3566\n","Epoch 84 Batch 1200 Loss 0.4183\n","Epoch 84 Batch 1300 Loss 0.3887\n","Epoch 84 Batch 1400 Loss 0.3669\n","Epoch 84 Batch 1500 Loss 0.4065\n","Epoch 84 Batch 1600 Loss 0.3963\n","Epoch 84 Batch 1700 Loss 0.3777\n","Epoch 84 Batch 1800 Loss 0.3665\n","Epoch 84 Batch 1900 Loss 0.4183\n","Epoch 84 Batch 2000 Loss 0.3736\n","Epoch 84 Batch 2100 Loss 0.4096\n","Epoch 84 Batch 2200 Loss 0.4017\n","Epoch 84 Batch 2300 Loss 0.3514\n","Epoch 84 Batch 2400 Loss 0.3942\n","Epoch 84 Batch 2500 Loss 0.4137\n","Epoch 84 Batch 2600 Loss 0.3618\n","Epoch 84 Batch 2700 Loss 0.3762\n","Epoch 84 Batch 2800 Loss 0.4331\n","Epoch 84 Batch 2900 Loss 0.4026\n","Epoch 84 Batch 3000 Loss 0.3906\n","Epoch 84 Batch 3100 Loss 0.3363\n","Epoch 84 Batch 3200 Loss 0.3807\n","Epoch 84 Batch 3300 Loss 0.4223\n","Epoch 84 Batch 3400 Loss 0.3784\n","Epoch 84 Batch 3500 Loss 0.3913\n","Epoch 84 Batch 3600 Loss 0.3632\n","Epoch 84 Batch 3700 Loss 0.3809\n","Epoch 84 Batch 3800 Loss 0.3878\n","Epoch 84 Batch 3900 Loss 0.3776\n","Epoch 84 Batch 4000 Loss 0.3638\n","Epoch 84 Batch 4100 Loss 0.3996\n","Epoch 84 Loss 0.386935\n","Time taken for 1 epoch 578.99 sec\n","\n","Epoch 85 Batch 0 Loss 0.3469\n","Epoch 85 Batch 100 Loss 0.4097\n","Epoch 85 Batch 200 Loss 0.4281\n","Epoch 85 Batch 300 Loss 0.3986\n","Epoch 85 Batch 400 Loss 0.3486\n","Epoch 85 Batch 500 Loss 0.3750\n","Epoch 85 Batch 600 Loss 0.3594\n","Epoch 85 Batch 700 Loss 0.3858\n","Epoch 85 Batch 800 Loss 0.4078\n","Epoch 85 Batch 900 Loss 0.3599\n","Epoch 85 Batch 1000 Loss 0.3601\n","Epoch 85 Batch 1100 Loss 0.3760\n","Epoch 85 Batch 1200 Loss 0.4068\n","Epoch 85 Batch 1300 Loss 0.4473\n","Epoch 85 Batch 1400 Loss 0.3768\n","Epoch 85 Batch 1500 Loss 0.3757\n","Epoch 85 Batch 1600 Loss 0.3830\n","Epoch 85 Batch 1700 Loss 0.3846\n","Epoch 85 Batch 1800 Loss 0.3678\n","Epoch 85 Batch 1900 Loss 0.3801\n","Epoch 85 Batch 2000 Loss 0.3657\n","Epoch 85 Batch 2100 Loss 0.4104\n","Epoch 85 Batch 2200 Loss 0.3686\n","Epoch 85 Batch 2300 Loss 0.3663\n","Epoch 85 Batch 2400 Loss 0.3796\n","Epoch 85 Batch 2500 Loss 0.3729\n","Epoch 85 Batch 2600 Loss 0.3638\n","Epoch 85 Batch 2700 Loss 0.3401\n","Epoch 85 Batch 2800 Loss 0.3726\n","Epoch 85 Batch 2900 Loss 0.3846\n","Epoch 85 Batch 3000 Loss 0.3932\n","Epoch 85 Batch 3100 Loss 0.3512\n","Epoch 85 Batch 3200 Loss 0.4108\n","Epoch 85 Batch 3300 Loss 0.3623\n","Epoch 85 Batch 3400 Loss 0.3914\n","Epoch 85 Batch 3500 Loss 0.4613\n","Epoch 85 Batch 3600 Loss 0.3740\n","Epoch 85 Batch 3700 Loss 0.3849\n","Epoch 85 Batch 3800 Loss 0.3687\n","Epoch 85 Batch 3900 Loss 0.4429\n","Epoch 85 Batch 4000 Loss 0.3824\n","Epoch 85 Batch 4100 Loss 0.3945\n","Epoch 85 Loss 0.387074\n","Time taken for 1 epoch 578.52 sec\n","\n","Epoch 86 Batch 0 Loss 0.3504\n","Epoch 86 Batch 100 Loss 0.4226\n","Epoch 86 Batch 200 Loss 0.3579\n","Epoch 86 Batch 300 Loss 0.4192\n","Epoch 86 Batch 400 Loss 0.3957\n","Epoch 86 Batch 500 Loss 0.4009\n","Epoch 86 Batch 600 Loss 0.3675\n","Epoch 86 Batch 700 Loss 0.3759\n","Epoch 86 Batch 800 Loss 0.3535\n","Epoch 86 Batch 900 Loss 0.4125\n","Epoch 86 Batch 1000 Loss 0.3011\n","Epoch 86 Batch 1100 Loss 0.4141\n","Epoch 86 Batch 1200 Loss 0.4163\n","Epoch 86 Batch 1300 Loss 0.3551\n","Epoch 86 Batch 1400 Loss 0.3897\n","Epoch 86 Batch 1500 Loss 0.4237\n","Epoch 86 Batch 1600 Loss 0.3738\n","Epoch 86 Batch 1700 Loss 0.3589\n","Epoch 86 Batch 1800 Loss 0.3943\n","Epoch 86 Batch 1900 Loss 0.3519\n","Epoch 86 Batch 2000 Loss 0.3946\n","Epoch 86 Batch 2100 Loss 0.3704\n","Epoch 86 Batch 2200 Loss 0.3980\n","Epoch 86 Batch 2300 Loss 0.3548\n","Epoch 86 Batch 2400 Loss 0.3723\n","Epoch 86 Batch 2500 Loss 0.3895\n","Epoch 86 Batch 2600 Loss 0.3967\n","Epoch 86 Batch 2700 Loss 0.3842\n","Epoch 86 Batch 2800 Loss 0.3952\n","Epoch 86 Batch 2900 Loss 0.3966\n","Epoch 86 Batch 3000 Loss 0.3802\n","Epoch 86 Batch 3100 Loss 0.3759\n","Epoch 86 Batch 3200 Loss 0.4023\n","Epoch 86 Batch 3300 Loss 0.3911\n","Epoch 86 Batch 3400 Loss 0.3732\n","Epoch 86 Batch 3500 Loss 0.3591\n","Epoch 86 Batch 3600 Loss 0.3814\n","Epoch 86 Batch 3700 Loss 0.3826\n","Epoch 86 Batch 3800 Loss 0.3762\n","Epoch 86 Batch 3900 Loss 0.3501\n","Epoch 86 Batch 4000 Loss 0.3678\n","Epoch 86 Batch 4100 Loss 0.3736\n","Epoch 86 Loss 0.386462\n","Time taken for 1 epoch 579.98 sec\n","\n","Epoch 87 Batch 0 Loss 0.3894\n","Epoch 87 Batch 100 Loss 0.3769\n","Epoch 87 Batch 200 Loss 0.3725\n","Epoch 87 Batch 300 Loss 0.3770\n","Epoch 87 Batch 400 Loss 0.3919\n","Epoch 87 Batch 500 Loss 0.3648\n","Epoch 87 Batch 600 Loss 0.4267\n","Epoch 87 Batch 700 Loss 0.3755\n","Epoch 87 Batch 800 Loss 0.3908\n","Epoch 87 Batch 900 Loss 0.3968\n","Epoch 87 Batch 1000 Loss 0.3871\n","Epoch 87 Batch 1100 Loss 0.3609\n","Epoch 87 Batch 1200 Loss 0.3524\n","Epoch 87 Batch 1300 Loss 0.4128\n","Epoch 87 Batch 1400 Loss 0.3970\n","Epoch 87 Batch 1500 Loss 0.3964\n","Epoch 87 Batch 1600 Loss 0.3409\n","Epoch 87 Batch 1700 Loss 0.3824\n","Epoch 87 Batch 1800 Loss 0.4076\n","Epoch 87 Batch 1900 Loss 0.3403\n","Epoch 87 Batch 2000 Loss 0.3534\n","Epoch 87 Batch 2100 Loss 0.4409\n","Epoch 87 Batch 2200 Loss 0.4090\n","Epoch 87 Batch 2300 Loss 0.3672\n","Epoch 87 Batch 2400 Loss 0.3826\n","Epoch 87 Batch 2500 Loss 0.4202\n","Epoch 87 Batch 2600 Loss 0.4208\n","Epoch 87 Batch 2700 Loss 0.3796\n","Epoch 87 Batch 2800 Loss 0.3898\n","Epoch 87 Batch 2900 Loss 0.3961\n","Epoch 87 Batch 3000 Loss 0.3571\n","Epoch 87 Batch 3100 Loss 0.3804\n","Epoch 87 Batch 3200 Loss 0.4391\n","Epoch 87 Batch 3300 Loss 0.3866\n","Epoch 87 Batch 3400 Loss 0.3575\n","Epoch 87 Batch 3500 Loss 0.4135\n","Epoch 87 Batch 3600 Loss 0.3601\n","Epoch 87 Batch 3700 Loss 0.4053\n","Epoch 87 Batch 3800 Loss 0.3685\n","Epoch 87 Batch 3900 Loss 0.3651\n","Epoch 87 Batch 4000 Loss 0.3636\n","Epoch 87 Batch 4100 Loss 0.3552\n","Epoch 87 Loss 0.386263\n","Time taken for 1 epoch 579.14 sec\n","\n","Epoch 88 Batch 0 Loss 0.3531\n","Epoch 88 Batch 100 Loss 0.3977\n","Epoch 88 Batch 200 Loss 0.3451\n","Epoch 88 Batch 300 Loss 0.4086\n","Epoch 88 Batch 400 Loss 0.4091\n","Epoch 88 Batch 500 Loss 0.3946\n","Epoch 88 Batch 600 Loss 0.4304\n","Epoch 88 Batch 700 Loss 0.3597\n","Epoch 88 Batch 800 Loss 0.3750\n","Epoch 88 Batch 900 Loss 0.3684\n","Epoch 88 Batch 1000 Loss 0.3332\n","Epoch 88 Batch 1100 Loss 0.3655\n","Epoch 88 Batch 1200 Loss 0.3755\n","Epoch 88 Batch 1300 Loss 0.3798\n","Epoch 88 Batch 1400 Loss 0.3731\n","Epoch 88 Batch 1500 Loss 0.3484\n","Epoch 88 Batch 1600 Loss 0.4204\n","Epoch 88 Batch 1700 Loss 0.3324\n","Epoch 88 Batch 1800 Loss 0.4461\n","Epoch 88 Batch 1900 Loss 0.3706\n","Epoch 88 Batch 2000 Loss 0.4164\n","Epoch 88 Batch 2100 Loss 0.4462\n","Epoch 88 Batch 2200 Loss 0.4339\n","Epoch 88 Batch 2300 Loss 0.4072\n","Epoch 88 Batch 2400 Loss 0.3949\n","Epoch 88 Batch 2500 Loss 0.3694\n","Epoch 88 Batch 2600 Loss 0.4078\n","Epoch 88 Batch 2700 Loss 0.3858\n","Epoch 88 Batch 2800 Loss 0.4202\n","Epoch 88 Batch 2900 Loss 0.3769\n","Epoch 88 Batch 3000 Loss 0.3846\n","Epoch 88 Batch 3100 Loss 0.3655\n","Epoch 88 Batch 3200 Loss 0.3734\n","Epoch 88 Batch 3300 Loss 0.4089\n","Epoch 88 Batch 3400 Loss 0.3981\n","Epoch 88 Batch 3500 Loss 0.3866\n","Epoch 88 Batch 3600 Loss 0.3912\n","Epoch 88 Batch 3700 Loss 0.4483\n","Epoch 88 Batch 3800 Loss 0.3731\n","Epoch 88 Batch 3900 Loss 0.3581\n","Epoch 88 Batch 4000 Loss 0.4223\n","Epoch 88 Batch 4100 Loss 0.3957\n","Epoch 88 Loss 0.386548\n","Time taken for 1 epoch 579.35 sec\n","\n","Epoch 89 Batch 0 Loss 0.3866\n","Epoch 89 Batch 100 Loss 0.3996\n","Epoch 89 Batch 200 Loss 0.4077\n","Epoch 89 Batch 300 Loss 0.3616\n","Epoch 89 Batch 400 Loss 0.3918\n","Epoch 89 Batch 500 Loss 0.3680\n","Epoch 89 Batch 600 Loss 0.4110\n","Epoch 89 Batch 700 Loss 0.3678\n","Epoch 89 Batch 800 Loss 0.3724\n","Epoch 89 Batch 900 Loss 0.3429\n","Epoch 89 Batch 1000 Loss 0.3368\n","Epoch 89 Batch 1100 Loss 0.3911\n","Epoch 89 Batch 1200 Loss 0.4098\n","Epoch 89 Batch 1300 Loss 0.4000\n","Epoch 89 Batch 1400 Loss 0.3375\n","Epoch 89 Batch 1500 Loss 0.4211\n","Epoch 89 Batch 1600 Loss 0.3652\n","Epoch 89 Batch 1700 Loss 0.4047\n","Epoch 89 Batch 1800 Loss 0.3707\n","Epoch 89 Batch 1900 Loss 0.3679\n","Epoch 89 Batch 2000 Loss 0.3942\n","Epoch 89 Batch 2100 Loss 0.3603\n","Epoch 89 Batch 2200 Loss 0.3554\n","Epoch 89 Batch 2300 Loss 0.3875\n","Epoch 89 Batch 2400 Loss 0.3641\n","Epoch 89 Batch 2500 Loss 0.3963\n","Epoch 89 Batch 2600 Loss 0.4192\n","Epoch 89 Batch 2700 Loss 0.3895\n","Epoch 89 Batch 2800 Loss 0.4145\n","Epoch 89 Batch 2900 Loss 0.4275\n","Epoch 89 Batch 3000 Loss 0.3894\n","Epoch 89 Batch 3100 Loss 0.3689\n","Epoch 89 Batch 3200 Loss 0.4016\n","Epoch 89 Batch 3300 Loss 0.4014\n","Epoch 89 Batch 3400 Loss 0.3847\n","Epoch 89 Batch 3500 Loss 0.4252\n","Epoch 89 Batch 3600 Loss 0.3951\n","Epoch 89 Batch 3700 Loss 0.3955\n","Epoch 89 Batch 3800 Loss 0.3867\n","Epoch 89 Batch 3900 Loss 0.3773\n","Epoch 89 Batch 4000 Loss 0.3513\n","Epoch 89 Batch 4100 Loss 0.4044\n","Epoch 89 Loss 0.386111\n","Time taken for 1 epoch 578.77 sec\n","\n","Epoch 90 Batch 0 Loss 0.4013\n","Epoch 90 Batch 100 Loss 0.4028\n","Epoch 90 Batch 200 Loss 0.3784\n","Epoch 90 Batch 300 Loss 0.4364\n","Epoch 90 Batch 400 Loss 0.3680\n","Epoch 90 Batch 500 Loss 0.3278\n","Epoch 90 Batch 600 Loss 0.3520\n","Epoch 90 Batch 700 Loss 0.3947\n","Epoch 90 Batch 800 Loss 0.3908\n","Epoch 90 Batch 900 Loss 0.3881\n","Epoch 90 Batch 1000 Loss 0.3458\n","Epoch 90 Batch 1100 Loss 0.3879\n","Epoch 90 Batch 1200 Loss 0.3941\n","Epoch 90 Batch 1300 Loss 0.4108\n","Epoch 90 Batch 1400 Loss 0.3691\n","Epoch 90 Batch 1500 Loss 0.4392\n","Epoch 90 Batch 1600 Loss 0.3758\n","Epoch 90 Batch 1700 Loss 0.3909\n","Epoch 90 Batch 1800 Loss 0.4140\n","Epoch 90 Batch 1900 Loss 0.3687\n","Epoch 90 Batch 2000 Loss 0.3585\n","Epoch 90 Batch 2100 Loss 0.3834\n","Epoch 90 Batch 2200 Loss 0.3805\n","Epoch 90 Batch 2300 Loss 0.3681\n","Epoch 90 Batch 2400 Loss 0.4318\n","Epoch 90 Batch 2500 Loss 0.3456\n","Epoch 90 Batch 2600 Loss 0.3830\n","Epoch 90 Batch 2700 Loss 0.3921\n","Epoch 90 Batch 2800 Loss 0.3825\n","Epoch 90 Batch 2900 Loss 0.3643\n","Epoch 90 Batch 3000 Loss 0.3867\n","Epoch 90 Batch 3100 Loss 0.3153\n","Epoch 90 Batch 3200 Loss 0.3729\n","Epoch 90 Batch 3300 Loss 0.4680\n","Epoch 90 Batch 3400 Loss 0.3744\n","Epoch 90 Batch 3500 Loss 0.4478\n","Epoch 90 Batch 3600 Loss 0.3974\n","Epoch 90 Batch 3700 Loss 0.3770\n","Epoch 90 Batch 3800 Loss 0.3488\n","Epoch 90 Batch 3900 Loss 0.3667\n","Epoch 90 Batch 4000 Loss 0.3663\n","Epoch 90 Batch 4100 Loss 0.3842\n","Epoch 90 Loss 0.385513\n","Time taken for 1 epoch 579.50 sec\n","\n","Epoch 91 Batch 0 Loss 0.3677\n","Epoch 91 Batch 100 Loss 0.3936\n","Epoch 91 Batch 200 Loss 0.3921\n","Epoch 91 Batch 300 Loss 0.3983\n","Epoch 91 Batch 400 Loss 0.4259\n","Epoch 91 Batch 500 Loss 0.3659\n","Epoch 91 Batch 600 Loss 0.3874\n","Epoch 91 Batch 700 Loss 0.3835\n","Epoch 91 Batch 800 Loss 0.3550\n","Epoch 91 Batch 900 Loss 0.3836\n","Epoch 91 Batch 1000 Loss 0.3345\n","Epoch 91 Batch 1100 Loss 0.3770\n","Epoch 91 Batch 1200 Loss 0.3964\n","Epoch 91 Batch 1300 Loss 0.4067\n","Epoch 91 Batch 1400 Loss 0.3513\n","Epoch 91 Batch 1500 Loss 0.3915\n","Epoch 91 Batch 1600 Loss 0.4027\n","Epoch 91 Batch 1700 Loss 0.3959\n","Epoch 91 Batch 1800 Loss 0.3795\n","Epoch 91 Batch 1900 Loss 0.3782\n","Epoch 91 Batch 2000 Loss 0.3871\n","Epoch 91 Batch 2100 Loss 0.4061\n","Epoch 91 Batch 2200 Loss 0.3563\n","Epoch 91 Batch 2300 Loss 0.3710\n","Epoch 91 Batch 2400 Loss 0.3979\n","Epoch 91 Batch 2500 Loss 0.4051\n","Epoch 91 Batch 2600 Loss 0.4089\n","Epoch 91 Batch 2700 Loss 0.3957\n","Epoch 91 Batch 2800 Loss 0.3792\n","Epoch 91 Batch 2900 Loss 0.3882\n","Epoch 91 Batch 3000 Loss 0.3897\n","Epoch 91 Batch 3100 Loss 0.3671\n","Epoch 91 Batch 3200 Loss 0.4141\n","Epoch 91 Batch 3300 Loss 0.4000\n","Epoch 91 Batch 3400 Loss 0.3756\n","Epoch 91 Batch 3500 Loss 0.3971\n","Epoch 91 Batch 3600 Loss 0.3596\n","Epoch 91 Batch 3700 Loss 0.4361\n","Epoch 91 Batch 3800 Loss 0.3984\n","Epoch 91 Batch 3900 Loss 0.4779\n","Epoch 91 Batch 4000 Loss 0.3907\n","Epoch 91 Batch 4100 Loss 0.4157\n","Epoch 91 Loss 0.385350\n","Time taken for 1 epoch 579.38 sec\n","\n","Epoch 92 Batch 0 Loss 0.3984\n","Epoch 92 Batch 100 Loss 0.3539\n","Epoch 92 Batch 200 Loss 0.3960\n","Epoch 92 Batch 300 Loss 0.3858\n","Epoch 92 Batch 400 Loss 0.3808\n","Epoch 92 Batch 500 Loss 0.4125\n","Epoch 92 Batch 600 Loss 0.3886\n","Epoch 92 Batch 700 Loss 0.3831\n","Epoch 92 Batch 800 Loss 0.3702\n","Epoch 92 Batch 900 Loss 0.3688\n","Epoch 92 Batch 1000 Loss 0.3548\n","Epoch 92 Batch 1100 Loss 0.3914\n","Epoch 92 Batch 1200 Loss 0.3912\n","Epoch 92 Batch 1300 Loss 0.3977\n","Epoch 92 Batch 1400 Loss 0.3853\n","Epoch 92 Batch 1500 Loss 0.3826\n","Epoch 92 Batch 1600 Loss 0.4145\n","Epoch 92 Batch 1700 Loss 0.3857\n","Epoch 92 Batch 1800 Loss 0.3966\n","Epoch 92 Batch 1900 Loss 0.4228\n","Epoch 92 Batch 2000 Loss 0.4052\n","Epoch 92 Batch 2100 Loss 0.4444\n","Epoch 92 Batch 2200 Loss 0.3872\n","Epoch 92 Batch 2300 Loss 0.3774\n","Epoch 92 Batch 2400 Loss 0.3868\n","Epoch 92 Batch 2500 Loss 0.4054\n","Epoch 92 Batch 2600 Loss 0.4292\n","Epoch 92 Batch 2700 Loss 0.4326\n","Epoch 92 Batch 2800 Loss 0.4192\n","Epoch 92 Batch 2900 Loss 0.3635\n","Epoch 92 Batch 3000 Loss 0.3514\n","Epoch 92 Batch 3100 Loss 0.4064\n","Epoch 92 Batch 3200 Loss 0.3592\n","Epoch 92 Batch 3300 Loss 0.3718\n","Epoch 92 Batch 3400 Loss 0.3965\n","Epoch 92 Batch 3500 Loss 0.3922\n","Epoch 92 Batch 3600 Loss 0.4368\n","Epoch 92 Batch 3700 Loss 0.3875\n","Epoch 92 Batch 3800 Loss 0.3652\n","Epoch 92 Batch 3900 Loss 0.3386\n","Epoch 92 Batch 4000 Loss 0.3803\n","Epoch 92 Batch 4100 Loss 0.4393\n","Epoch 92 Loss 0.385810\n","Time taken for 1 epoch 579.01 sec\n","\n","Epoch 93 Batch 0 Loss 0.3751\n","Epoch 93 Batch 100 Loss 0.4056\n","Epoch 93 Batch 200 Loss 0.4127\n","Epoch 93 Batch 300 Loss 0.3760\n","Epoch 93 Batch 400 Loss 0.3466\n","Epoch 93 Batch 500 Loss 0.3637\n","Epoch 93 Batch 600 Loss 0.3883\n","Epoch 93 Batch 700 Loss 0.3890\n","Epoch 93 Batch 800 Loss 0.4255\n","Epoch 93 Batch 900 Loss 0.3866\n","Epoch 93 Batch 1000 Loss 0.3944\n","Epoch 93 Batch 1100 Loss 0.4496\n","Epoch 93 Batch 1200 Loss 0.3637\n","Epoch 93 Batch 1300 Loss 0.4397\n","Epoch 93 Batch 1400 Loss 0.3456\n","Epoch 93 Batch 1500 Loss 0.3896\n","Epoch 93 Batch 1600 Loss 0.4232\n","Epoch 93 Batch 1700 Loss 0.3781\n","Epoch 93 Batch 1800 Loss 0.3133\n","Epoch 93 Batch 1900 Loss 0.3927\n","Epoch 93 Batch 2000 Loss 0.3434\n","Epoch 93 Batch 2100 Loss 0.4337\n","Epoch 93 Batch 2200 Loss 0.3637\n","Epoch 93 Batch 2300 Loss 0.3779\n","Epoch 93 Batch 2400 Loss 0.4110\n","Epoch 93 Batch 2500 Loss 0.3768\n","Epoch 93 Batch 2600 Loss 0.3986\n","Epoch 93 Batch 2700 Loss 0.4065\n","Epoch 93 Batch 2800 Loss 0.3900\n","Epoch 93 Batch 2900 Loss 0.4180\n","Epoch 93 Batch 3000 Loss 0.3448\n","Epoch 93 Batch 3100 Loss 0.3716\n","Epoch 93 Batch 3200 Loss 0.3591\n","Epoch 93 Batch 3300 Loss 0.4048\n","Epoch 93 Batch 3400 Loss 0.3271\n","Epoch 93 Batch 3500 Loss 0.3987\n","Epoch 93 Batch 3600 Loss 0.3677\n","Epoch 93 Batch 3700 Loss 0.4055\n","Epoch 93 Batch 3800 Loss 0.3843\n","Epoch 93 Batch 3900 Loss 0.4084\n","Epoch 93 Batch 4000 Loss 0.3757\n","Epoch 93 Batch 4100 Loss 0.3787\n","Epoch 93 Loss 0.384765\n","Time taken for 1 epoch 578.85 sec\n","\n","Epoch 94 Batch 0 Loss 0.3996\n","Epoch 94 Batch 100 Loss 0.4233\n","Epoch 94 Batch 200 Loss 0.3791\n","Epoch 94 Batch 300 Loss 0.3902\n","Epoch 94 Batch 400 Loss 0.3494\n","Epoch 94 Batch 500 Loss 0.3628\n","Epoch 94 Batch 600 Loss 0.3489\n","Epoch 94 Batch 700 Loss 0.3855\n","Epoch 94 Batch 800 Loss 0.3965\n","Epoch 94 Batch 900 Loss 0.4027\n","Epoch 94 Batch 1000 Loss 0.3845\n","Epoch 94 Batch 1100 Loss 0.3781\n","Epoch 94 Batch 1200 Loss 0.4068\n","Epoch 94 Batch 1300 Loss 0.4109\n","Epoch 94 Batch 1400 Loss 0.3965\n","Epoch 94 Batch 1500 Loss 0.3539\n","Epoch 94 Batch 1600 Loss 0.3871\n","Epoch 94 Batch 1700 Loss 0.3769\n","Epoch 94 Batch 1800 Loss 0.3564\n","Epoch 94 Batch 1900 Loss 0.3632\n","Epoch 94 Batch 2000 Loss 0.4009\n","Epoch 94 Batch 2100 Loss 0.3980\n","Epoch 94 Batch 2200 Loss 0.3883\n","Epoch 94 Batch 2300 Loss 0.3290\n","Epoch 94 Batch 2400 Loss 0.3945\n","Epoch 94 Batch 2500 Loss 0.3778\n","Epoch 94 Batch 2600 Loss 0.3861\n","Epoch 94 Batch 2700 Loss 0.4091\n","Epoch 94 Batch 2800 Loss 0.3940\n","Epoch 94 Batch 2900 Loss 0.3901\n","Epoch 94 Batch 3000 Loss 0.3781\n","Epoch 94 Batch 3100 Loss 0.3875\n","Epoch 94 Batch 3200 Loss 0.3504\n","Epoch 94 Batch 3300 Loss 0.4118\n","Epoch 94 Batch 3400 Loss 0.4081\n","Epoch 94 Batch 3500 Loss 0.3592\n","Epoch 94 Batch 3600 Loss 0.4048\n","Epoch 94 Batch 3700 Loss 0.3655\n","Epoch 94 Batch 3800 Loss 0.3971\n","Epoch 94 Batch 3900 Loss 0.3770\n","Epoch 94 Batch 4000 Loss 0.3619\n","Epoch 94 Batch 4100 Loss 0.3704\n","Epoch 94 Loss 0.384711\n","Time taken for 1 epoch 578.77 sec\n","\n","Epoch 95 Batch 0 Loss 0.3608\n","Epoch 95 Batch 100 Loss 0.3848\n","Epoch 95 Batch 200 Loss 0.3818\n","Epoch 95 Batch 300 Loss 0.3965\n","Epoch 95 Batch 400 Loss 0.4194\n","Epoch 95 Batch 500 Loss 0.3855\n","Epoch 95 Batch 600 Loss 0.3666\n","Epoch 95 Batch 700 Loss 0.3725\n","Epoch 95 Batch 800 Loss 0.4018\n","Epoch 95 Batch 900 Loss 0.4098\n","Epoch 95 Batch 1000 Loss 0.3582\n","Epoch 95 Batch 1100 Loss 0.3535\n","Epoch 95 Batch 1200 Loss 0.4045\n","Epoch 95 Batch 1300 Loss 0.3633\n","Epoch 95 Batch 1400 Loss 0.3754\n","Epoch 95 Batch 1500 Loss 0.3701\n","Epoch 95 Batch 1600 Loss 0.3889\n","Epoch 95 Batch 1700 Loss 0.3765\n","Epoch 95 Batch 1800 Loss 0.3436\n","Epoch 95 Batch 1900 Loss 0.4040\n","Epoch 95 Batch 2000 Loss 0.3862\n","Epoch 95 Batch 2100 Loss 0.4056\n","Epoch 95 Batch 2200 Loss 0.4294\n","Epoch 95 Batch 2300 Loss 0.4167\n","Epoch 95 Batch 2400 Loss 0.3799\n","Epoch 95 Batch 2500 Loss 0.4326\n","Epoch 95 Batch 2600 Loss 0.3444\n","Epoch 95 Batch 2700 Loss 0.4072\n","Epoch 95 Batch 2800 Loss 0.3694\n","Epoch 95 Batch 2900 Loss 0.3667\n","Epoch 95 Batch 3000 Loss 0.4176\n","Epoch 95 Batch 3100 Loss 0.3631\n","Epoch 95 Batch 3200 Loss 0.3921\n","Epoch 95 Batch 3300 Loss 0.3690\n","Epoch 95 Batch 3400 Loss 0.3441\n","Epoch 95 Batch 3500 Loss 0.3437\n","Epoch 95 Batch 3600 Loss 0.3660\n","Epoch 95 Batch 3700 Loss 0.3908\n","Epoch 95 Batch 3800 Loss 0.3481\n","Epoch 95 Batch 3900 Loss 0.3500\n","Epoch 95 Batch 4000 Loss 0.3721\n","Epoch 95 Batch 4100 Loss 0.3775\n","Epoch 95 Loss 0.384622\n","Time taken for 1 epoch 579.22 sec\n","\n","Epoch 96 Batch 0 Loss 0.3652\n","Epoch 96 Batch 100 Loss 0.3500\n","Epoch 96 Batch 200 Loss 0.4427\n","Epoch 96 Batch 300 Loss 0.3864\n","Epoch 96 Batch 400 Loss 0.3921\n","Epoch 96 Batch 500 Loss 0.3944\n","Epoch 96 Batch 600 Loss 0.3871\n","Epoch 96 Batch 700 Loss 0.3710\n","Epoch 96 Batch 800 Loss 0.4233\n","Epoch 96 Batch 900 Loss 0.3751\n","Epoch 96 Batch 1000 Loss 0.3533\n","Epoch 96 Batch 1100 Loss 0.3707\n","Epoch 96 Batch 1200 Loss 0.3760\n","Epoch 96 Batch 1300 Loss 0.3993\n","Epoch 96 Batch 1400 Loss 0.3333\n","Epoch 96 Batch 1500 Loss 0.3635\n","Epoch 96 Batch 1600 Loss 0.3731\n","Epoch 96 Batch 1700 Loss 0.3812\n","Epoch 96 Batch 1800 Loss 0.3820\n","Epoch 96 Batch 1900 Loss 0.4032\n","Epoch 96 Batch 2000 Loss 0.3760\n","Epoch 96 Batch 2100 Loss 0.3786\n","Epoch 96 Batch 2200 Loss 0.3648\n","Epoch 96 Batch 2300 Loss 0.3848\n","Epoch 96 Batch 2400 Loss 0.3962\n","Epoch 96 Batch 2500 Loss 0.4131\n","Epoch 96 Batch 2600 Loss 0.3288\n","Epoch 96 Batch 2700 Loss 0.3811\n","Epoch 96 Batch 2800 Loss 0.3868\n","Epoch 96 Batch 2900 Loss 0.4292\n","Epoch 96 Batch 3000 Loss 0.3792\n","Epoch 96 Batch 3100 Loss 0.3771\n","Epoch 96 Batch 3200 Loss 0.3497\n","Epoch 96 Batch 3300 Loss 0.4185\n","Epoch 96 Batch 3400 Loss 0.3722\n","Epoch 96 Batch 3500 Loss 0.3850\n","Epoch 96 Batch 3600 Loss 0.3860\n","Epoch 96 Batch 3700 Loss 0.3785\n","Epoch 96 Batch 3800 Loss 0.3879\n","Epoch 96 Batch 3900 Loss 0.3731\n","Epoch 96 Batch 4000 Loss 0.3532\n","Epoch 96 Batch 4100 Loss 0.3966\n","Epoch 96 Loss 0.384105\n","Time taken for 1 epoch 580.27 sec\n","\n","Epoch 97 Batch 0 Loss 0.3887\n","Epoch 97 Batch 100 Loss 0.3598\n","Epoch 97 Batch 200 Loss 0.4106\n","Epoch 97 Batch 300 Loss 0.3821\n","Epoch 97 Batch 400 Loss 0.3895\n","Epoch 97 Batch 500 Loss 0.3644\n","Epoch 97 Batch 600 Loss 0.3442\n","Epoch 97 Batch 700 Loss 0.3700\n","Epoch 97 Batch 800 Loss 0.3953\n","Epoch 97 Batch 900 Loss 0.3950\n","Epoch 97 Batch 1000 Loss 0.3823\n","Epoch 97 Batch 1100 Loss 0.3805\n","Epoch 97 Batch 1200 Loss 0.3883\n","Epoch 97 Batch 1300 Loss 0.4135\n","Epoch 97 Batch 1400 Loss 0.3672\n","Epoch 97 Batch 1500 Loss 0.3747\n","Epoch 97 Batch 1600 Loss 0.3936\n","Epoch 97 Batch 1700 Loss 0.3384\n","Epoch 97 Batch 1800 Loss 0.4037\n","Epoch 97 Batch 1900 Loss 0.4001\n","Epoch 97 Batch 2000 Loss 0.4019\n","Epoch 97 Batch 2100 Loss 0.3905\n","Epoch 97 Batch 2200 Loss 0.3968\n","Epoch 97 Batch 2300 Loss 0.3745\n","Epoch 97 Batch 2400 Loss 0.3991\n","Epoch 97 Batch 2500 Loss 0.3779\n","Epoch 97 Batch 2600 Loss 0.4277\n","Epoch 97 Batch 2700 Loss 0.4004\n","Epoch 97 Batch 2800 Loss 0.4178\n","Epoch 97 Batch 2900 Loss 0.4758\n","Epoch 97 Batch 3000 Loss 0.3788\n","Epoch 97 Batch 3100 Loss 0.3609\n","Epoch 97 Batch 3200 Loss 0.3952\n","Epoch 97 Batch 3300 Loss 0.3994\n","Epoch 97 Batch 3400 Loss 0.3749\n","Epoch 97 Batch 3500 Loss 0.3969\n","Epoch 97 Batch 3600 Loss 0.3730\n","Epoch 97 Batch 3700 Loss 0.3363\n","Epoch 97 Batch 3800 Loss 0.3725\n","Epoch 97 Batch 3900 Loss 0.4272\n","Epoch 97 Batch 4000 Loss 0.3629\n","Epoch 97 Batch 4100 Loss 0.3766\n","Epoch 97 Loss 0.383155\n","Time taken for 1 epoch 579.15 sec\n","\n","Epoch 98 Batch 0 Loss 0.3658\n","Epoch 98 Batch 100 Loss 0.3749\n","Epoch 98 Batch 200 Loss 0.4218\n","Epoch 98 Batch 300 Loss 0.3625\n","Epoch 98 Batch 400 Loss 0.3753\n","Epoch 98 Batch 500 Loss 0.4063\n","Epoch 98 Batch 600 Loss 0.4087\n","Epoch 98 Batch 700 Loss 0.3965\n","Epoch 98 Batch 800 Loss 0.3684\n","Epoch 98 Batch 900 Loss 0.4116\n","Epoch 98 Batch 1000 Loss 0.3875\n","Epoch 98 Batch 1100 Loss 0.3846\n","Epoch 98 Batch 1200 Loss 0.3826\n","Epoch 98 Batch 1300 Loss 0.4096\n","Epoch 98 Batch 1400 Loss 0.3657\n","Epoch 98 Batch 1500 Loss 0.4035\n","Epoch 98 Batch 1600 Loss 0.4122\n","Epoch 98 Batch 1700 Loss 0.3632\n","Epoch 98 Batch 1800 Loss 0.3627\n","Epoch 98 Batch 1900 Loss 0.3551\n","Epoch 98 Batch 2000 Loss 0.3273\n","Epoch 98 Batch 2100 Loss 0.4110\n","Epoch 98 Batch 2200 Loss 0.3674\n","Epoch 98 Batch 2300 Loss 0.3659\n","Epoch 98 Batch 2400 Loss 0.4029\n","Epoch 98 Batch 2500 Loss 0.4157\n","Epoch 98 Batch 2600 Loss 0.4332\n","Epoch 98 Batch 2700 Loss 0.3249\n","Epoch 98 Batch 2800 Loss 0.3682\n","Epoch 98 Batch 2900 Loss 0.3994\n","Epoch 98 Batch 3000 Loss 0.3880\n","Epoch 98 Batch 3100 Loss 0.3677\n","Epoch 98 Batch 3200 Loss 0.3444\n","Epoch 98 Batch 3300 Loss 0.3875\n","Epoch 98 Batch 3400 Loss 0.3727\n","Epoch 98 Batch 3500 Loss 0.3627\n","Epoch 98 Batch 3600 Loss 0.4028\n","Epoch 98 Batch 3700 Loss 0.3567\n","Epoch 98 Batch 3800 Loss 0.3890\n","Epoch 98 Batch 3900 Loss 0.3744\n","Epoch 98 Batch 4000 Loss 0.4265\n","Epoch 98 Batch 4100 Loss 0.4187\n","Epoch 98 Loss 0.383011\n","Time taken for 1 epoch 578.61 sec\n","\n","Epoch 99 Batch 0 Loss 0.3542\n","Epoch 99 Batch 100 Loss 0.3816\n","Epoch 99 Batch 200 Loss 0.3884\n","Epoch 99 Batch 300 Loss 0.3386\n","Epoch 99 Batch 400 Loss 0.4072\n","Epoch 99 Batch 500 Loss 0.3706\n","Epoch 99 Batch 600 Loss 0.3708\n","Epoch 99 Batch 700 Loss 0.3886\n","Epoch 99 Batch 800 Loss 0.3970\n","Epoch 99 Batch 900 Loss 0.3819\n","Epoch 99 Batch 1000 Loss 0.3732\n","Epoch 99 Batch 1100 Loss 0.3110\n","Epoch 99 Batch 1200 Loss 0.3803\n","Epoch 99 Batch 1300 Loss 0.4143\n","Epoch 99 Batch 1400 Loss 0.4450\n","Epoch 99 Batch 1500 Loss 0.3678\n","Epoch 99 Batch 1600 Loss 0.4149\n","Epoch 99 Batch 1700 Loss 0.3957\n","Epoch 99 Batch 1800 Loss 0.3798\n","Epoch 99 Batch 1900 Loss 0.3671\n","Epoch 99 Batch 2000 Loss 0.3742\n","Epoch 99 Batch 2100 Loss 0.3661\n","Epoch 99 Batch 2200 Loss 0.3401\n","Epoch 99 Batch 2300 Loss 0.3695\n","Epoch 99 Batch 2400 Loss 0.4126\n","Epoch 99 Batch 2500 Loss 0.4205\n","Epoch 99 Batch 2600 Loss 0.3218\n","Epoch 99 Batch 2700 Loss 0.4273\n","Epoch 99 Batch 2800 Loss 0.3856\n","Epoch 99 Batch 2900 Loss 0.4025\n","Epoch 99 Batch 3000 Loss 0.3987\n","Epoch 99 Batch 3100 Loss 0.3955\n","Epoch 99 Batch 3200 Loss 0.3995\n","Epoch 99 Batch 3300 Loss 0.4212\n","Epoch 99 Batch 3400 Loss 0.3662\n","Epoch 99 Batch 3500 Loss 0.3724\n","Epoch 99 Batch 3600 Loss 0.3910\n","Epoch 99 Batch 3700 Loss 0.3494\n","Epoch 99 Batch 3800 Loss 0.3738\n","Epoch 99 Batch 3900 Loss 0.4076\n","Epoch 99 Batch 4000 Loss 0.4236\n","Epoch 99 Batch 4100 Loss 0.4033\n","Epoch 99 Loss 0.385941\n","Time taken for 1 epoch 579.71 sec\n","\n","Epoch 100 Batch 0 Loss 0.3169\n","Epoch 100 Batch 100 Loss 0.3806\n","Epoch 100 Batch 200 Loss 0.3739\n","Epoch 100 Batch 300 Loss 0.4498\n","Epoch 100 Batch 400 Loss 0.3368\n","Epoch 100 Batch 500 Loss 0.3738\n","Epoch 100 Batch 600 Loss 0.3829\n","Epoch 100 Batch 700 Loss 0.4049\n","Epoch 100 Batch 800 Loss 0.3741\n","Epoch 100 Batch 900 Loss 0.4029\n","Epoch 100 Batch 1000 Loss 0.3792\n","Epoch 100 Batch 1100 Loss 0.3801\n","Epoch 100 Batch 1200 Loss 0.4162\n","Epoch 100 Batch 1300 Loss 0.3924\n","Epoch 100 Batch 1400 Loss 0.3628\n","Epoch 100 Batch 1500 Loss 0.3323\n","Epoch 100 Batch 1600 Loss 0.4616\n","Epoch 100 Batch 1700 Loss 0.4097\n","Epoch 100 Batch 1800 Loss 0.3581\n","Epoch 100 Batch 1900 Loss 0.3670\n","Epoch 100 Batch 2000 Loss 0.3460\n","Epoch 100 Batch 2100 Loss 0.4363\n","Epoch 100 Batch 2200 Loss 0.3871\n","Epoch 100 Batch 2300 Loss 0.3630\n","Epoch 100 Batch 2400 Loss 0.3807\n","Epoch 100 Batch 2500 Loss 0.4008\n","Epoch 100 Batch 2600 Loss 0.3911\n","Epoch 100 Batch 2700 Loss 0.3030\n","Epoch 100 Batch 2800 Loss 0.3744\n","Epoch 100 Batch 2900 Loss 0.3690\n","Epoch 100 Batch 3000 Loss 0.3697\n","Epoch 100 Batch 3100 Loss 0.3729\n","Epoch 100 Batch 3200 Loss 0.3675\n","Epoch 100 Batch 3300 Loss 0.3871\n","Epoch 100 Batch 3400 Loss 0.3783\n","Epoch 100 Batch 3500 Loss 0.3739\n","Epoch 100 Batch 3600 Loss 0.3386\n","Epoch 100 Batch 3700 Loss 0.3477\n","Epoch 100 Batch 3800 Loss 0.3728\n","Epoch 100 Batch 3900 Loss 0.3681\n","Epoch 100 Batch 4000 Loss 0.4077\n","Epoch 100 Batch 4100 Loss 0.4264\n","Epoch 100 Loss 0.382988\n","Time taken for 1 epoch 579.07 sec\n","\n"]}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"95RbKEP4d0LO","outputId":"51b77938-c41a-4307-cf56-96be4423c6c7"},"source":["print(loss_plot)"],"execution_count":null,"outputs":[{"output_type":"stream","name":"stdout","text":["[<tf.Tensor: shape=(), dtype=float32, numpy=0.90094453>, <tf.Tensor: shape=(), dtype=float32, numpy=0.7638719>, <tf.Tensor: shape=(), dtype=float32, numpy=0.7232207>, <tf.Tensor: shape=(), dtype=float32, numpy=0.6939957>, <tf.Tensor: shape=(), dtype=float32, numpy=0.6689623>, <tf.Tensor: shape=(), dtype=float32, numpy=0.6453086>, <tf.Tensor: shape=(), dtype=float32, numpy=0.62263566>, <tf.Tensor: shape=(), dtype=float32, numpy=0.601515>, <tf.Tensor: shape=(), dtype=float32, numpy=0.5814615>, <tf.Tensor: shape=(), dtype=float32, numpy=0.56332266>, <tf.Tensor: shape=(), dtype=float32, numpy=0.5462771>, <tf.Tensor: shape=(), dtype=float32, numpy=0.5313508>, <tf.Tensor: shape=(), dtype=float32, numpy=0.51784617>, <tf.Tensor: shape=(), dtype=float32, numpy=0.5054475>, <tf.Tensor: shape=(), dtype=float32, numpy=0.49425602>, <tf.Tensor: shape=(), dtype=float32, numpy=0.4842329>, <tf.Tensor: shape=(), dtype=float32, numpy=0.4756532>, <tf.Tensor: shape=(), dtype=float32, numpy=0.46715477>, <tf.Tensor: shape=(), dtype=float32, numpy=0.46056616>, <tf.Tensor: shape=(), dtype=float32, numpy=0.4530077>, <tf.Tensor: shape=(), dtype=float32, numpy=0.4472476>, <tf.Tensor: shape=(), dtype=float32, numpy=0.44146737>, <tf.Tensor: shape=(), dtype=float32, numpy=0.43622562>, <tf.Tensor: shape=(), dtype=float32, numpy=0.43108135>, <tf.Tensor: shape=(), dtype=float32, numpy=0.426814>, <tf.Tensor: shape=(), dtype=float32, numpy=0.4227815>, <tf.Tensor: shape=(), dtype=float32, numpy=0.4185351>, <tf.Tensor: shape=(), dtype=float32, numpy=0.4155181>, <tf.Tensor: shape=(), dtype=float32, numpy=0.4115897>, <tf.Tensor: shape=(), dtype=float32, numpy=0.40839642>, <tf.Tensor: shape=(), dtype=float32, numpy=0.40564784>, <tf.Tensor: shape=(), dtype=float32, numpy=0.40245503>, <tf.Tensor: shape=(), dtype=float32, numpy=0.4007412>, <tf.Tensor: shape=(), dtype=float32, numpy=0.39717722>, <tf.Tensor: shape=(), dtype=float32, numpy=0.3950015>, <tf.Tensor: shape=(), dtype=float32, numpy=0.3930092>, <tf.Tensor: shape=(), dtype=float32, numpy=0.39063314>, <tf.Tensor: shape=(), dtype=float32, numpy=0.38842526>, <tf.Tensor: shape=(), dtype=float32, numpy=0.38690692>, <tf.Tensor: shape=(), dtype=float32, numpy=0.38542557>, <tf.Tensor: shape=(), dtype=float32, numpy=0.3829183>, <tf.Tensor: shape=(), dtype=float32, numpy=0.3824856>, <tf.Tensor: shape=(), dtype=float32, numpy=0.3801944>, <tf.Tensor: shape=(), dtype=float32, numpy=0.37789112>, <tf.Tensor: shape=(), dtype=float32, numpy=0.3772205>, <tf.Tensor: shape=(), dtype=float32, numpy=0.3783176>, <tf.Tensor: shape=(), dtype=float32, numpy=0.37521744>, <tf.Tensor: shape=(), dtype=float32, numpy=0.37328938>, <tf.Tensor: shape=(), dtype=float32, numpy=0.37215248>, <tf.Tensor: shape=(), dtype=float32, numpy=0.3711378>, <tf.Tensor: shape=(), dtype=float32, numpy=0.36901116>, <tf.Tensor: shape=(), dtype=float32, numpy=0.3686585>, <tf.Tensor: shape=(), dtype=float32, numpy=0.36683074>, <tf.Tensor: shape=(), dtype=float32, numpy=0.36590007>, <tf.Tensor: shape=(), dtype=float32, numpy=0.36573526>, <tf.Tensor: shape=(), dtype=float32, numpy=0.36425334>, <tf.Tensor: shape=(), dtype=float32, numpy=0.36401367>, <tf.Tensor: shape=(), dtype=float32, numpy=0.36275458>, <tf.Tensor: shape=(), dtype=float32, numpy=0.36212325>, <tf.Tensor: shape=(), dtype=float32, numpy=0.36061552>, <tf.Tensor: shape=(), dtype=float32, numpy=0.360097>, <tf.Tensor: shape=(), dtype=float32, numpy=0.36055627>, <tf.Tensor: shape=(), dtype=float32, numpy=0.35835904>, <tf.Tensor: shape=(), dtype=float32, numpy=0.3578668>, <tf.Tensor: shape=(), dtype=float32, numpy=0.35790023>, <tf.Tensor: shape=(), dtype=float32, numpy=0.35824588>, <tf.Tensor: shape=(), dtype=float32, numpy=0.3549051>, <tf.Tensor: shape=(), dtype=float32, numpy=0.35475862>, <tf.Tensor: shape=(), dtype=float32, numpy=0.35446295>, <tf.Tensor: shape=(), dtype=float32, numpy=0.35482848>, <tf.Tensor: shape=(), dtype=float32, numpy=0.3533987>, <tf.Tensor: shape=(), dtype=float32, numpy=0.35587886>, <tf.Tensor: shape=(), dtype=float32, numpy=0.3556172>, <tf.Tensor: shape=(), dtype=float32, numpy=0.3516192>, <tf.Tensor: shape=(), dtype=float32, numpy=0.35174045>, <tf.Tensor: shape=(), dtype=float32, numpy=0.3519186>, <tf.Tensor: shape=(), dtype=float32, numpy=0.34994477>, <tf.Tensor: shape=(), dtype=float32, numpy=0.35042673>, <tf.Tensor: shape=(), dtype=float32, numpy=0.34909573>, <tf.Tensor: shape=(), dtype=float32, numpy=0.34852022>, <tf.Tensor: shape=(), dtype=float32, numpy=0.34908685>, <tf.Tensor: shape=(), dtype=float32, numpy=0.34680203>, <tf.Tensor: shape=(), dtype=float32, numpy=0.3475141>, <tf.Tensor: shape=(), dtype=float32, numpy=0.3476748>, <tf.Tensor: shape=(), dtype=float32, numpy=0.34696>, <tf.Tensor: shape=(), dtype=float32, numpy=0.34611368>, <tf.Tensor: shape=(), dtype=float32, numpy=0.34693038>, <tf.Tensor: shape=(), dtype=float32, numpy=0.344908>, <tf.Tensor: shape=(), dtype=float32, numpy=0.344427>, <tf.Tensor: shape=(), dtype=float32, numpy=0.3448892>, <tf.Tensor: shape=(), dtype=float32, numpy=0.34601542>, <tf.Tensor: shape=(), dtype=float32, numpy=0.34875157>, <tf.Tensor: shape=(), dtype=float32, numpy=0.34323218>, <tf.Tensor: shape=(), dtype=float32, numpy=0.3429761>, <tf.Tensor: shape=(), dtype=float32, numpy=0.34427267>, <tf.Tensor: shape=(), dtype=float32, numpy=0.3429716>, <tf.Tensor: shape=(), dtype=float32, numpy=0.3422915>, <tf.Tensor: shape=(), dtype=float32, numpy=0.3419776>, <tf.Tensor: shape=(), dtype=float32, numpy=0.34273538>, <tf.Tensor: shape=(), dtype=float32, numpy=0.3413331>]\n"]}]},{"cell_type":"code","metadata":{"id":"Y3hFMjsIt7qG"},"source":["val_dataset = tf.data.Dataset.from_tensor_slices((img_name_val, cap_val))\n","\n","# Use map to load the numpy files in parallel\n","val_dataset = val_dataset.map(lambda item1, item2: tf.numpy_function(\n","          map_func, [item1, item2], [tf.float32, tf.int32]),\n","          num_parallel_calls=tf.data.AUTOTUNE)\n","\n","# Shuffle and batch\n","val_dataset = val_dataset.shuffle(BUFFER_SIZE).batch(BATCH_SIZE)\n","val_dataset = val_dataset.prefetch(buffer_size=tf.data.AUTOTUNE)"],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"Ce9E24kmtcIh","outputId":"715f077c-dc8c-42e3-d337-6ade9666adf1"},"source":["### val epoch\n","val_loss_plot = []\n","start_epoch = 0\n","EPOCHS = 100\n","\n","for epoch in range(start_epoch, EPOCHS):\n","    start = time.time()\n","    total_loss = 0\n","\n","    for (batch, (img_tensor, target)) in enumerate(val_dataset):\n","        batch_loss, t_loss = train_step(img_tensor, target)\n","        total_loss += t_loss\n","\n","        if batch % 100 == 0:\n","            average_batch_loss = batch_loss.numpy()/int(target.shape[1])\n","            print(f'Epoch {epoch+1} Batch {batch} Loss {average_batch_loss:.4f}')\n","    # storing the epoch end loss value to plot later\n","    val_loss_plot.append(total_loss / num_steps)\n","\n","    if epoch % 5 == 0:\n","      ckpt_manager.save()\n","\n","    print(f'Epoch {epoch+1} Loss {total_loss/num_steps:.6f}')\n","    print(f'Time taken for 1 epoch {time.time()-start:.2f} sec\\n')"],"execution_count":null,"outputs":[{"output_type":"stream","name":"stdout","text":["Epoch 1 Batch 0 Loss 1.8246\n","Epoch 1 Batch 100 Loss 0.9224\n","Epoch 1 Batch 200 Loss 1.0037\n","Epoch 1 Batch 300 Loss 0.9701\n","Epoch 1 Batch 400 Loss 0.8850\n","Epoch 1 Batch 500 Loss 0.9055\n","Epoch 1 Batch 600 Loss 0.8569\n","Epoch 1 Batch 700 Loss 1.0189\n","Epoch 1 Batch 800 Loss 0.8572\n","Epoch 1 Batch 900 Loss 0.8787\n","Epoch 1 Batch 1000 Loss 0.8215\n","Epoch 1 Loss 0.235051\n","Time taken for 1 epoch 172.46 sec\n","\n","Epoch 2 Batch 0 Loss 0.7103\n","Epoch 2 Batch 100 Loss 0.8328\n","Epoch 2 Batch 200 Loss 0.7351\n","Epoch 2 Batch 300 Loss 0.8091\n","Epoch 2 Batch 400 Loss 0.7944\n","Epoch 2 Batch 500 Loss 0.7153\n","Epoch 2 Batch 600 Loss 0.7318\n","Epoch 2 Batch 700 Loss 0.8070\n","Epoch 2 Batch 800 Loss 0.7399\n","Epoch 2 Batch 900 Loss 0.7468\n","Epoch 2 Batch 1000 Loss 0.7485\n","Epoch 2 Loss 0.196696\n","Time taken for 1 epoch 147.55 sec\n","\n","Epoch 3 Batch 0 Loss 0.7211\n","Epoch 3 Batch 100 Loss 0.7356\n","Epoch 3 Batch 200 Loss 0.7259\n","Epoch 3 Batch 300 Loss 0.7005\n","Epoch 3 Batch 400 Loss 0.7177\n","Epoch 3 Batch 500 Loss 0.7299\n","Epoch 3 Batch 600 Loss 0.6999\n","Epoch 3 Batch 700 Loss 0.7568\n","Epoch 3 Batch 800 Loss 0.7067\n","Epoch 3 Batch 900 Loss 0.7184\n","Epoch 3 Batch 1000 Loss 0.7475\n","Epoch 3 Loss 0.181533\n","Time taken for 1 epoch 148.31 sec\n","\n","Epoch 4 Batch 0 Loss 0.7294\n","Epoch 4 Batch 100 Loss 0.6568\n","Epoch 4 Batch 200 Loss 0.6783\n","Epoch 4 Batch 300 Loss 0.6628\n","Epoch 4 Batch 400 Loss 0.7069\n","Epoch 4 Batch 500 Loss 0.7536\n","Epoch 4 Batch 600 Loss 0.6210\n","Epoch 4 Batch 700 Loss 0.6344\n","Epoch 4 Batch 800 Loss 0.6707\n","Epoch 4 Batch 900 Loss 0.6750\n","Epoch 4 Batch 1000 Loss 0.6201\n","Epoch 4 Loss 0.168681\n","Time taken for 1 epoch 147.19 sec\n","\n","Epoch 5 Batch 0 Loss 0.6636\n","Epoch 5 Batch 100 Loss 0.5871\n","Epoch 5 Batch 200 Loss 0.6335\n","Epoch 5 Batch 300 Loss 0.6269\n","Epoch 5 Batch 400 Loss 0.5882\n","Epoch 5 Batch 500 Loss 0.6125\n","Epoch 5 Batch 600 Loss 0.6076\n","Epoch 5 Batch 700 Loss 0.5885\n","Epoch 5 Batch 800 Loss 0.5897\n","Epoch 5 Batch 900 Loss 0.6219\n","Epoch 5 Batch 1000 Loss 0.6465\n","Epoch 5 Loss 0.156871\n","Time taken for 1 epoch 147.71 sec\n","\n","Epoch 6 Batch 0 Loss 0.5755\n","Epoch 6 Batch 100 Loss 0.6374\n","Epoch 6 Batch 200 Loss 0.5681\n","Epoch 6 Batch 300 Loss 0.5830\n","Epoch 6 Batch 400 Loss 0.5239\n","Epoch 6 Batch 500 Loss 0.5762\n","Epoch 6 Batch 600 Loss 0.6220\n","Epoch 6 Batch 700 Loss 0.5894\n","Epoch 6 Batch 800 Loss 0.5437\n","Epoch 6 Batch 900 Loss 0.6008\n","Epoch 6 Batch 1000 Loss 0.6124\n","Epoch 6 Loss 0.145888\n","Time taken for 1 epoch 148.09 sec\n","\n","Epoch 7 Batch 0 Loss 0.5853\n","Epoch 7 Batch 100 Loss 0.4974\n","Epoch 7 Batch 200 Loss 0.5241\n","Epoch 7 Batch 300 Loss 0.5249\n","Epoch 7 Batch 400 Loss 0.5365\n","Epoch 7 Batch 500 Loss 0.5315\n","Epoch 7 Batch 600 Loss 0.5226\n","Epoch 7 Batch 700 Loss 0.5469\n","Epoch 7 Batch 800 Loss 0.5195\n","Epoch 7 Batch 900 Loss 0.5861\n","Epoch 7 Batch 1000 Loss 0.5591\n","Epoch 7 Loss 0.135991\n","Time taken for 1 epoch 147.71 sec\n","\n","Epoch 8 Batch 0 Loss 0.5264\n","Epoch 8 Batch 100 Loss 0.5044\n","Epoch 8 Batch 200 Loss 0.5052\n","Epoch 8 Batch 300 Loss 0.5160\n","Epoch 8 Batch 400 Loss 0.5257\n","Epoch 8 Batch 500 Loss 0.5589\n","Epoch 8 Batch 600 Loss 0.5173\n","Epoch 8 Batch 700 Loss 0.5614\n","Epoch 8 Batch 800 Loss 0.4747\n","Epoch 8 Batch 900 Loss 0.5391\n","Epoch 8 Batch 1000 Loss 0.5166\n","Epoch 8 Loss 0.127188\n","Time taken for 1 epoch 146.66 sec\n","\n","Epoch 9 Batch 0 Loss 0.4374\n","Epoch 9 Batch 100 Loss 0.5054\n","Epoch 9 Batch 200 Loss 0.5062\n","Epoch 9 Batch 300 Loss 0.4953\n","Epoch 9 Batch 400 Loss 0.4714\n","Epoch 9 Batch 500 Loss 0.4995\n","Epoch 9 Batch 600 Loss 0.4546\n","Epoch 9 Batch 700 Loss 0.5190\n","Epoch 9 Batch 800 Loss 0.4505\n","Epoch 9 Batch 900 Loss 0.4497\n","Epoch 9 Batch 1000 Loss 0.4719\n","Epoch 9 Loss 0.119466\n","Time taken for 1 epoch 147.60 sec\n","\n","Epoch 10 Batch 0 Loss 0.4589\n","Epoch 10 Batch 100 Loss 0.4188\n","Epoch 10 Batch 200 Loss 0.4204\n","Epoch 10 Batch 300 Loss 0.4163\n","Epoch 10 Batch 400 Loss 0.4601\n","Epoch 10 Batch 500 Loss 0.4727\n","Epoch 10 Batch 600 Loss 0.4179\n","Epoch 10 Batch 700 Loss 0.4407\n","Epoch 10 Batch 800 Loss 0.4270\n","Epoch 10 Batch 900 Loss 0.5212\n","Epoch 10 Batch 1000 Loss 0.4809\n","Epoch 10 Loss 0.112666\n","Time taken for 1 epoch 148.04 sec\n","\n","Epoch 11 Batch 0 Loss 0.4621\n","Epoch 11 Batch 100 Loss 0.4421\n","Epoch 11 Batch 200 Loss 0.4097\n","Epoch 11 Batch 300 Loss 0.4187\n","Epoch 11 Batch 400 Loss 0.4310\n","Epoch 11 Batch 500 Loss 0.4444\n","Epoch 11 Batch 600 Loss 0.3867\n","Epoch 11 Batch 700 Loss 0.4332\n","Epoch 11 Batch 800 Loss 0.4146\n","Epoch 11 Batch 900 Loss 0.4257\n","Epoch 11 Batch 1000 Loss 0.4282\n","Epoch 11 Loss 0.106596\n","Time taken for 1 epoch 146.76 sec\n","\n","Epoch 12 Batch 0 Loss 0.4082\n","Epoch 12 Batch 100 Loss 0.3425\n","Epoch 12 Batch 200 Loss 0.3568\n","Epoch 12 Batch 300 Loss 0.3586\n","Epoch 12 Batch 400 Loss 0.3658\n","Epoch 12 Batch 500 Loss 0.3915\n","Epoch 12 Batch 600 Loss 0.4281\n","Epoch 12 Batch 700 Loss 0.4193\n","Epoch 12 Batch 800 Loss 0.4165\n","Epoch 12 Batch 900 Loss 0.3736\n","Epoch 12 Batch 1000 Loss 0.3925\n","Epoch 12 Loss 0.101284\n","Time taken for 1 epoch 147.63 sec\n","\n","Epoch 13 Batch 0 Loss 0.3896\n","Epoch 13 Batch 100 Loss 0.3737\n","Epoch 13 Batch 200 Loss 0.4076\n","Epoch 13 Batch 300 Loss 0.3371\n","Epoch 13 Batch 400 Loss 0.3783\n","Epoch 13 Batch 500 Loss 0.3939\n","Epoch 13 Batch 600 Loss 0.3519\n","Epoch 13 Batch 700 Loss 0.4165\n","Epoch 13 Batch 800 Loss 0.3582\n","Epoch 13 Batch 900 Loss 0.3838\n","Epoch 13 Batch 1000 Loss 0.3850\n","Epoch 13 Loss 0.096571\n","Time taken for 1 epoch 147.92 sec\n","\n","Epoch 14 Batch 0 Loss 0.3838\n","Epoch 14 Batch 100 Loss 0.3783\n","Epoch 14 Batch 200 Loss 0.3433\n","Epoch 14 Batch 300 Loss 0.3802\n","Epoch 14 Batch 400 Loss 0.3596\n","Epoch 14 Batch 500 Loss 0.3765\n","Epoch 14 Batch 600 Loss 0.3696\n","Epoch 14 Batch 700 Loss 0.3876\n","Epoch 14 Batch 800 Loss 0.3857\n","Epoch 14 Batch 900 Loss 0.3653\n","Epoch 14 Batch 1000 Loss 0.3864\n","Epoch 14 Loss 0.092474\n","Time taken for 1 epoch 147.28 sec\n","\n","Epoch 15 Batch 0 Loss 0.3524\n","Epoch 15 Batch 100 Loss 0.2837\n","Epoch 15 Batch 200 Loss 0.3678\n","Epoch 15 Batch 300 Loss 0.3795\n","Epoch 15 Batch 400 Loss 0.3633\n","Epoch 15 Batch 500 Loss 0.3780\n","Epoch 15 Batch 600 Loss 0.2994\n","Epoch 15 Batch 700 Loss 0.3715\n","Epoch 15 Batch 800 Loss 0.3817\n","Epoch 15 Batch 900 Loss 0.3306\n","Epoch 15 Batch 1000 Loss 0.3485\n","Epoch 15 Loss 0.088728\n","Time taken for 1 epoch 147.41 sec\n","\n","Epoch 16 Batch 0 Loss 0.3567\n","Epoch 16 Batch 100 Loss 0.3408\n","Epoch 16 Batch 200 Loss 0.3635\n","Epoch 16 Batch 300 Loss 0.3102\n","Epoch 16 Batch 400 Loss 0.3383\n","Epoch 16 Batch 500 Loss 0.3343\n","Epoch 16 Batch 600 Loss 0.3462\n","Epoch 16 Batch 700 Loss 0.3426\n","Epoch 16 Batch 800 Loss 0.3355\n","Epoch 16 Batch 900 Loss 0.3775\n","Epoch 16 Batch 1000 Loss 0.3357\n","Epoch 16 Loss 0.085556\n","Time taken for 1 epoch 148.04 sec\n","\n","Epoch 17 Batch 0 Loss 0.3469\n","Epoch 17 Batch 100 Loss 0.3172\n","Epoch 17 Batch 200 Loss 0.3349\n","Epoch 17 Batch 300 Loss 0.3167\n","Epoch 17 Batch 400 Loss 0.3439\n","Epoch 17 Batch 500 Loss 0.3276\n","Epoch 17 Batch 600 Loss 0.3176\n","Epoch 17 Batch 700 Loss 0.3443\n","Epoch 17 Batch 800 Loss 0.3140\n","Epoch 17 Batch 900 Loss 0.3115\n","Epoch 17 Batch 1000 Loss 0.3604\n","Epoch 17 Loss 0.082527\n","Time taken for 1 epoch 147.31 sec\n","\n","Epoch 18 Batch 0 Loss 0.3189\n","Epoch 18 Batch 100 Loss 0.3138\n","Epoch 18 Batch 200 Loss 0.3198\n","Epoch 18 Batch 300 Loss 0.3097\n","Epoch 18 Batch 400 Loss 0.3081\n","Epoch 18 Batch 500 Loss 0.3321\n","Epoch 18 Batch 600 Loss 0.3239\n","Epoch 18 Batch 700 Loss 0.3958\n","Epoch 18 Batch 800 Loss 0.3394\n","Epoch 18 Batch 900 Loss 0.3457\n","Epoch 18 Batch 1000 Loss 0.3331\n","Epoch 18 Loss 0.079792\n","Time taken for 1 epoch 147.00 sec\n","\n","Epoch 19 Batch 0 Loss 0.3069\n","Epoch 19 Batch 100 Loss 0.3784\n","Epoch 19 Batch 200 Loss 0.3126\n","Epoch 19 Batch 300 Loss 0.3158\n","Epoch 19 Batch 400 Loss 0.3262\n","Epoch 19 Batch 500 Loss 0.2903\n","Epoch 19 Batch 600 Loss 0.3611\n","Epoch 19 Batch 700 Loss 0.3288\n","Epoch 19 Batch 800 Loss 0.3077\n","Epoch 19 Batch 900 Loss 0.3349\n","Epoch 19 Batch 1000 Loss 0.3338\n","Epoch 19 Loss 0.077289\n","Time taken for 1 epoch 146.94 sec\n","\n","Epoch 20 Batch 0 Loss 0.3110\n","Epoch 20 Batch 100 Loss 0.2781\n","Epoch 20 Batch 200 Loss 0.2855\n","Epoch 20 Batch 300 Loss 0.2759\n","Epoch 20 Batch 400 Loss 0.2982\n","Epoch 20 Batch 500 Loss 0.2770\n","Epoch 20 Batch 600 Loss 0.2662\n","Epoch 20 Batch 700 Loss 0.2746\n","Epoch 20 Batch 800 Loss 0.2905\n","Epoch 20 Batch 900 Loss 0.2858\n","Epoch 20 Batch 1000 Loss 0.3185\n","Epoch 20 Loss 0.075064\n","Time taken for 1 epoch 146.32 sec\n","\n","Epoch 21 Batch 0 Loss 0.3172\n","Epoch 21 Batch 100 Loss 0.2852\n","Epoch 21 Batch 200 Loss 0.3046\n","Epoch 21 Batch 300 Loss 0.3189\n","Epoch 21 Batch 400 Loss 0.2742\n","Epoch 21 Batch 500 Loss 0.3159\n","Epoch 21 Batch 600 Loss 0.2699\n","Epoch 21 Batch 700 Loss 0.2476\n","Epoch 21 Batch 800 Loss 0.3110\n","Epoch 21 Batch 900 Loss 0.2943\n","Epoch 21 Batch 1000 Loss 0.2984\n","Epoch 21 Loss 0.073126\n","Time taken for 1 epoch 147.57 sec\n","\n","Epoch 22 Batch 0 Loss 0.2649\n","Epoch 22 Batch 100 Loss 0.2862\n","Epoch 22 Batch 200 Loss 0.2839\n","Epoch 22 Batch 300 Loss 0.2679\n","Epoch 22 Batch 400 Loss 0.2944\n","Epoch 22 Batch 500 Loss 0.2723\n","Epoch 22 Batch 600 Loss 0.3136\n","Epoch 22 Batch 700 Loss 0.3215\n","Epoch 22 Batch 800 Loss 0.2818\n","Epoch 22 Batch 900 Loss 0.2640\n","Epoch 22 Batch 1000 Loss 0.2750\n","Epoch 22 Loss 0.071204\n","Time taken for 1 epoch 147.53 sec\n","\n","Epoch 23 Batch 0 Loss 0.3059\n","Epoch 23 Batch 100 Loss 0.2457\n","Epoch 23 Batch 200 Loss 0.2765\n","Epoch 23 Batch 300 Loss 0.2492\n","Epoch 23 Batch 400 Loss 0.2955\n","Epoch 23 Batch 500 Loss 0.2448\n","Epoch 23 Batch 600 Loss 0.2975\n","Epoch 23 Batch 700 Loss 0.2770\n","Epoch 23 Batch 800 Loss 0.2854\n","Epoch 23 Batch 900 Loss 0.2804\n","Epoch 23 Batch 1000 Loss 0.2981\n","Epoch 23 Loss 0.069499\n","Time taken for 1 epoch 147.16 sec\n","\n","Epoch 24 Batch 0 Loss 0.2873\n","Epoch 24 Batch 100 Loss 0.2428\n","Epoch 24 Batch 200 Loss 0.2738\n","Epoch 24 Batch 300 Loss 0.2614\n","Epoch 24 Batch 400 Loss 0.2358\n","Epoch 24 Batch 500 Loss 0.2926\n","Epoch 24 Batch 600 Loss 0.2743\n","Epoch 24 Batch 700 Loss 0.2832\n","Epoch 24 Batch 800 Loss 0.2628\n","Epoch 24 Batch 900 Loss 0.2746\n","Epoch 24 Batch 1000 Loss 0.2842\n","Epoch 24 Loss 0.067786\n","Time taken for 1 epoch 146.93 sec\n","\n","Epoch 25 Batch 0 Loss 0.2756\n","Epoch 25 Batch 100 Loss 0.2394\n","Epoch 25 Batch 200 Loss 0.2802\n","Epoch 25 Batch 300 Loss 0.2549\n","Epoch 25 Batch 400 Loss 0.2589\n","Epoch 25 Batch 500 Loss 0.2502\n","Epoch 25 Batch 600 Loss 0.2646\n","Epoch 25 Batch 700 Loss 0.3095\n","Epoch 25 Batch 800 Loss 0.2544\n","Epoch 25 Batch 900 Loss 0.2602\n","Epoch 25 Batch 1000 Loss 0.2999\n","Epoch 25 Loss 0.066483\n","Time taken for 1 epoch 146.95 sec\n","\n","Epoch 26 Batch 0 Loss 0.2851\n","Epoch 26 Batch 100 Loss 0.2205\n","Epoch 26 Batch 200 Loss 0.2358\n","Epoch 26 Batch 300 Loss 0.2647\n","Epoch 26 Batch 400 Loss 0.2481\n","Epoch 26 Batch 500 Loss 0.2413\n","Epoch 26 Batch 600 Loss 0.2298\n","Epoch 26 Batch 700 Loss 0.2641\n","Epoch 26 Batch 800 Loss 0.2574\n","Epoch 26 Batch 900 Loss 0.2811\n","Epoch 26 Batch 1000 Loss 0.2747\n","Epoch 26 Loss 0.065098\n","Time taken for 1 epoch 147.31 sec\n","\n","Epoch 27 Batch 0 Loss 0.2589\n","Epoch 27 Batch 100 Loss 0.2930\n","Epoch 27 Batch 200 Loss 0.2591\n","Epoch 27 Batch 300 Loss 0.2656\n","Epoch 27 Batch 400 Loss 0.2836\n","Epoch 27 Batch 500 Loss 0.2890\n","Epoch 27 Batch 600 Loss 0.2608\n","Epoch 27 Batch 700 Loss 0.2497\n","Epoch 27 Batch 800 Loss 0.2522\n","Epoch 27 Batch 900 Loss 0.2765\n","Epoch 27 Batch 1000 Loss 0.2721\n","Epoch 27 Loss 0.063770\n","Time taken for 1 epoch 147.09 sec\n","\n","Epoch 28 Batch 0 Loss 0.2660\n","Epoch 28 Batch 100 Loss 0.2363\n","Epoch 28 Batch 200 Loss 0.2762\n","Epoch 28 Batch 300 Loss 0.2924\n","Epoch 28 Batch 400 Loss 0.2827\n","Epoch 28 Batch 500 Loss 0.2804\n","Epoch 28 Batch 600 Loss 0.2428\n","Epoch 28 Batch 700 Loss 0.2712\n","Epoch 28 Batch 800 Loss 0.2536\n","Epoch 28 Batch 900 Loss 0.2522\n","Epoch 28 Batch 1000 Loss 0.2653\n","Epoch 28 Loss 0.062586\n","Time taken for 1 epoch 147.12 sec\n","\n","Epoch 29 Batch 0 Loss 0.2306\n","Epoch 29 Batch 100 Loss 0.2330\n","Epoch 29 Batch 200 Loss 0.2528\n","Epoch 29 Batch 300 Loss 0.2367\n","Epoch 29 Batch 400 Loss 0.2534\n","Epoch 29 Batch 500 Loss 0.2707\n","Epoch 29 Batch 600 Loss 0.2258\n","Epoch 29 Batch 700 Loss 0.2557\n","Epoch 29 Batch 800 Loss 0.2516\n","Epoch 29 Batch 900 Loss 0.2289\n","Epoch 29 Batch 1000 Loss 0.2393\n","Epoch 29 Loss 0.061399\n","Time taken for 1 epoch 147.02 sec\n","\n","Epoch 30 Batch 0 Loss 0.2849\n","Epoch 30 Batch 100 Loss 0.2340\n","Epoch 30 Batch 200 Loss 0.2220\n","Epoch 30 Batch 300 Loss 0.2454\n","Epoch 30 Batch 400 Loss 0.2289\n","Epoch 30 Batch 500 Loss 0.2592\n","Epoch 30 Batch 600 Loss 0.1974\n","Epoch 30 Batch 700 Loss 0.2302\n","Epoch 30 Batch 800 Loss 0.2373\n","Epoch 30 Batch 900 Loss 0.2417\n","Epoch 30 Batch 1000 Loss 0.2445\n","Epoch 30 Loss 0.060463\n","Time taken for 1 epoch 147.58 sec\n","\n","Epoch 31 Batch 0 Loss 0.2612\n","Epoch 31 Batch 100 Loss 0.2305\n","Epoch 31 Batch 200 Loss 0.2083\n","Epoch 31 Batch 300 Loss 0.2509\n","Epoch 31 Batch 400 Loss 0.2378\n","Epoch 31 Batch 500 Loss 0.2193\n","Epoch 31 Batch 600 Loss 0.2790\n","Epoch 31 Batch 700 Loss 0.2403\n","Epoch 31 Batch 800 Loss 0.2589\n","Epoch 31 Batch 900 Loss 0.2232\n","Epoch 31 Batch 1000 Loss 0.2448\n","Epoch 31 Loss 0.059457\n","Time taken for 1 epoch 147.48 sec\n","\n","Epoch 32 Batch 0 Loss 0.2317\n","Epoch 32 Batch 100 Loss 0.2651\n","Epoch 32 Batch 200 Loss 0.2283\n","Epoch 32 Batch 300 Loss 0.2388\n","Epoch 32 Batch 400 Loss 0.2497\n","Epoch 32 Batch 500 Loss 0.2313\n","Epoch 32 Batch 600 Loss 0.2313\n","Epoch 32 Batch 700 Loss 0.2321\n","Epoch 32 Batch 800 Loss 0.2291\n","Epoch 32 Batch 900 Loss 0.2452\n","Epoch 32 Batch 1000 Loss 0.2165\n","Epoch 32 Loss 0.058425\n","Time taken for 1 epoch 147.69 sec\n","\n","Epoch 33 Batch 0 Loss 0.2797\n","Epoch 33 Batch 100 Loss 0.2103\n","Epoch 33 Batch 200 Loss 0.2155\n","Epoch 33 Batch 300 Loss 0.2283\n","Epoch 33 Batch 400 Loss 0.2358\n","Epoch 33 Batch 500 Loss 0.2149\n","Epoch 33 Batch 600 Loss 0.2193\n","Epoch 33 Batch 700 Loss 0.2375\n","Epoch 33 Batch 800 Loss 0.2544\n","Epoch 33 Batch 900 Loss 0.2528\n","Epoch 33 Batch 1000 Loss 0.2557\n","Epoch 33 Loss 0.057478\n","Time taken for 1 epoch 147.08 sec\n","\n","Epoch 34 Batch 0 Loss 0.2517\n","Epoch 34 Batch 100 Loss 0.2055\n","Epoch 34 Batch 200 Loss 0.2398\n","Epoch 34 Batch 300 Loss 0.2270\n","Epoch 34 Batch 400 Loss 0.2423\n","Epoch 34 Batch 500 Loss 0.1996\n","Epoch 34 Batch 600 Loss 0.1993\n","Epoch 34 Batch 700 Loss 0.2192\n","Epoch 34 Batch 800 Loss 0.2314\n","Epoch 34 Batch 900 Loss 0.2461\n","Epoch 34 Batch 1000 Loss 0.2356\n","Epoch 34 Loss 0.056673\n","Time taken for 1 epoch 147.54 sec\n","\n","Epoch 35 Batch 0 Loss 0.2303\n","Epoch 35 Batch 100 Loss 0.2196\n","Epoch 35 Batch 200 Loss 0.2340\n","Epoch 35 Batch 300 Loss 0.2608\n","Epoch 35 Batch 400 Loss 0.2238\n","Epoch 35 Batch 500 Loss 0.2291\n","Epoch 35 Batch 600 Loss 0.2559\n","Epoch 35 Batch 700 Loss 0.2271\n","Epoch 35 Batch 800 Loss 0.2557\n","Epoch 35 Batch 900 Loss 0.2303\n","Epoch 35 Batch 1000 Loss 0.2100\n","Epoch 35 Loss 0.055901\n","Time taken for 1 epoch 146.67 sec\n","\n","Epoch 36 Batch 0 Loss 0.2274\n","Epoch 36 Batch 100 Loss 0.2127\n","Epoch 36 Batch 200 Loss 0.2164\n","Epoch 36 Batch 300 Loss 0.2334\n","Epoch 36 Batch 400 Loss 0.2372\n","Epoch 36 Batch 500 Loss 0.2396\n","Epoch 36 Batch 600 Loss 0.2108\n","Epoch 36 Batch 700 Loss 0.2335\n","Epoch 36 Batch 800 Loss 0.1980\n","Epoch 36 Batch 900 Loss 0.2494\n","Epoch 36 Batch 1000 Loss 0.2185\n","Epoch 36 Loss 0.055326\n","Time taken for 1 epoch 147.55 sec\n","\n","Epoch 37 Batch 0 Loss 0.2484\n","Epoch 37 Batch 100 Loss 0.2356\n","Epoch 37 Batch 200 Loss 0.2218\n","Epoch 37 Batch 300 Loss 0.2525\n","Epoch 37 Batch 400 Loss 0.2113\n","Epoch 37 Batch 500 Loss 0.2294\n","Epoch 37 Batch 600 Loss 0.2164\n","Epoch 37 Batch 700 Loss 0.2436\n","Epoch 37 Batch 800 Loss 0.2425\n","Epoch 37 Batch 900 Loss 0.2121\n","Epoch 37 Batch 1000 Loss 0.2024\n","Epoch 37 Loss 0.054377\n","Time taken for 1 epoch 147.30 sec\n","\n","Epoch 38 Batch 0 Loss 0.2360\n","Epoch 38 Batch 100 Loss 0.2457\n","Epoch 38 Batch 200 Loss 0.2243\n","Epoch 38 Batch 300 Loss 0.2102\n","Epoch 38 Batch 400 Loss 0.2117\n","Epoch 38 Batch 500 Loss 0.2040\n","Epoch 38 Batch 600 Loss 0.2086\n","Epoch 38 Batch 700 Loss 0.2186\n","Epoch 38 Batch 800 Loss 0.2073\n","Epoch 38 Batch 900 Loss 0.2292\n","Epoch 38 Batch 1000 Loss 0.2084\n","Epoch 38 Loss 0.053874\n","Time taken for 1 epoch 146.94 sec\n","\n","Epoch 39 Batch 0 Loss 0.2430\n","Epoch 39 Batch 100 Loss 0.2123\n","Epoch 39 Batch 200 Loss 0.2202\n","Epoch 39 Batch 300 Loss 0.1956\n","Epoch 39 Batch 400 Loss 0.1995\n","Epoch 39 Batch 500 Loss 0.2114\n","Epoch 39 Batch 600 Loss 0.1949\n","Epoch 39 Batch 700 Loss 0.2203\n","Epoch 39 Batch 800 Loss 0.2031\n","Epoch 39 Batch 900 Loss 0.2186\n","Epoch 39 Batch 1000 Loss 0.2430\n","Epoch 39 Loss 0.052969\n","Time taken for 1 epoch 147.64 sec\n","\n","Epoch 40 Batch 0 Loss 0.2277\n","Epoch 40 Batch 100 Loss 0.2028\n","Epoch 40 Batch 200 Loss 0.2005\n","Epoch 40 Batch 300 Loss 0.2178\n","Epoch 40 Batch 400 Loss 0.2429\n","Epoch 40 Batch 500 Loss 0.2306\n","Epoch 40 Batch 600 Loss 0.2002\n","Epoch 40 Batch 700 Loss 0.2195\n","Epoch 40 Batch 800 Loss 0.2041\n","Epoch 40 Batch 900 Loss 0.2089\n","Epoch 40 Batch 1000 Loss 0.2150\n","Epoch 40 Loss 0.052495\n","Time taken for 1 epoch 146.99 sec\n","\n","Epoch 41 Batch 0 Loss 0.2066\n","Epoch 41 Batch 100 Loss 0.2055\n","Epoch 41 Batch 200 Loss 0.1965\n","Epoch 41 Batch 300 Loss 0.1993\n","Epoch 41 Batch 400 Loss 0.2157\n","Epoch 41 Batch 500 Loss 0.2124\n","Epoch 41 Batch 600 Loss 0.2029\n","Epoch 41 Batch 700 Loss 0.1880\n","Epoch 41 Batch 800 Loss 0.1666\n","Epoch 41 Batch 900 Loss 0.2070\n","Epoch 41 Batch 1000 Loss 0.2119\n","Epoch 41 Loss 0.051869\n","Time taken for 1 epoch 148.05 sec\n","\n","Epoch 42 Batch 0 Loss 0.2058\n","Epoch 42 Batch 100 Loss 0.2050\n","Epoch 42 Batch 200 Loss 0.2299\n","Epoch 42 Batch 300 Loss 0.2208\n","Epoch 42 Batch 400 Loss 0.1923\n","Epoch 42 Batch 500 Loss 0.2107\n","Epoch 42 Batch 600 Loss 0.2292\n","Epoch 42 Batch 700 Loss 0.2057\n","Epoch 42 Batch 800 Loss 0.2262\n","Epoch 42 Batch 900 Loss 0.2447\n","Epoch 42 Batch 1000 Loss 0.2161\n","Epoch 42 Loss 0.051327\n","Time taken for 1 epoch 146.56 sec\n","\n","Epoch 43 Batch 0 Loss 0.2136\n","Epoch 43 Batch 100 Loss 0.2097\n","Epoch 43 Batch 200 Loss 0.2016\n","Epoch 43 Batch 300 Loss 0.2011\n","Epoch 43 Batch 400 Loss 0.2220\n","Epoch 43 Batch 500 Loss 0.1961\n","Epoch 43 Batch 600 Loss 0.2247\n","Epoch 43 Batch 700 Loss 0.2070\n","Epoch 43 Batch 800 Loss 0.2002\n","Epoch 43 Batch 900 Loss 0.2034\n","Epoch 43 Batch 1000 Loss 0.2198\n","Epoch 43 Loss 0.050628\n","Time taken for 1 epoch 147.86 sec\n","\n","Epoch 44 Batch 0 Loss 0.2089\n","Epoch 44 Batch 100 Loss 0.1744\n","Epoch 44 Batch 200 Loss 0.2124\n","Epoch 44 Batch 300 Loss 0.2049\n","Epoch 44 Batch 400 Loss 0.1798\n","Epoch 44 Batch 500 Loss 0.1690\n","Epoch 44 Batch 600 Loss 0.1900\n","Epoch 44 Batch 700 Loss 0.1909\n","Epoch 44 Batch 800 Loss 0.1879\n","Epoch 44 Batch 900 Loss 0.2044\n","Epoch 44 Batch 1000 Loss 0.2068\n","Epoch 44 Loss 0.050399\n","Time taken for 1 epoch 147.60 sec\n","\n","Epoch 45 Batch 0 Loss 0.2194\n","Epoch 45 Batch 100 Loss 0.1881\n","Epoch 45 Batch 200 Loss 0.2003\n","Epoch 45 Batch 300 Loss 0.1948\n","Epoch 45 Batch 400 Loss 0.2075\n","Epoch 45 Batch 500 Loss 0.2161\n","Epoch 45 Batch 600 Loss 0.2120\n","Epoch 45 Batch 700 Loss 0.1777\n","Epoch 45 Batch 800 Loss 0.2028\n","Epoch 45 Batch 900 Loss 0.2047\n","Epoch 45 Batch 1000 Loss 0.2026\n","Epoch 45 Loss 0.049721\n","Time taken for 1 epoch 147.39 sec\n","\n","Epoch 46 Batch 0 Loss 0.2656\n","Epoch 46 Batch 100 Loss 0.1886\n","Epoch 46 Batch 200 Loss 0.2000\n","Epoch 46 Batch 300 Loss 0.1960\n","Epoch 46 Batch 400 Loss 0.1962\n","Epoch 46 Batch 500 Loss 0.1915\n","Epoch 46 Batch 600 Loss 0.2133\n","Epoch 46 Batch 700 Loss 0.1816\n","Epoch 46 Batch 800 Loss 0.2003\n","Epoch 46 Batch 900 Loss 0.2158\n","Epoch 46 Batch 1000 Loss 0.1834\n","Epoch 46 Loss 0.049241\n","Time taken for 1 epoch 148.11 sec\n","\n","Epoch 47 Batch 0 Loss 0.2303\n","Epoch 47 Batch 100 Loss 0.2043\n","Epoch 47 Batch 200 Loss 0.1946\n","Epoch 47 Batch 300 Loss 0.2098\n","Epoch 47 Batch 400 Loss 0.2127\n","Epoch 47 Batch 500 Loss 0.2085\n","Epoch 47 Batch 600 Loss 0.1762\n","Epoch 47 Batch 700 Loss 0.1988\n","Epoch 47 Batch 800 Loss 0.1696\n","Epoch 47 Batch 900 Loss 0.1863\n","Epoch 47 Batch 1000 Loss 0.2024\n","Epoch 47 Loss 0.048754\n","Time taken for 1 epoch 147.49 sec\n","\n","Epoch 48 Batch 0 Loss 0.2051\n","Epoch 48 Batch 100 Loss 0.1994\n","Epoch 48 Batch 200 Loss 0.2094\n","Epoch 48 Batch 300 Loss 0.2057\n","Epoch 48 Batch 400 Loss 0.1957\n","Epoch 48 Batch 500 Loss 0.1982\n","Epoch 48 Batch 600 Loss 0.1940\n","Epoch 48 Batch 700 Loss 0.1858\n","Epoch 48 Batch 800 Loss 0.1928\n","Epoch 48 Batch 900 Loss 0.1823\n","Epoch 48 Batch 1000 Loss 0.1850\n","Epoch 48 Loss 0.048369\n","Time taken for 1 epoch 147.17 sec\n","\n","Epoch 49 Batch 0 Loss 0.2108\n","Epoch 49 Batch 100 Loss 0.1575\n","Epoch 49 Batch 200 Loss 0.1822\n","Epoch 49 Batch 300 Loss 0.1958\n","Epoch 49 Batch 400 Loss 0.1797\n","Epoch 49 Batch 500 Loss 0.2061\n","Epoch 49 Batch 600 Loss 0.1957\n","Epoch 49 Batch 700 Loss 0.1954\n","Epoch 49 Batch 800 Loss 0.1914\n","Epoch 49 Batch 900 Loss 0.1957\n","Epoch 49 Batch 1000 Loss 0.1918\n","Epoch 49 Loss 0.048040\n","Time taken for 1 epoch 147.63 sec\n","\n","Epoch 50 Batch 0 Loss 0.2251\n","Epoch 50 Batch 100 Loss 0.1875\n","Epoch 50 Batch 200 Loss 0.1822\n","Epoch 50 Batch 300 Loss 0.1865\n","Epoch 50 Batch 400 Loss 0.1941\n","Epoch 50 Batch 500 Loss 0.2008\n","Epoch 50 Batch 600 Loss 0.2205\n","Epoch 50 Batch 700 Loss 0.1882\n","Epoch 50 Batch 800 Loss 0.1800\n","Epoch 50 Batch 900 Loss 0.1974\n","Epoch 50 Batch 1000 Loss 0.1622\n","Epoch 50 Loss 0.047523\n","Time taken for 1 epoch 147.07 sec\n","\n","Epoch 51 Batch 0 Loss 0.2117\n","Epoch 51 Batch 100 Loss 0.1769\n","Epoch 51 Batch 200 Loss 0.1647\n","Epoch 51 Batch 300 Loss 0.1857\n","Epoch 51 Batch 400 Loss 0.2052\n","Epoch 51 Batch 500 Loss 0.1802\n","Epoch 51 Batch 600 Loss 0.1965\n","Epoch 51 Batch 700 Loss 0.2150\n","Epoch 51 Batch 800 Loss 0.1979\n","Epoch 51 Batch 900 Loss 0.2085\n","Epoch 51 Batch 1000 Loss 0.1862\n","Epoch 51 Loss 0.047089\n","Time taken for 1 epoch 147.65 sec\n","\n","Epoch 52 Batch 0 Loss 0.2116\n","Epoch 52 Batch 100 Loss 0.1778\n","Epoch 52 Batch 200 Loss 0.1822\n","Epoch 52 Batch 300 Loss 0.2028\n","Epoch 52 Batch 400 Loss 0.1982\n","Epoch 52 Batch 500 Loss 0.1821\n","Epoch 52 Batch 600 Loss 0.1964\n","Epoch 52 Batch 700 Loss 0.1966\n","Epoch 52 Batch 800 Loss 0.2017\n","Epoch 52 Batch 900 Loss 0.1826\n","Epoch 52 Batch 1000 Loss 0.1752\n","Epoch 52 Loss 0.046610\n","Time taken for 1 epoch 147.77 sec\n","\n","Epoch 53 Batch 0 Loss 0.1844\n","Epoch 53 Batch 100 Loss 0.2400\n","Epoch 53 Batch 200 Loss 0.1954\n","Epoch 53 Batch 300 Loss 0.1582\n","Epoch 53 Batch 400 Loss 0.1659\n","Epoch 53 Batch 500 Loss 0.1845\n","Epoch 53 Batch 600 Loss 0.1983\n","Epoch 53 Batch 700 Loss 0.2188\n","Epoch 53 Batch 800 Loss 0.1615\n","Epoch 53 Batch 900 Loss 0.2241\n","Epoch 53 Batch 1000 Loss 0.2090\n","Epoch 53 Loss 0.046385\n","Time taken for 1 epoch 148.13 sec\n","\n","Epoch 54 Batch 0 Loss 0.1844\n","Epoch 54 Batch 100 Loss 0.1599\n","Epoch 54 Batch 200 Loss 0.1784\n","Epoch 54 Batch 300 Loss 0.1887\n","Epoch 54 Batch 400 Loss 0.1706\n","Epoch 54 Batch 500 Loss 0.2021\n","Epoch 54 Batch 600 Loss 0.2069\n","Epoch 54 Batch 700 Loss 0.2063\n","Epoch 54 Batch 800 Loss 0.1917\n","Epoch 54 Batch 900 Loss 0.2014\n","Epoch 54 Batch 1000 Loss 0.1924\n","Epoch 54 Loss 0.046004\n","Time taken for 1 epoch 147.83 sec\n","\n","Epoch 55 Batch 0 Loss 0.1956\n","Epoch 55 Batch 100 Loss 0.1760\n","Epoch 55 Batch 200 Loss 0.2008\n","Epoch 55 Batch 300 Loss 0.2006\n","Epoch 55 Batch 400 Loss 0.1833\n","Epoch 55 Batch 500 Loss 0.1699\n","Epoch 55 Batch 600 Loss 0.1882\n","Epoch 55 Batch 700 Loss 0.1763\n","Epoch 55 Batch 800 Loss 0.1932\n","Epoch 55 Batch 900 Loss 0.2028\n","Epoch 55 Batch 1000 Loss 0.1782\n","Epoch 55 Loss 0.045805\n","Time taken for 1 epoch 147.25 sec\n","\n","Epoch 56 Batch 0 Loss 0.1808\n","Epoch 56 Batch 100 Loss 0.1904\n","Epoch 56 Batch 200 Loss 0.1560\n","Epoch 56 Batch 300 Loss 0.1922\n","Epoch 56 Batch 400 Loss 0.1855\n","Epoch 56 Batch 500 Loss 0.1971\n","Epoch 56 Batch 600 Loss 0.1771\n","Epoch 56 Batch 700 Loss 0.1971\n","Epoch 56 Batch 800 Loss 0.1799\n","Epoch 56 Batch 900 Loss 0.1891\n","Epoch 56 Batch 1000 Loss 0.1891\n","Epoch 56 Loss 0.045393\n","Time taken for 1 epoch 147.14 sec\n","\n","Epoch 57 Batch 0 Loss 0.1656\n","Epoch 57 Batch 100 Loss 0.1993\n","Epoch 57 Batch 200 Loss 0.1709\n","Epoch 57 Batch 300 Loss 0.2013\n","Epoch 57 Batch 400 Loss 0.1565\n","Epoch 57 Batch 500 Loss 0.1735\n","Epoch 57 Batch 600 Loss 0.1720\n","Epoch 57 Batch 700 Loss 0.1928\n","Epoch 57 Batch 800 Loss 0.1960\n","Epoch 57 Batch 900 Loss 0.1785\n","Epoch 57 Batch 1000 Loss 0.1736\n","Epoch 57 Loss 0.044855\n","Time taken for 1 epoch 147.31 sec\n","\n","Epoch 58 Batch 0 Loss 0.1877\n","Epoch 58 Batch 100 Loss 0.1480\n","Epoch 58 Batch 200 Loss 0.1602\n","Epoch 58 Batch 300 Loss 0.1978\n","Epoch 58 Batch 400 Loss 0.1681\n","Epoch 58 Batch 500 Loss 0.1747\n","Epoch 58 Batch 600 Loss 0.1697\n","Epoch 58 Batch 700 Loss 0.1785\n","Epoch 58 Batch 800 Loss 0.1797\n","Epoch 58 Batch 900 Loss 0.1656\n","Epoch 58 Batch 1000 Loss 0.1669\n","Epoch 58 Loss 0.044566\n","Time taken for 1 epoch 147.52 sec\n","\n","Epoch 59 Batch 0 Loss 0.1757\n","Epoch 59 Batch 100 Loss 0.1712\n","Epoch 59 Batch 200 Loss 0.1722\n","Epoch 59 Batch 300 Loss 0.1868\n","Epoch 59 Batch 400 Loss 0.1745\n","Epoch 59 Batch 500 Loss 0.1742\n","Epoch 59 Batch 600 Loss 0.1875\n","Epoch 59 Batch 700 Loss 0.2023\n","Epoch 59 Batch 800 Loss 0.1789\n","Epoch 59 Batch 900 Loss 0.1788\n","Epoch 59 Batch 1000 Loss 0.1799\n","Epoch 59 Loss 0.044486\n","Time taken for 1 epoch 147.05 sec\n","\n","Epoch 60 Batch 0 Loss 0.1867\n","Epoch 60 Batch 100 Loss 0.1801\n","Epoch 60 Batch 200 Loss 0.1807\n","Epoch 60 Batch 300 Loss 0.1884\n","Epoch 60 Batch 400 Loss 0.1939\n","Epoch 60 Batch 500 Loss 0.1682\n","Epoch 60 Batch 600 Loss 0.1581\n","Epoch 60 Batch 700 Loss 0.2006\n","Epoch 60 Batch 800 Loss 0.1490\n","Epoch 60 Batch 900 Loss 0.1804\n","Epoch 60 Batch 1000 Loss 0.1724\n","Epoch 60 Loss 0.044132\n","Time taken for 1 epoch 147.43 sec\n","\n","Epoch 61 Batch 0 Loss 0.2226\n","Epoch 61 Batch 100 Loss 0.1489\n","Epoch 61 Batch 200 Loss 0.1660\n","Epoch 61 Batch 300 Loss 0.1735\n","Epoch 61 Batch 400 Loss 0.1807\n","Epoch 61 Batch 500 Loss 0.1865\n","Epoch 61 Batch 600 Loss 0.1800\n","Epoch 61 Batch 700 Loss 0.1965\n","Epoch 61 Batch 800 Loss 0.1778\n","Epoch 61 Batch 900 Loss 0.1992\n","Epoch 61 Batch 1000 Loss 0.1904\n","Epoch 61 Loss 0.044023\n","Time taken for 1 epoch 147.57 sec\n","\n","Epoch 62 Batch 0 Loss 0.1940\n","Epoch 62 Batch 100 Loss 0.1726\n","Epoch 62 Batch 200 Loss 0.1859\n","Epoch 62 Batch 300 Loss 0.1763\n","Epoch 62 Batch 400 Loss 0.1737\n","Epoch 62 Batch 500 Loss 0.1900\n","Epoch 62 Batch 600 Loss 0.1630\n","Epoch 62 Batch 700 Loss 0.1856\n","Epoch 62 Batch 800 Loss 0.1666\n","Epoch 62 Batch 900 Loss 0.1732\n","Epoch 62 Batch 1000 Loss 0.1877\n","Epoch 62 Loss 0.044138\n","Time taken for 1 epoch 147.80 sec\n","\n","Epoch 63 Batch 0 Loss 0.1775\n","Epoch 63 Batch 100 Loss 0.1579\n","Epoch 63 Batch 200 Loss 0.1759\n","Epoch 63 Batch 300 Loss 0.1455\n","Epoch 63 Batch 400 Loss 0.1851\n","Epoch 63 Batch 500 Loss 0.2034\n","Epoch 63 Batch 600 Loss 0.1612\n","Epoch 63 Batch 700 Loss 0.1693\n","Epoch 63 Batch 800 Loss 0.1833\n","Epoch 63 Batch 900 Loss 0.1753\n","Epoch 63 Batch 1000 Loss 0.1695\n","Epoch 63 Loss 0.043099\n","Time taken for 1 epoch 147.16 sec\n","\n","Epoch 64 Batch 0 Loss 0.1940\n","Epoch 64 Batch 100 Loss 0.1678\n","Epoch 64 Batch 200 Loss 0.1597\n","Epoch 64 Batch 300 Loss 0.1900\n","Epoch 64 Batch 400 Loss 0.1504\n","Epoch 64 Batch 500 Loss 0.1654\n","Epoch 64 Batch 600 Loss 0.1655\n","Epoch 64 Batch 700 Loss 0.1649\n","Epoch 64 Batch 800 Loss 0.1965\n","Epoch 64 Batch 900 Loss 0.1735\n","Epoch 64 Batch 1000 Loss 0.1737\n","Epoch 64 Loss 0.042740\n","Time taken for 1 epoch 146.94 sec\n","\n","Epoch 65 Batch 0 Loss 0.2076\n","Epoch 65 Batch 100 Loss 0.1770\n","Epoch 65 Batch 200 Loss 0.1686\n","Epoch 65 Batch 300 Loss 0.1766\n","Epoch 65 Batch 400 Loss 0.1771\n","Epoch 65 Batch 500 Loss 0.1565\n","Epoch 65 Batch 600 Loss 0.1411\n","Epoch 65 Batch 700 Loss 0.1784\n","Epoch 65 Batch 800 Loss 0.1598\n","Epoch 65 Batch 900 Loss 0.1742\n","Epoch 65 Batch 1000 Loss 0.1743\n","Epoch 65 Loss 0.042697\n","Time taken for 1 epoch 147.54 sec\n","\n","Epoch 66 Batch 0 Loss 0.1699\n","Epoch 66 Batch 100 Loss 0.1691\n","Epoch 66 Batch 200 Loss 0.1762\n","Epoch 66 Batch 300 Loss 0.1771\n","Epoch 66 Batch 400 Loss 0.1919\n","Epoch 66 Batch 500 Loss 0.1635\n","Epoch 66 Batch 600 Loss 0.1399\n","Epoch 66 Batch 700 Loss 0.1735\n","Epoch 66 Batch 800 Loss 0.1661\n","Epoch 66 Batch 900 Loss 0.1902\n","Epoch 66 Batch 1000 Loss 0.1717\n","Epoch 66 Loss 0.042544\n","Time taken for 1 epoch 147.96 sec\n","\n","Epoch 67 Batch 0 Loss 0.1861\n","Epoch 67 Batch 100 Loss 0.1602\n","Epoch 67 Batch 200 Loss 0.1855\n","Epoch 67 Batch 300 Loss 0.1808\n","Epoch 67 Batch 400 Loss 0.1689\n","Epoch 67 Batch 500 Loss 0.1409\n","Epoch 67 Batch 600 Loss 0.1764\n","Epoch 67 Batch 700 Loss 0.1671\n","Epoch 67 Batch 800 Loss 0.1565\n","Epoch 67 Batch 900 Loss 0.1750\n","Epoch 67 Batch 1000 Loss 0.1629\n","Epoch 67 Loss 0.042212\n","Time taken for 1 epoch 146.80 sec\n","\n","Epoch 68 Batch 0 Loss 0.2208\n","Epoch 68 Batch 100 Loss 0.1515\n","Epoch 68 Batch 200 Loss 0.1873\n","Epoch 68 Batch 300 Loss 0.1572\n","Epoch 68 Batch 400 Loss 0.1785\n","Epoch 68 Batch 500 Loss 0.1604\n","Epoch 68 Batch 600 Loss 0.1346\n","Epoch 68 Batch 700 Loss 0.1880\n","Epoch 68 Batch 800 Loss 0.1548\n","Epoch 68 Batch 900 Loss 0.1652\n","Epoch 68 Batch 1000 Loss 0.1650\n","Epoch 68 Loss 0.042043\n","Time taken for 1 epoch 147.28 sec\n","\n","Epoch 69 Batch 0 Loss 0.1766\n","Epoch 69 Batch 100 Loss 0.1578\n","Epoch 69 Batch 200 Loss 0.1623\n","Epoch 69 Batch 300 Loss 0.1637\n","Epoch 69 Batch 400 Loss 0.1734\n","Epoch 69 Batch 500 Loss 0.1576\n","Epoch 69 Batch 600 Loss 0.2109\n","Epoch 69 Batch 700 Loss 0.1580\n","Epoch 69 Batch 800 Loss 0.1669\n","Epoch 69 Batch 900 Loss 0.1704\n","Epoch 69 Batch 1000 Loss 0.1597\n","Epoch 69 Loss 0.041900\n","Time taken for 1 epoch 146.72 sec\n","\n","Epoch 70 Batch 0 Loss 0.1710\n","Epoch 70 Batch 100 Loss 0.1768\n","Epoch 70 Batch 200 Loss 0.1497\n","Epoch 70 Batch 300 Loss 0.1565\n","Epoch 70 Batch 400 Loss 0.1770\n","Epoch 70 Batch 500 Loss 0.1684\n","Epoch 70 Batch 600 Loss 0.1638\n","Epoch 70 Batch 700 Loss 0.1837\n","Epoch 70 Batch 800 Loss 0.1600\n","Epoch 70 Batch 900 Loss 0.1565\n","Epoch 70 Batch 1000 Loss 0.1586\n","Epoch 70 Loss 0.041435\n","Time taken for 1 epoch 147.84 sec\n","\n","Epoch 71 Batch 0 Loss 0.1916\n","Epoch 71 Batch 100 Loss 0.1494\n","Epoch 71 Batch 200 Loss 0.1462\n","Epoch 71 Batch 300 Loss 0.1833\n","Epoch 71 Batch 400 Loss 0.1586\n","Epoch 71 Batch 500 Loss 0.1552\n","Epoch 71 Batch 600 Loss 0.1666\n","Epoch 71 Batch 700 Loss 0.1845\n","Epoch 71 Batch 800 Loss 0.1404\n","Epoch 71 Batch 900 Loss 0.1678\n","Epoch 71 Batch 1000 Loss 0.1747\n","Epoch 71 Loss 0.041376\n","Time taken for 1 epoch 147.70 sec\n","\n","Epoch 72 Batch 0 Loss 0.1688\n","Epoch 72 Batch 100 Loss 0.1767\n","Epoch 72 Batch 200 Loss 0.1593\n","Epoch 72 Batch 300 Loss 0.1644\n","Epoch 72 Batch 400 Loss 0.1391\n","Epoch 72 Batch 500 Loss 0.1764\n","Epoch 72 Batch 600 Loss 0.1632\n","Epoch 72 Batch 700 Loss 0.1696\n","Epoch 72 Batch 800 Loss 0.1728\n","Epoch 72 Batch 900 Loss 0.1576\n","Epoch 72 Batch 1000 Loss 0.1838\n","Epoch 72 Loss 0.041097\n","Time taken for 1 epoch 147.86 sec\n","\n","Epoch 73 Batch 0 Loss 0.1962\n","Epoch 73 Batch 100 Loss 0.2106\n","Epoch 73 Batch 200 Loss 0.1899\n","Epoch 73 Batch 300 Loss 0.1569\n","Epoch 73 Batch 400 Loss 0.1790\n","Epoch 73 Batch 500 Loss 0.1739\n","Epoch 73 Batch 600 Loss 0.1672\n","Epoch 73 Batch 700 Loss 0.1652\n","Epoch 73 Batch 800 Loss 0.1753\n","Epoch 73 Batch 900 Loss 0.1800\n","Epoch 73 Batch 1000 Loss 0.1509\n","Epoch 73 Loss 0.040984\n","Time taken for 1 epoch 147.68 sec\n","\n","Epoch 74 Batch 0 Loss 0.1762\n","Epoch 74 Batch 100 Loss 0.1544\n","Epoch 74 Batch 200 Loss 0.1788\n","Epoch 74 Batch 300 Loss 0.1689\n","Epoch 74 Batch 400 Loss 0.1539\n","Epoch 74 Batch 500 Loss 0.1408\n","Epoch 74 Batch 600 Loss 0.1705\n","Epoch 74 Batch 700 Loss 0.1858\n","Epoch 74 Batch 800 Loss 0.1606\n","Epoch 74 Batch 900 Loss 0.1559\n","Epoch 74 Batch 1000 Loss 0.1668\n","Epoch 74 Loss 0.040812\n","Time taken for 1 epoch 147.75 sec\n","\n","Epoch 75 Batch 0 Loss 0.1944\n","Epoch 75 Batch 100 Loss 0.1596\n","Epoch 75 Batch 200 Loss 0.1428\n","Epoch 75 Batch 300 Loss 0.1720\n","Epoch 75 Batch 400 Loss 0.1747\n","Epoch 75 Batch 500 Loss 0.1635\n","Epoch 75 Batch 600 Loss 0.1810\n","Epoch 75 Batch 700 Loss 0.1890\n","Epoch 75 Batch 800 Loss 0.1721\n","Epoch 75 Batch 900 Loss 0.1744\n","Epoch 75 Batch 1000 Loss 0.1575\n","Epoch 75 Loss 0.040523\n","Time taken for 1 epoch 146.66 sec\n","\n","Epoch 76 Batch 0 Loss 0.1790\n","Epoch 76 Batch 100 Loss 0.1683\n","Epoch 76 Batch 200 Loss 0.1691\n","Epoch 76 Batch 300 Loss 0.1432\n","Epoch 76 Batch 400 Loss 0.1832\n","Epoch 76 Batch 500 Loss 0.1452\n","Epoch 76 Batch 600 Loss 0.1773\n","Epoch 76 Batch 700 Loss 0.1714\n","Epoch 76 Batch 800 Loss 0.1652\n","Epoch 76 Batch 900 Loss 0.1429\n","Epoch 76 Batch 1000 Loss 0.1499\n","Epoch 76 Loss 0.040397\n","Time taken for 1 epoch 147.26 sec\n","\n","Epoch 77 Batch 0 Loss 0.1601\n","Epoch 77 Batch 100 Loss 0.1641\n","Epoch 77 Batch 200 Loss 0.1490\n","Epoch 77 Batch 300 Loss 0.1700\n","Epoch 77 Batch 400 Loss 0.1621\n","Epoch 77 Batch 500 Loss 0.1575\n","Epoch 77 Batch 600 Loss 0.1744\n","Epoch 77 Batch 700 Loss 0.1596\n","Epoch 77 Batch 800 Loss 0.1666\n","Epoch 77 Batch 900 Loss 0.1651\n","Epoch 77 Batch 1000 Loss 0.1362\n","Epoch 77 Loss 0.040444\n","Time taken for 1 epoch 147.46 sec\n","\n","Epoch 78 Batch 0 Loss 0.1610\n","Epoch 78 Batch 100 Loss 0.1471\n","Epoch 78 Batch 200 Loss 0.1721\n","Epoch 78 Batch 300 Loss 0.1546\n","Epoch 78 Batch 400 Loss 0.1602\n","Epoch 78 Batch 500 Loss 0.1588\n","Epoch 78 Batch 600 Loss 0.1693\n","Epoch 78 Batch 700 Loss 0.1757\n","Epoch 78 Batch 800 Loss 0.1515\n","Epoch 78 Batch 900 Loss 0.1632\n","Epoch 78 Batch 1000 Loss 0.1675\n","Epoch 78 Loss 0.039975\n","Time taken for 1 epoch 147.25 sec\n","\n","Epoch 79 Batch 0 Loss 0.1861\n","Epoch 79 Batch 100 Loss 0.1393\n","Epoch 79 Batch 200 Loss 0.1479\n","Epoch 79 Batch 300 Loss 0.1447\n","Epoch 79 Batch 400 Loss 0.1428\n","Epoch 79 Batch 500 Loss 0.1451\n","Epoch 79 Batch 600 Loss 0.1528\n","Epoch 79 Batch 700 Loss 0.1638\n","Epoch 79 Batch 800 Loss 0.1548\n","Epoch 79 Batch 900 Loss 0.1804\n","Epoch 79 Batch 1000 Loss 0.1654\n","Epoch 79 Loss 0.039485\n","Time taken for 1 epoch 147.32 sec\n","\n","Epoch 80 Batch 0 Loss 0.1730\n","Epoch 80 Batch 100 Loss 0.1599\n","Epoch 80 Batch 200 Loss 0.1514\n","Epoch 80 Batch 300 Loss 0.1565\n","Epoch 80 Batch 400 Loss 0.1532\n","Epoch 80 Batch 500 Loss 0.1702\n","Epoch 80 Batch 600 Loss 0.1704\n","Epoch 80 Batch 700 Loss 0.1565\n","Epoch 80 Batch 800 Loss 0.1587\n","Epoch 80 Batch 900 Loss 0.1425\n","Epoch 80 Batch 1000 Loss 0.1628\n","Epoch 80 Loss 0.039402\n","Time taken for 1 epoch 147.12 sec\n","\n","Epoch 81 Batch 0 Loss 0.1628\n","Epoch 81 Batch 100 Loss 0.1598\n","Epoch 81 Batch 200 Loss 0.1718\n","Epoch 81 Batch 300 Loss 0.1427\n","Epoch 81 Batch 400 Loss 0.1623\n","Epoch 81 Batch 500 Loss 0.1768\n","Epoch 81 Batch 600 Loss 0.1820\n","Epoch 81 Batch 700 Loss 0.1665\n","Epoch 81 Batch 800 Loss 0.1544\n","Epoch 81 Batch 900 Loss 0.1339\n","Epoch 81 Batch 1000 Loss 0.1762\n","Epoch 81 Loss 0.039423\n","Time taken for 1 epoch 147.91 sec\n","\n","Epoch 82 Batch 0 Loss 0.1772\n","Epoch 82 Batch 100 Loss 0.1525\n","Epoch 82 Batch 200 Loss 0.1800\n","Epoch 82 Batch 300 Loss 0.1710\n","Epoch 82 Batch 400 Loss 0.1416\n","Epoch 82 Batch 500 Loss 0.1525\n","Epoch 82 Batch 600 Loss 0.1467\n","Epoch 82 Batch 700 Loss 0.1562\n","Epoch 82 Batch 800 Loss 0.1484\n","Epoch 82 Batch 900 Loss 0.1660\n","Epoch 82 Batch 1000 Loss 0.1499\n","Epoch 82 Loss 0.039372\n","Time taken for 1 epoch 147.79 sec\n","\n","Epoch 83 Batch 0 Loss 0.1905\n","Epoch 83 Batch 100 Loss 0.1606\n","Epoch 83 Batch 200 Loss 0.1650\n","Epoch 83 Batch 300 Loss 0.1514\n","Epoch 83 Batch 400 Loss 0.1511\n","Epoch 83 Batch 500 Loss 0.1568\n","Epoch 83 Batch 600 Loss 0.1786\n","Epoch 83 Batch 700 Loss 0.1628\n","Epoch 83 Batch 800 Loss 0.1483\n","Epoch 83 Batch 900 Loss 0.1880\n","Epoch 83 Batch 1000 Loss 0.1432\n","Epoch 83 Loss 0.039155\n","Time taken for 1 epoch 146.87 sec\n","\n","Epoch 84 Batch 0 Loss 0.1628\n","Epoch 84 Batch 100 Loss 0.1419\n","Epoch 84 Batch 200 Loss 0.1538\n","Epoch 84 Batch 300 Loss 0.1435\n","Epoch 84 Batch 400 Loss 0.1624\n","Epoch 84 Batch 500 Loss 0.1574\n","Epoch 84 Batch 600 Loss 0.1578\n","Epoch 84 Batch 700 Loss 0.1418\n","Epoch 84 Batch 800 Loss 0.1556\n","Epoch 84 Batch 900 Loss 0.1543\n","Epoch 84 Batch 1000 Loss 0.1649\n","Epoch 84 Loss 0.038826\n","Time taken for 1 epoch 147.77 sec\n","\n","Epoch 85 Batch 0 Loss 0.1847\n","Epoch 85 Batch 100 Loss 0.1405\n","Epoch 85 Batch 200 Loss 0.1475\n","Epoch 85 Batch 300 Loss 0.1644\n","Epoch 85 Batch 400 Loss 0.1480\n","Epoch 85 Batch 500 Loss 0.1587\n","Epoch 85 Batch 600 Loss 0.1550\n","Epoch 85 Batch 700 Loss 0.1540\n","Epoch 85 Batch 800 Loss 0.1865\n","Epoch 85 Batch 900 Loss 0.1691\n","Epoch 85 Batch 1000 Loss 0.1676\n","Epoch 85 Loss 0.038769\n","Time taken for 1 epoch 147.35 sec\n","\n","Epoch 86 Batch 0 Loss 0.1543\n","Epoch 86 Batch 100 Loss 0.1469\n","Epoch 86 Batch 200 Loss 0.1491\n","Epoch 86 Batch 300 Loss 0.1675\n","Epoch 86 Batch 400 Loss 0.1408\n","Epoch 86 Batch 500 Loss 0.1428\n","Epoch 86 Batch 600 Loss 0.1644\n","Epoch 86 Batch 700 Loss 0.1539\n","Epoch 86 Batch 800 Loss 0.1452\n","Epoch 86 Batch 900 Loss 0.1537\n","Epoch 86 Batch 1000 Loss 0.1461\n","Epoch 86 Loss 0.038436\n","Time taken for 1 epoch 148.35 sec\n","\n","Epoch 87 Batch 0 Loss 0.1876\n","Epoch 87 Batch 100 Loss 0.1430\n","Epoch 87 Batch 200 Loss 0.1515\n","Epoch 87 Batch 300 Loss 0.1585\n","Epoch 87 Batch 400 Loss 0.1524\n","Epoch 87 Batch 500 Loss 0.1488\n","Epoch 87 Batch 600 Loss 0.1364\n","Epoch 87 Batch 700 Loss 0.1731\n","Epoch 87 Batch 800 Loss 0.1443\n","Epoch 87 Batch 900 Loss 0.1590\n","Epoch 87 Batch 1000 Loss 0.1701\n","Epoch 87 Loss 0.038307\n","Time taken for 1 epoch 147.02 sec\n","\n","Epoch 88 Batch 0 Loss 0.1666\n","Epoch 88 Batch 100 Loss 0.1515\n","Epoch 88 Batch 200 Loss 0.1585\n","Epoch 88 Batch 300 Loss 0.1870\n","Epoch 88 Batch 400 Loss 0.1577\n","Epoch 88 Batch 500 Loss 0.1631\n","Epoch 88 Batch 600 Loss 0.1732\n","Epoch 88 Batch 700 Loss 0.1553\n","Epoch 88 Batch 800 Loss 0.1660\n","Epoch 88 Batch 900 Loss 0.1333\n","Epoch 88 Batch 1000 Loss 0.1529\n","Epoch 88 Loss 0.038426\n","Time taken for 1 epoch 147.77 sec\n","\n","Epoch 89 Batch 0 Loss 0.1699\n","Epoch 89 Batch 100 Loss 0.1390\n","Epoch 89 Batch 200 Loss 0.1391\n","Epoch 89 Batch 300 Loss 0.1522\n","Epoch 89 Batch 400 Loss 0.1514\n","Epoch 89 Batch 500 Loss 0.1414\n","Epoch 89 Batch 600 Loss 0.1748\n","Epoch 89 Batch 700 Loss 0.1454\n","Epoch 89 Batch 800 Loss 0.1581\n","Epoch 89 Batch 900 Loss 0.1544\n","Epoch 89 Batch 1000 Loss 0.1437\n","Epoch 89 Loss 0.037981\n","Time taken for 1 epoch 147.49 sec\n","\n","Epoch 90 Batch 0 Loss 0.1589\n","Epoch 90 Batch 100 Loss 0.1520\n","Epoch 90 Batch 200 Loss 0.1386\n","Epoch 90 Batch 300 Loss 0.1331\n","Epoch 90 Batch 400 Loss 0.1309\n","Epoch 90 Batch 500 Loss 0.1535\n","Epoch 90 Batch 600 Loss 0.1582\n","Epoch 90 Batch 700 Loss 0.1458\n","Epoch 90 Batch 800 Loss 0.1528\n","Epoch 90 Batch 900 Loss 0.1530\n","Epoch 90 Batch 1000 Loss 0.1691\n","Epoch 90 Loss 0.038113\n","Time taken for 1 epoch 147.74 sec\n","\n","Epoch 91 Batch 0 Loss 0.1808\n","Epoch 91 Batch 100 Loss 0.1580\n","Epoch 91 Batch 200 Loss 0.1542\n","Epoch 91 Batch 300 Loss 0.1542\n","Epoch 91 Batch 400 Loss 0.1460\n","Epoch 91 Batch 500 Loss 0.1418\n","Epoch 91 Batch 600 Loss 0.1481\n","Epoch 91 Batch 700 Loss 0.1550\n","Epoch 91 Batch 800 Loss 0.1744\n","Epoch 91 Batch 900 Loss 0.1594\n","Epoch 91 Batch 1000 Loss 0.1575\n","Epoch 91 Loss 0.037948\n","Time taken for 1 epoch 148.46 sec\n","\n","Epoch 92 Batch 0 Loss 0.1544\n","Epoch 92 Batch 100 Loss 0.1566\n","Epoch 92 Batch 200 Loss 0.1432\n","Epoch 92 Batch 300 Loss 0.1662\n","Epoch 92 Batch 400 Loss 0.1501\n","Epoch 92 Batch 500 Loss 0.1482\n","Epoch 92 Batch 600 Loss 0.1625\n","Epoch 92 Batch 700 Loss 0.1551\n","Epoch 92 Batch 800 Loss 0.1454\n","Epoch 92 Batch 900 Loss 0.1695\n","Epoch 92 Batch 1000 Loss 0.1392\n","Epoch 92 Loss 0.037753\n","Time taken for 1 epoch 147.05 sec\n","\n","Epoch 93 Batch 0 Loss 0.1723\n","Epoch 93 Batch 100 Loss 0.1517\n","Epoch 93 Batch 200 Loss 0.1562\n","Epoch 93 Batch 300 Loss 0.1599\n","Epoch 93 Batch 400 Loss 0.1459\n","Epoch 93 Batch 500 Loss 0.1717\n","Epoch 93 Batch 600 Loss 0.1602\n","Epoch 93 Batch 700 Loss 0.1523\n","Epoch 93 Batch 800 Loss 0.1560\n","Epoch 93 Batch 900 Loss 0.1532\n","Epoch 93 Batch 1000 Loss 0.1666\n","Epoch 93 Loss 0.037418\n","Time taken for 1 epoch 146.94 sec\n","\n","Epoch 94 Batch 0 Loss 0.1352\n","Epoch 94 Batch 100 Loss 0.1621\n","Epoch 94 Batch 200 Loss 0.1424\n","Epoch 94 Batch 300 Loss 0.1403\n","Epoch 94 Batch 400 Loss 0.1499\n","Epoch 94 Batch 500 Loss 0.1538\n","Epoch 94 Batch 600 Loss 0.1619\n","Epoch 94 Batch 700 Loss 0.1640\n","Epoch 94 Batch 800 Loss 0.1515\n","Epoch 94 Batch 900 Loss 0.1558\n","Epoch 94 Batch 1000 Loss 0.1568\n","Epoch 94 Loss 0.037591\n","Time taken for 1 epoch 147.29 sec\n","\n","Epoch 95 Batch 0 Loss 0.1715\n","Epoch 95 Batch 100 Loss 0.1489\n","Epoch 95 Batch 200 Loss 0.1388\n","Epoch 95 Batch 300 Loss 0.1503\n","Epoch 95 Batch 400 Loss 0.1667\n","Epoch 95 Batch 500 Loss 0.1472\n","Epoch 95 Batch 600 Loss 0.1392\n","Epoch 95 Batch 700 Loss 0.1491\n","Epoch 95 Batch 800 Loss 0.1510\n","Epoch 95 Batch 900 Loss 0.1501\n","Epoch 95 Batch 1000 Loss 0.1625\n","Epoch 95 Loss 0.037153\n","Time taken for 1 epoch 147.57 sec\n","\n","Epoch 96 Batch 0 Loss 0.1567\n","Epoch 96 Batch 100 Loss 0.1445\n","Epoch 96 Batch 200 Loss 0.1313\n","Epoch 96 Batch 300 Loss 0.1519\n","Epoch 96 Batch 400 Loss 0.1786\n","Epoch 96 Batch 500 Loss 0.1421\n","Epoch 96 Batch 600 Loss 0.1536\n","Epoch 96 Batch 700 Loss 0.1655\n","Epoch 96 Batch 800 Loss 0.1325\n","Epoch 96 Batch 900 Loss 0.1411\n","Epoch 96 Batch 1000 Loss 0.1514\n","Epoch 96 Loss 0.037105\n","Time taken for 1 epoch 148.15 sec\n","\n","Epoch 97 Batch 0 Loss 0.1728\n","Epoch 97 Batch 100 Loss 0.1442\n","Epoch 97 Batch 200 Loss 0.1545\n","Epoch 97 Batch 300 Loss 0.1389\n","Epoch 97 Batch 400 Loss 0.1486\n","Epoch 97 Batch 500 Loss 0.1506\n","Epoch 97 Batch 600 Loss 0.1339\n","Epoch 97 Batch 700 Loss 0.1492\n","Epoch 97 Batch 800 Loss 0.1548\n","Epoch 97 Batch 900 Loss 0.1515\n","Epoch 97 Batch 1000 Loss 0.1544\n","Epoch 97 Loss 0.037056\n","Time taken for 1 epoch 147.14 sec\n","\n","Epoch 98 Batch 0 Loss 0.1507\n","Epoch 98 Batch 100 Loss 0.1435\n","Epoch 98 Batch 200 Loss 0.1449\n","Epoch 98 Batch 300 Loss 0.1471\n","Epoch 98 Batch 400 Loss 0.1428\n","Epoch 98 Batch 500 Loss 0.1467\n","Epoch 98 Batch 600 Loss 0.1482\n","Epoch 98 Batch 700 Loss 0.1425\n","Epoch 98 Batch 800 Loss 0.1765\n","Epoch 98 Batch 900 Loss 0.1245\n","Epoch 98 Batch 1000 Loss 0.1518\n","Epoch 98 Loss 0.036969\n","Time taken for 1 epoch 148.58 sec\n","\n","Epoch 99 Batch 0 Loss 0.1926\n","Epoch 99 Batch 100 Loss 0.1390\n","Epoch 99 Batch 200 Loss 0.1505\n","Epoch 99 Batch 300 Loss 0.1385\n","Epoch 99 Batch 400 Loss 0.1458\n","Epoch 99 Batch 500 Loss 0.1457\n","Epoch 99 Batch 600 Loss 0.1319\n","Epoch 99 Batch 700 Loss 0.1759\n","Epoch 99 Batch 800 Loss 0.1481\n","Epoch 99 Batch 900 Loss 0.1578\n","Epoch 99 Batch 1000 Loss 0.1464\n","Epoch 99 Loss 0.036869\n","Time taken for 1 epoch 146.28 sec\n","\n","Epoch 100 Batch 0 Loss 0.1545\n","Epoch 100 Batch 100 Loss 0.1620\n","Epoch 100 Batch 200 Loss 0.1483\n","Epoch 100 Batch 300 Loss 0.1505\n","Epoch 100 Batch 400 Loss 0.1486\n","Epoch 100 Batch 500 Loss 0.1612\n","Epoch 100 Batch 600 Loss 0.1483\n","Epoch 100 Batch 700 Loss 0.1454\n","Epoch 100 Batch 800 Loss 0.1688\n","Epoch 100 Batch 900 Loss 0.1341\n","Epoch 100 Batch 1000 Loss 0.1522\n","Epoch 100 Loss 0.037031\n","Time taken for 1 epoch 147.42 sec\n","\n"]}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"ojjAGC6Yd6lm","outputId":"fba0bcce-e3d1-4a08-a0dc-312b3e827559"},"source":["print(val_loss_plot)"],"execution_count":null,"outputs":[{"output_type":"stream","name":"stdout","text":["[<tf.Tensor: shape=(), dtype=float32, numpy=0.23036732>, <tf.Tensor: shape=(), dtype=float32, numpy=0.1923207>, <tf.Tensor: shape=(), dtype=float32, numpy=0.17641157>, <tf.Tensor: shape=(), dtype=float32, numpy=0.1626834>, <tf.Tensor: shape=(), dtype=float32, numpy=0.14980216>, <tf.Tensor: shape=(), dtype=float32, numpy=0.13769525>, <tf.Tensor: shape=(), dtype=float32, numpy=0.12663545>, <tf.Tensor: shape=(), dtype=float32, numpy=0.11675103>, <tf.Tensor: shape=(), dtype=float32, numpy=0.108159766>, <tf.Tensor: shape=(), dtype=float32, numpy=0.10069339>, <tf.Tensor: shape=(), dtype=float32, numpy=0.0943507>, <tf.Tensor: shape=(), dtype=float32, numpy=0.08848612>, <tf.Tensor: shape=(), dtype=float32, numpy=0.08349086>, <tf.Tensor: shape=(), dtype=float32, numpy=0.07959088>, <tf.Tensor: shape=(), dtype=float32, numpy=0.07546627>, <tf.Tensor: shape=(), dtype=float32, numpy=0.07188845>, <tf.Tensor: shape=(), dtype=float32, numpy=0.06901307>, <tf.Tensor: shape=(), dtype=float32, numpy=0.06645331>, <tf.Tensor: shape=(), dtype=float32, numpy=0.06532782>, <tf.Tensor: shape=(), dtype=float32, numpy=0.06182967>, <tf.Tensor: shape=(), dtype=float32, numpy=0.059647538>, <tf.Tensor: shape=(), dtype=float32, numpy=0.057864588>, <tf.Tensor: shape=(), dtype=float32, numpy=0.057636652>, <tf.Tensor: shape=(), dtype=float32, numpy=0.05519193>, <tf.Tensor: shape=(), dtype=float32, numpy=0.053402495>, <tf.Tensor: shape=(), dtype=float32, numpy=0.053098965>, <tf.Tensor: shape=(), dtype=float32, numpy=0.05123428>, <tf.Tensor: shape=(), dtype=float32, numpy=0.050058056>, <tf.Tensor: shape=(), dtype=float32, numpy=0.05234058>, <tf.Tensor: shape=(), dtype=float32, numpy=0.052986722>, <tf.Tensor: shape=(), dtype=float32, numpy=0.046475474>, <tf.Tensor: shape=(), dtype=float32, numpy=0.04505856>, <tf.Tensor: shape=(), dtype=float32, numpy=0.047280744>, <tf.Tensor: shape=(), dtype=float32, numpy=0.04691777>, <tf.Tensor: shape=(), dtype=float32, numpy=0.04432003>, <tf.Tensor: shape=(), dtype=float32, numpy=0.04302533>, <tf.Tensor: shape=(), dtype=float32, numpy=0.04318342>, <tf.Tensor: shape=(), dtype=float32, numpy=0.043717165>, <tf.Tensor: shape=(), dtype=float32, numpy=0.04302818>, <tf.Tensor: shape=(), dtype=float32, numpy=0.041885372>, <tf.Tensor: shape=(), dtype=float32, numpy=0.040922392>, <tf.Tensor: shape=(), dtype=float32, numpy=0.040105846>, <tf.Tensor: shape=(), dtype=float32, numpy=0.04052717>, <tf.Tensor: shape=(), dtype=float32, numpy=0.039789036>, <tf.Tensor: shape=(), dtype=float32, numpy=0.040002283>, <tf.Tensor: shape=(), dtype=float32, numpy=0.03928763>, <tf.Tensor: shape=(), dtype=float32, numpy=0.03847522>, <tf.Tensor: shape=(), dtype=float32, numpy=0.03881884>, <tf.Tensor: shape=(), dtype=float32, numpy=0.037286486>, <tf.Tensor: shape=(), dtype=float32, numpy=0.03836844>, <tf.Tensor: shape=(), dtype=float32, numpy=0.0387118>, <tf.Tensor: shape=(), dtype=float32, numpy=0.036991186>, <tf.Tensor: shape=(), dtype=float32, numpy=0.036453426>, <tf.Tensor: shape=(), dtype=float32, numpy=0.036054384>, <tf.Tensor: shape=(), dtype=float32, numpy=0.0371168>, <tf.Tensor: shape=(), dtype=float32, numpy=0.03551669>, <tf.Tensor: shape=(), dtype=float32, numpy=0.035523377>, <tf.Tensor: shape=(), dtype=float32, numpy=0.035846394>, <tf.Tensor: shape=(), dtype=float32, numpy=0.035118084>, <tf.Tensor: shape=(), dtype=float32, numpy=0.03499495>, <tf.Tensor: shape=(), dtype=float32, numpy=0.03713759>, <tf.Tensor: shape=(), dtype=float32, numpy=0.035949625>, <tf.Tensor: shape=(), dtype=float32, numpy=0.033005595>, <tf.Tensor: shape=(), dtype=float32, numpy=0.033122677>, <tf.Tensor: shape=(), dtype=float32, numpy=0.034359172>, <tf.Tensor: shape=(), dtype=float32, numpy=0.033425238>, <tf.Tensor: shape=(), dtype=float32, numpy=0.032933034>, <tf.Tensor: shape=(), dtype=float32, numpy=0.03356567>, <tf.Tensor: shape=(), dtype=float32, numpy=0.03323962>, <tf.Tensor: shape=(), dtype=float32, numpy=0.033326507>, <tf.Tensor: shape=(), dtype=float32, numpy=0.03417777>, <tf.Tensor: shape=(), dtype=float32, numpy=0.031903964>, <tf.Tensor: shape=(), dtype=float32, numpy=0.03363083>, <tf.Tensor: shape=(), dtype=float32, numpy=0.0323785>, <tf.Tensor: shape=(), dtype=float32, numpy=0.031435344>, <tf.Tensor: shape=(), dtype=float32, numpy=0.031719204>, <tf.Tensor: shape=(), dtype=float32, numpy=0.03296215>, <tf.Tensor: shape=(), dtype=float32, numpy=0.031540766>, <tf.Tensor: shape=(), dtype=float32, numpy=0.031127105>, <tf.Tensor: shape=(), dtype=float32, numpy=0.03149339>, <tf.Tensor: shape=(), dtype=float32, numpy=0.031848643>, <tf.Tensor: shape=(), dtype=float32, numpy=0.031359773>, <tf.Tensor: shape=(), dtype=float32, numpy=0.032338638>, <tf.Tensor: shape=(), dtype=float32, numpy=0.030206027>, <tf.Tensor: shape=(), dtype=float32, numpy=0.030389525>, <tf.Tensor: shape=(), dtype=float32, numpy=0.03083208>, <tf.Tensor: shape=(), dtype=float32, numpy=0.030877851>, <tf.Tensor: shape=(), dtype=float32, numpy=0.030174538>, <tf.Tensor: shape=(), dtype=float32, numpy=0.03062648>, <tf.Tensor: shape=(), dtype=float32, numpy=0.030977225>, <tf.Tensor: shape=(), dtype=float32, numpy=0.03062937>, <tf.Tensor: shape=(), dtype=float32, numpy=0.030067489>, <tf.Tensor: shape=(), dtype=float32, numpy=0.029919483>, <tf.Tensor: shape=(), dtype=float32, numpy=0.029894644>, <tf.Tensor: shape=(), dtype=float32, numpy=0.029973866>, <tf.Tensor: shape=(), dtype=float32, numpy=0.030021654>, <tf.Tensor: shape=(), dtype=float32, numpy=0.029744348>, <tf.Tensor: shape=(), dtype=float32, numpy=0.029217536>, <tf.Tensor: shape=(), dtype=float32, numpy=0.029283587>, <tf.Tensor: shape=(), dtype=float32, numpy=0.030614179>]\n"]}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/","height":294},"id":"sM16k_UpppDu","outputId":"1ee2267f-264b-4279-bdc5-e47b9259402d"},"source":["plt.plot(loss_plot, label = 'train loss')\n","plt.plot(val_loss_plot, label = 'val loss')\n","plt.xlabel('Epochs')\n","plt.ylabel('Loss')\n","plt.title('Kor Full dataset Train & Validation Loss Plot')\n","plt.legend()\n","plt.show()SA"],"execution_count":null,"outputs":[{"output_type":"display_data","data":{"image/png":"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\n","text/plain":["<Figure size 432x288 with 1 Axes>"]},"metadata":{"needs_background":"light"}}]},{"cell_type":"markdown","metadata":{"id":"xGvOcLQKghXN"},"source":["## 캡셔닝\n","\n","- 각 타임스텝에서 decoder는 Decoder의 hidden state, encoder 출력값 그리고 이전 예측값을 입력값으로 받습니다.\n","- 모델이 end 토큰을 예측하면 예측이 중지됩니다.\n","- 그리고 타임스텝마다 attention 가중치를 저장합니다."]},{"cell_type":"code","metadata":{"id":"RCWpDtyNRPGs"},"source":["def evaluate(image):\n","    attention_plot = np.zeros((max_length, attention_features_shape))\n","\n","    hidden = decoder.reset_state(batch_size=1)\n","\n","    temp_input = tf.expand_dims(load_image(image)[0], 0)\n","    img_tensor_val = image_features_extract_model(temp_input)\n","    img_tensor_val = tf.reshape(img_tensor_val, (img_tensor_val.shape[0],\n","                                                 -1,\n","                                                 img_tensor_val.shape[3]))\n","\n","    features = encoder(img_tensor_val)\n","\n","    dec_input = tf.expand_dims([tokenizer.word_index['<start>']], 0)\n","    result = []\n","\n","    for i in range(max_length):\n","        predictions, hidden, attention_weights = decoder(dec_input,\n","                                                         features,\n","                                                         hidden)\n","\n","        attention_plot[i] = tf.reshape(attention_weights, (-1, )).numpy()\n","\n","        predicted_id = tf.random.categorical(predictions, 1)[0][0].numpy()\n","        result.append(tokenizer.index_word[predicted_id])\n","\n","        if tokenizer.index_word[predicted_id] == '<end>':\n","            return result, attention_plot\n","\n","        dec_input = tf.expand_dims([predicted_id], 0)\n","\n","    attention_plot = attention_plot[:len(result), :]\n","    return result, attention_plot"],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"id":"mXYnXwIZqBcf"},"source":["def plot_attention(image, result, attention_plot):\n","    temp_image = np.array(Image.open(image))\n","\n","    fig = plt.figure(figsize=(10, 10))\n","\n","    len_result = len(result)\n","    for i in range(len_result):\n","        temp_att = np.resize(attention_plot[i], (8, 8))\n","        \n","        #grid_size = max(np.ceil(len_result/2), 2)\n","        grid_size = max(np.ceil(len_result/4), 4)\n","        ax = fig.add_subplot(grid_size, grid_size, i+1)\n","        ax.set_title(result[i])\n","        img = ax.imshow(temp_image)\n","        ax.imshow(temp_att, cmap='gray', alpha=0.5, extent=img.get_extent())\n","\n","    plt.tight_layout()\n","    plt.show()"],"execution_count":null,"outputs":[]},{"cell_type":"markdown","metadata":{"id":"IzKsrhLFJw99"},"source":["# TEST 단계\n"]},{"cell_type":"code","metadata":{"id":"IVvdIeelx0w5","colab":{"base_uri":"https://localhost:8080/"},"outputId":"c905e976-bb27-4cb8-9c56-d8c93d3dff70"},"source":["# Download image files\n","image_folder = '/val2014/'\n","if not os.path.exists(os.path.abspath('.') + image_folder):\n","  image_zip = tf.keras.utils.get_file('val2014.zip',\n","                                      cache_subdir=os.path.abspath('.'),\n","                                      origin='http://images.cocodataset.org/zips/val2014.zip',\n","                                      extract=True)\n","  PATH = os.path.dirname(image_zip) + image_folder\n","  os.remove(image_zip)\n","else:\n","  PATH = os.path.abspath('.') + image_folder"],"execution_count":null,"outputs":[{"output_type":"stream","name":"stdout","text":["Downloading data from http://images.cocodataset.org/zips/val2014.zip\n","6645014528/6645013297 [==============================] - 139s 0us/step\n","6645022720/6645013297 [==============================] - 139s 0us/step\n"]}]},{"cell_type":"code","metadata":{"id":"z0YyAitYKSlK"},"source":["annotation_file = 'captions_kor_validation.json'#s.path.dirname(annotation_zip)+'/annotations/captions_train2014.json'\n","with open(annotation_file, 'r') as f:\n","    annotations = json.load(f)\n","\n","kor_json = annotations\n","\n","kor_val_caption = []\n","for i in range(len(kor_json['annotations'])):\n","    for j in range(len(kor_json['annotations'][i])):\n","      caption = {'caption': kor_json['annotations'][i]['caption'],\n","                 'id': str(kor_json['annotations'][i]['id']).zfill(12), #'{0:12d}'.format(kor_json[i]['id']),\n","                 #'id' : kor_json[i]['id'],\n","                 'image_id' : kor_json['annotations'][i]['id']}\n","      kor_val_caption.append(caption)\n","print(len(kor_val_caption))\n","annotations['annotations'] = kor_val_caption"],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"id":"kKokV9UhYOQ8"},"source":["df = pd.DataFrame(kor_val_caption)\n","df.drop_duplicates(subset=None, keep='first', inplace=False, ignore_index=False)\n","grouped_df = df.groupby(\"id\")\n","\n","grouped_lists = grouped_df[\"caption\"].agg(lambda column: \"^\".join(column))\n","\n","df = grouped_lists.to_frame()\n","captions = df['caption'].str.split('^').tolist()\n","val_df = pd.DataFrame(captions)\n","val_df.index = grouped_lists.index\n","\n","kor_val_caption = val_df.values.tolist()\n","\n","val_df = val_df.loc[:, ~val_df.T.duplicated()]"],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"id":"zF_zb3ssc0cT"},"source":["## BLEU 계산\n","\n","bleu_avg = 0\n","bleu_1 = 0\n","bleu_2 = 0\n","bleu_3 = 0\n","bleu_4 = 0\n","\n","for i in range(len(kor_val_caption)):\n","  image_path = 'val2014/COCO_val2014_{}.jpg'.format(val_df.index[i])\n","\n","  result, attention_plot = evaluate(image_path)\n","\n","  reference = list(map(lambda ref: ref.split(), kor_val_caption[i]))\n","\n","  candidate =  result[:-1]\n","\n","  smooth = SmoothingFunction()\n","  score_avg = sentence_bleu(reference,candidate, smoothing_function=smooth.method1, weights=(0.25,0.25,0.25,0.25))\n","  score_1 = sentence_bleu(reference,candidate, smoothing_function=smooth.method1, weights=(1,0,0,0))\n","  score_2 = sentence_bleu(reference,candidate, smoothing_function=smooth.method1, weights=(0,1,0,0))\n","  score_3 = sentence_bleu(reference,candidate, smoothing_function=smooth.method1, weights=(0,0,1,0))\n","  score_4 = sentence_bleu(reference,candidate, smoothing_function=smooth.method1, weights=(0,0,0,1))\n","\n","  bleu_1 = bleu_1 + score_1\n","  bleu_2 = bleu_2 + score_2\n","  bleu_3 = bleu_3 + score_3\n","  bleu_4 = bleu_4 + score_4\n","  bleu_avg = bleu_avg + score_avg\n","\n","bleu_1 = bleu_1 / len(kor_val_caption) * 100\n","bleu_2 = bleu_2 / len(kor_val_caption) * 100\n","bleu_3 = bleu_3 / len(kor_val_caption) * 100\n","bleu_4 = bleu_4 / len(kor_val_caption) * 100\n","bleu_avg = bleu_avg / len(kor_val_caption) * 100\n","\n","print('bleu_1 : ' , bleu_1)\n","print('bleu_2 : ' , bleu_2)\n","print('bleu_3 : ' , bleu_3)\n","print('bleu_4 : ' , bleu_4)\n","print('bleu_avg : ', bleu_avg)"],"execution_count":null,"outputs":[]},{"cell_type":"code","source":["values = [bleu_1, bleu_2, bleu_3, bleu_4, bleu_avg]\n","values"],"metadata":{"id":"W3R_CXjhpJhN"},"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/","height":541},"id":"AVVhpnFiXVOZ","outputId":"f7e7c799-cf8d-436d-9c57-7e4c15ad2a67"},"source":["# BLEU visualizatoin\n","\n","column = ['BLEU-1', 'BLEU-2', 'BLEU-3', 'BLEU-4', 'BLEU-avg']\n","values = [bleu_1, bleu_2, bleu_3, bleu_4, bleu_avg]\n","fig = plt.figure(figsize=(10, 8))\n","ax = fig.add_subplot(111)\n","bar = plt.bar(column, values, align='center')\n","plt.title('BLEU score of Kor test set', fontsize = 15)\n","plt.ylabel('% of BLEU score', fontsize = 15)\n","plt.xticks(fontsize = 15)\n","plt.yticks(fontsize = 15)"],"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["(array([ 0. ,  2.5,  5. ,  7.5, 10. , 12.5, 15. , 17.5, 20. , 22.5]),\n"," <a list of 10 Text major ticklabel objects>)"]},"metadata":{},"execution_count":108},{"output_type":"display_data","data":{"image/png":"iVBORw0KGgoAAAANSUhEUgAAAnYAAAHqCAYAAACJGANcAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nO3dedxcZX338c+XHSSlSELrholWpWrR2miLVQQVXKj6YBFtH33EVnEv1hULamwVcKEuXaxIFXCpca+A+4JiwSVsWsXiwiJqMRHZSSDwe/44JzgOM/d9T3LfmZmTz/v1Oq/JnHOu6/xmjphvrnOuM6kqJEmSNP22GncBkiRJmh8GO0mSpI4w2EmSJHWEwU6SJKkjDHaSJEkdYbCTJEnqCIOdJG0GSe6d5HNJLkvy7HHXI6mbDHbSFi7J6UmqZ1md5GNJ7tazz6FJ1s/Qx6F9ffQup7b7rEjywyHtv5DkxHn/cJPlJGANsC/wif6N7Xf11L51D02yNskzNk+JkqadwU4SNKFjEfBbwD7A9sCnk2w7Qh83t330LwfPb6lT637ASVX1w6q6fLadk9yLJgC+qareM5+FJNknyYvmuc/HJzl0Pvvs63+vJCsWqn+pKwx2kgDWV9W1VXVNVV0AvAq4J3CvUTpp++hf1i5IxdNnW2DdXHZMsgT4FPDpqnrVAtSyDzCvwQ54PHDoPPfZay/gNQvYv9QJBjtJg2zTvl451iqAJPdN8uUk1yT5eZIPJ1nWs/2xSf4ryfVJfpHk/W0wIslDknwtyQ1J1iQ5PsmuPW1PT/LBJMcluWrDpeIkWyc5MslFSdYl+U6SA2eo8T5JPp3k2iRXJlmZ5M7tthVJNvx245fbS65LZ+hrB+A/gcuAv+rbtlWSV/bUdUGSZ/ZsX9r2/5wkX01yU5Kj+vq4GPgH4K4bLpf3bNsryWeSXJfkiiTvTrKoZ/thSX7YftffT/Kadn0Bfw08rO3z4iGfbWD7dtuOSd6c5Gft5eevJ/mTdtvpwHs3HGu271DakhnsJP2GJItpRuxOrqrLxl0P8H7gJ8DvAY8Cvgs8GCDJIcApwFfb7XsDq4ClSf4I+DzwOZqRx/2BewOnJun9/74nAL9LMyL0uHbdvwAvBl4C3A04GfhEkgf0F5fkLsAZwEU0l1v3phmdOz3JzsDRNJekAR7b/vmSIZ81NAFmMXBQVd3Yt/2NwAuA5wJLgdcCxyV5Tt9+xwLvBu7c9tfr3jTB7lJ+fbmcJPcAvgZc0X6O/YHlNJfpaT/7O9vvZSnwfOBPkmzd9nFS235Re4z+72mm9gAfo7ls/1TgHsBZwOfagPwYYEOA3VDzsO9Q2rJVlYuLyxa8AKcDa2lu7L8CKJq/oJf07HMozeXaYX0c2rZbM2A5oN1nBfDDIe2/AJw4ZNvVwEsHrA9wMfDuIe0+Any2b90ewHrgsT2f/efADj373B24BXhaX9svDzoW8Gbg+0B61t2u/S6f17OugH1n+A4L+B5wLfAz4Lf7tt8euBH4i771r2w/w1Y0gamAt8xyzo8CLu5bd3L7ObbtWfewtr+7Ak9s/7x4SJ8nAKfPcMyh7YGHt9se2rNuK+BHwKvb909t/soa/38zLi6TvDhiJwma0ZL7t8uDgZ8C57U38M/VzT199C5f3cTa3g38fZLX9l5GpRnVuSvtiNIAfwR8sXdFVV0K/IDmHrMNzqzfvA/wETQB6sN9/X0L+MMhx/lSVd16SbOqrgO+0Xecudi1PcYvaQJjrz+gGQn8Yt/6L9GMON6zb92oHgmsrKqbetZ9q329f3vcS4EzkvyfJBmx/5naPxK4pKrO2LCiqm4Bzmbwdy5pCIOdJIDrq+qyqrq0qs4C/pJm5OgfRumk7aN/2RCa1gNbD2m6NXDTkG1/C7wMeA5wSZKj2st3S9rtwy4X34FmxLDfapoAtcGv+rYvoQlQl7X35a1Jsobm0uEum3CcuTiqqn5Ac6n1GUn27TsOA461un2d6TPNxRLgxX2f+dJ22y5VdRXwIOCbNKOh5yd52Fw7n6X9EuCOvcduj/94Bn/nkobYZvZdJG1pqurmJOcz4F6pTfBLYLch2xbT/IU/qJYC/iXJCTTh7o00o4OfbHf5HZpLdv3+d8jxFjM8DEITiq5g8EjRoGf5bexxBlkHUFVfS3IS8K4ke1XVDe1xaI+1uqfN4vb1MoYH57n4FfAWbntPHjTfB9U8puXp7aSHtwCfSXLvqrpoLgcY1r499n/TBLl+c5pJLKnhiJ2k22gvk+0JDHyg8EY6E1i0YaZjz7Hu1B7rrJkaV9W6qnobzYzRvWnuB/sF8JQhTf6L5hJf77H2oLlkefYMh/oKTVjaYcDo4/8O2P+/gEf0XlpMcru2xpmOM5uX09xXt6J9fz5wHX2fqX1/Oc3l87kqmnsUe30FuPeQUdfrf6Nx1cXA/wV24NcBeFCfgw9+2/ZfoZn8ct2AY28Isc2NlaNfApa2KI7YSQLYpp3BCbA7cDhwH/qeddazT6/rZ9m+vqrWVtX5SVYCH0nzSwpn09wn926a+9E+2d+wDYEvA/4N+DawjOa+tTe3o4pHAe9IcjnwQZqRvEOAc2lmo65K8mrgPTQjXf9Ec9/YZ4Z9EVX13STvBz6a5CXtcZfQzJj9VVW9s6/JW2kunf5LkjfTPNz5dTSjUCcPO85sqmpNkle0n++DVXVukrcAb07yK+A8mskNrwBeWVW3jJB5/hfYPckD23p/Cvw98PUkPwLeRxMi9wJeUFWPS3IMzb2HH6OZ3PEM4AZ+fR/e/wKPTvIHNPf8nVtVt142nqX9T9vP8/Ekr6T5B8WdaEL7WVX1n/x6xPKAJFcAa6vqO3P9wNIWY9yzN1xcXMa70MwMrZ7lBprRtQN69jm0b5/e5b6zbD+1p5/taEagfkRzie0ymktyi4bU9lvA64EL27ouoXnEx1Y9+zyJJhjeQHOJ8iPA77bbHkozonYDzb1pJwC37/vsJww47rbtcS6mCSM/pQkk9xhS532Bz9Lcl3gVzcSLu/TtM5dZsU/tW5e2/rNpLrNuBfxdW9c6mlHLw3r2X9r285BZzvkuNLN8r2+/0/u36x9I83iYa2lmI38beFG77QHtd/BzmtD3DeARfcdeRTPD+vvAnfqOOVv7RcDbaELfurauk3rO5TbAR9vaLgcePe7/dlxcJnFJ1a0TuSRJkjTFvMdOkiSpIwx2kiRJHWGwkyRJ6giDnSRJUkf4uBNg8eLFtXTp0nGXIUmSNKuzzz57TVUtGbTNYAcsXbqUVatWjbsMSZKkWSW5ZNg2L8VKkiR1hMFOkiSpIwx2kiRJHWGwkyRJ6giDnSRJUkcY7CRJkjrCYCdJktQRBjtJkqSOMNhJkiR1hMFOkiSpIwx2kiRJHWGwkyRJ6giDnSRJUkcY7CRJkjrCYCdJktQRBjtJkqSOMNhJkiR1hMFOkiSpIwx2kiRJHbHNuAvYUiw94rRxl9A5Fx974LhLkCRpojhiJ0mS1BEGO0mSpI4w2EmSJHWEwU6SJKkjDHaSJEkdYbCTJEnqCIOdJElSRxjsJEmSOsJgJ0mS1BEGO0mSpI4w2EmSJHWEwU6SJKkjDHaSJEkdYbCTJEnqCIOdJElSRxjsJEmSOsJgJ0mS1BEGO0mSpI4w2EmSJHXEWIJdkl2S/GOSy5LckOT7Sf6mZ/tWSV6b5OdJrktyWpK7zqHfFya5qO3zjCR7LewnkSRJmhzjGrFbCSwBDgL2AF4DHJfkye32VwHPAv4cuBdwA/DZJNsN6zDJ04GjgRcAS4HzgC8kWbxAn0GSJGmijCvYvaSqnlZV36qq1VW1Evg8cGCSHYGXAq+qqjOr6jLgr4G7AE+aoc9XAW+tqtOq6nLgRcD1wLMX9qNIkiRNhrEEu6r67oDVNwI7AnsDOwOn9Ox/FXAGcMCg/pLcDbh7X5ubgc8MayNJktQ1EzF5IskdgP1pRu3uBVxVVb/o2+1C4J5DurhXzz5zbSNJktQpYw92SQL8O3AxcBKwK3DlgF2vbLcNsmH9VXNtk+SwJKuSrFq9evWoZUuSJE2csQc7mvvp9gX+sqrW0dR0y4D9bga2HtLHVkBVVc21TVUdX1XLq2r5kiVLNqpwSZKkSTLWYJfkscAxwDOr6vx29TU099j1WwRcPaSra5rustMIbSRJkjplbMEuyX2BDwKvq6oP9Gz6EbA4SX+4Wwb8eEh3P+rZZ65tJEmSOmVcDyjeHTgVOLWqVvRt/hqwnmYyxYb9twceBnyxZ13vJdbvAZdz2xmwB/S2kSRJ6rLNHuzahwx/Avgl8DdJdu5ZdqqqK4F3AG9OsjzJHYF30jyT7r1tHy8CrkvyIICqugV4A3Bkkocn2T3JMTQPP/7nzf0ZJUmSxmGbMRzzjjTPqgPon456Cc2vRryMZuLDp2jukzsDeGRVXdfudyOwjmZkb4O30jwH7ySaX7U4F9i/fcCxJElS5232YFdVFwOZZZ8bgRe3y6Dt/wr8a9+6ovlJsaPnpVBJkqQpMwmPO5EkSdI8MNhJkiR1hMFOkiSpIwx2kiRJHWGwkyRJ6giDnSRJUkcY7CRJkjrCYCdJktQRBjtJkqSOMNhJkiR1hMFOkiSpIwx2kiRJHWGwkyRJ6giDnSRJUkcY7CRJkjrCYCdJktQRBjtJkqSOMNhJkiR1hMFOkiSpIwx2kiRJHWGwkyRJ6giDnSRJUkcY7CRJkjrCYCdJktQRBjtJkqSOMNhJkiR1hMFOkiSpIwx2kiRJHWGwkyRJ6giDnSRJUkcY7CRJkjrCYCdJktQRBjtJkqSOMNhJkiR1hMFOkiSpI8Ya7JK8PEklOapn3Yp23aDl9Bn6OnFImyM2y4eRJEkas23GcdAkOwErgd8HrurbfDTw5gHN3g2snaXrY9r2vdZtTI2SJEnTZizBDtgBOBN4CvDd3g1VdSNwY++6JEuAxwP7z9LvjVV17TzWKUmSNDXGEuyq6gqa0TWSzKXJXwM/rKozFrIuSZKkaTbxkyeSbAU8G3jHuGuRJEmaZBMf7IDHAkuA985h3+cluSzJj5N8JsmTMmRIMMlhSVYlWbV69ep5LViSJGkcpiHYPQ/4QFVdPct+/wQ8GngocAiwCvgPbjuZAoCqOr6qllfV8iVLlsxnvZIkSWMxrskTc5LkbsCjgAfOtm9Vnd3z9iJgVZK1wJFJVlSVs2MlSVKnTfqI3XOAVVV1zka2P5dmBu6i+StJkiRpMk1ssEuyPfBXwL8N2b71HLrZD/hJVa2Zz9okSZIm0cQGO+DJNPV9sH9DkicCa9tXkuzVTpZ4bJI9kixrf83icOCVm7VqSZKkMZnke+yeC5xcVTcM2HYTza9Q3NS+vxD4Bs2z8e4KbEdzGfagqjp1M9QqSZI0dmMPdlW1dMj6vWdocwo9981V1VrgNe0iSZK0RZrkS7GSJEkagcFOkiSpIwx2kiRJHWGwkyRJ6giDnSRJUkcY7CRJkjrCYCdJktQRBjtJkqSOMNhJkiR1hMFOkiSpIwx2kiRJHWGwkyRJ6giDnSRJUkcY7CRJkjrCYCdJktQRBjtJkqSOMNhJkiR1hMFOkiSpIwx2kiRJHWGwkyRJ6giDnSRJUkcY7CRJkjrCYCdJktQRBjtJkqSOMNhJkiR1hMFOkiSpIwx2kiRJHWGwkyRJ6giDnSRJUkcY7CRJkjrCYCdJktQRBjtJkqSOMNhJkiR1hMFOkiSpIwx2kiRJHWGwkyRJ6oixBrskL09SSY7qW396u75/ecos/R2S5IIka5Ock2S/hf0EkiRJk2MswS7JTklOAQ4Drhqy23OARX3LR2bocz/gfcAxwF2ADwOnJdlzHkuXJEmaWOMasdsBOBO4H3DlkH3WVdW1fcv6Gfo8ElhZVSdX1eqqOgY4B3jJ/JYuSZI0mcYS7Krqiqo6pqqum4/+kuwA7AOc0rfpVOCA+TiGJEnSpOvK5IllwLbAhX3rLwTukmTHzV+SJEnS5jXJwe7oJD9LcmGSTyR55Az77tq+9l/WvRIIsEt/gySHJVmVZNXq1avnqWRJkqTxmdRgdxTwGOBPgUOBnwGfT/LsIftv+By39K2/uX3dur9BVR1fVcuravmSJUs2vWJJkqQx22bcBQxSVV/reXsRcGaS7YBXAO8c0OSa9nXnvvWL2ter57dCSZKkyTOpI3aDnAvsPmTbRUDR3GvXaxmwpqquuW0TSZKkbpmmYLcfTbgDIMmtl1er6mpgFbedAfso4IubpTpJkqQxm7hgl+TPkqxM8vAkd0xyryRvA55Ac+8dSZ4IrG1fNzgGOCzJwUkWJ3kBsD/wxs3+ISRJksZgEu+x+y/gEcC/AHdu130deFhVndm+vwlY274CUFUfT/JC4FiaX574H+CgqjpncxUuSZI0TmMPdlW1tO/9r4C/bZdhbU7h1xMjetefAJwwzyVKkiRNhYm7FCtJkqSNY7CTJEnqCIOdJElSRxjsJEmSOsJgJ0mS1BEGO0mSpI4w2EmSJHWEwU6SJKkjDHaSJEkdYbCTJEnqCIOdJElSRxjsJEmSOmKkYJfkjkmOSvLRJJ9Pslu7/nFJdl+YEiVJkjQXcw52Sf4YuAA4FFgP/Cmwfbv5wcAb5rs4SZIkzd0oI3ZvAf4duEdVPRm4rmfbSmDfeaxLkiRJI9pmhH3/EHhSVdWAbZcBvzs/JUmSJGljjDJidwXwe0O2/V67XZIkSWMySrD7D+A9SR7cs66S7AIcDXxoXiuTJEnSSEYJdkcB/wOckeTHwG8BpwCXtP0cNf/lSZIkaa7mfI9dVa0FHpPk4cAjgCXAr4BXA58ecu+dJEmSNpM5B7sknwOeX1VfAr60cCVJkiRpY4xyKfZPgWsXqhBJkiRtmlGC3YeBv12oQiRJkrRpRnmO3RXA85M8guZS7K/6d6iqo+erMEmSJI1mlGB3EPAL4PbAwQO2F81jTyRJkjQGo8yKXbaQhUiSJGnTjDJiB0CSPYF9gF2BnwBnV9X/zHdhkiRJGs0ojzvZHjgReDJwA7AGWAzskORDwNOr6saFKFKSJEmzG2VW7N8DewEPrqrbVdVdq+p2wEPb9f+wEAVKkiRpbkYJdv8XeGZVfb13ZVWdCRwGPHU+C5MkSdJoRgl2uwEXD9n2Y5rZspIkSRqTUYLd94DHDdn2BOCCTS9HkiRJG2uUWbF/D3wwyf1pHlC8hmYU75HA04GnzH95kiRJmqtRnmP3n0kOAV4NPLNtezPwDeDxVfWFhSlRkiRJczHq405WVdUDk2wFLAHWVNXNSX4/yU5Vdf2CVSpJkqQZjXKP3XHAcwGq6paquryqbm63PR1423wXJ0mSpLkbJdj9H+C9Q7b9O/CYUQ+e5OVJKslRfevvkOTdSX6R5Pok5yeZ8R6+JCe2ffUvR4xalyRJ0jQaZfLEbsCvhmxbywiPO0myE7AS+H3gqgG7fKFdHg6spvm1iw8kuaSqzpqh62OAo/vWrZtrXZIkSdNslBG7C4C/HLLtQODCEfraATgTuB9w5YDth1TV4VX13+0l37cD32f2UcEbq+ravuWmEeqSJEmaWqOM2L0VOCHJIuBk4KfA7jSPOXkd8JK5dlRVV9CMrpFk0PbvDmh2I7DjCPVKkiRtUUZ53MnJSX6H5nEnf9+z6SbgzVX1zvkuboP22Xl7AS9fqGNIkiRNu1FG7KiqNyV5B7A3zeNOrgS+3o7ALYj2MSvvAb5UVZ+bZffnJXkmzejehTSTOj5SVTWg38NofuOWPfbYY36LliRJGoM532OX5PZJ9mnvW/t8VX0AWAS8OslBC1ci/wjckeaRKjP5J+DRwEOBQ4BVwH9w28kUAFTV8VW1vKqWL1myZB7LlSRJGo9RRuzeCFwLfBUgybOBfwHOBp6Z5GVV9Y75LK49xrOA/avqpzPtW1Vn97y9CFiVZC1wZJIVVeXsWEmS1GmjzIp9JHASQJIdgdcCr6+qPwYOBl40n4UleQTNKNxhVfWVjezmXJoZuIvmrTBJkqQJNUqwWwxc0v75MGB74E3t+1XAneerqCT3Aj4CHFdVJw7ZZ+s5dLUf8JOqWjNftUmSJE2qUYLdd4EXJ3kizczY46rq2nbbnsDP56OgJLsCp9I85+7oJDv3LDu2+zwRWNu+kmSvJJ9J8tgkeyRZ1v6axeHAK+ejLkmSpEk3yj12LwU+TvMLE/9J+xy61jOA0+appvsBv9cuV/dt+wqwL80jVta2r9DMgP1GW9Ndge1oLsMeVFWnzlNdkiRJE22U59idkWR3YJeq6v9psU8BX9uYAqpqad/704HbPrX4N/c5hZ775qpqLfCadpEkSdoijfocu1sY8HuxVfXReatIkiRJG2WUe+wkSZI0wQx2kiRJHWGwkyRJ6ogZ77FL8uoZNt8EXAGcVVXfnteqJEmSNLLZJk88bYZtWwO7ATsn+Q/g0KpaP2+VSZIkaSQzBruqusdsHSR5EPAx4AjgdfNUlyRJkka0yffYVdU3aR5Q/PRNL0eSJEkba74mT8zrb8VKkiRpdPMV7O4JrJmnviRJkrQRNjnYJbkjcALwkU0vR5IkSRtrtsedfG6GzVsDi4HfB74JvGoe65IkSdKIZnvcyc+AGrJtPfAtmkB3SlUN20+SJEmbwWyPOzl0M9UhSZKkTTTjPXZJ3p5km75190yyVd+6bZO8fSEKlCRJ0tzMNnni+cB2feu+Bvxu37rt2n0lSZI0JrMFu8xxnSRJksZstmDnhAhJkqQpsTEjdpIkSZpAcxmx6x+1G7ROkiRJYzbbc+wCfDtJb5C7PfC1JDf37SdJkqQxmi3YvXazVCFJkqRNNtsDiucc7JLstOnlSJIkaWPNdo/dnCS5HXDNfPQlSZKkjTMvwa7lfXaSJEljNJ/BzpmykiRJYzSfwU6SJEljZLCTJEnqiBlnxSb5AXO7xOr9dZIkSWM223Ps3rdZqpAkSdImm7fn2EmSJGm8ZhuxA259Tt0NVXVL+3434GXAvYEfAO+sqgsXrEpJkiTNasbJE0mWJPkicDVwTZJXJ9kROAM4DPgd4GnAOUkesODVSpIkaajZRuzeCOwMHAxsC7wY2Av4GfDAqrouyfbAycDRwKMXsFZJkiTNYLZgtz9wYFWdD5Dk68DFwL5VdR1AVa1L8hrgmwtZqCRJkmY223Ps7gD8aMObqroUuAq4qG+/nwC3G/XgSV6epJIcNWDbC5NclOSGJGck2WsO/R2S5IIka5Ock2S/UWuSJEmaVrMFuwC39K1bD9y8KQdNslOSU2ju07tqwPan01zafQGwFDgP+EKSxTP0uR/N41mOAe4CfBg4Lcmem1KrJEnStJjLrNiXJbmp5/0OwN8kubpn3XYjHncH4EzgKcB3B2x/FfDWqjoNIMmLgMcBzwZeP6TPI4GVVXVy+/6YJAcCLwGeNWJ9kiRJU2e2YHcpcGjfujXAk4fsOydVdQXNyBrJb/5oRZK7AXcHTunZ/+YknwEOYECwS7IDsA/w1L5NpwLPnWtdkiRJ02y2BxQv3Ux19LpX+9r/XLwLgScMabOMZtbuoDZ3SbJjVd0wfyVKkiRNntnusRuHXdvX/nvvruzZNqzNlQPaBNilv0GSw5KsSrJq9erVG1urJEnSxJjEYLcVUFVVfetvBraeoQ3cdqLHhkket2lXVcdX1fKqWr5kyZKNLlaSJGlSTGKwuwZIkp361i+i+QWMYW2geZhyfxtmaCdJktQZQ4NdktcmuWPP+33SP9NhYWx4bt6yvvXLgB8PaXMRUEParKmqa27bRJIkqVtmGrF7Br/50OEvAzsubDkAfA+4nGYGbK8DgC9ueJPk1surVXU1sGpAm0f1tpEkSeqymWbFrgLekWQlsI5mEsJfJrlxWIOeZ8httKq6JckbgCOTnA/8N/C3wB7APwMkeSKwMsmTq+pjbdNjgA8kOQM4neYZefsDe29qTZIkSdNgpmD3AuBtwEtpHkBcwFHt6yAFbHKwa72VZnTwJGAJcC6wf1Vd1m6/CVjbvjYHr/p4khcCx9L88sT/AAdV1TnzVJMkSdJEGxrsqupnwJM2vE9yC3Dvqrp+PgsY9Ky8dkbs0e0yqM0p/HpiRO/6E4AT5rM+SZKkaTHKrNhn0IySSZIkaQLN5bdiAaiqk5LcIcnhND/ftStwGXA28PZ2hE+SJEljMucRuyT3Br5D87Ne5wEf4tc/8/XtdrskSZLGZM4jdsCbgJVV9fz+DUn+FTgOeMx8FSZJkqTRjBLsHkozU3aQN9OM4kmSJGlMRpk8cTOw7ZBt2/Hr32WVJEnSGIwS7D4HvD7J9r0rk+wA/EO7XZIkSWMyyqXYlwBfAX6a5OvAGmA3ml92uALYb/7LkyRJ0lzNecSu/dWH+wErgF8Av0MT7o4A7lNVP12IAiVJkjQ3o4zYUVXX0v5eqyRJkibLKPfYSZIkaYIZ7CRJkjrCYCdJktQRI91j1yvJVsCewNbABVW1ft6qkiRJ0sg2asQuyR8CP6R5/MnpwMVJHjKPdUmSJGlEMwa7JPccsuk9wIuraklV7Qa8Bjh+vouTJEnS3M02YvedJMcnuVPf+iU0I3UbnNGukyRJ0pjMFuyWA78L/CDJcUl2a9e/Hfh6kjcmeRPwaeBdC1inJEmSZjFjsKuq71TV44FHAA8AfpxkBfCvwAtpJl9sA/xtVf3dAtcqSZKkGcxpVmxVnQXsl2R/4PXA84FjgFdW1boFrE+SJElzNNvkiSVJ3pfk8iRX0IzSPRV4DvDXwA+TPCvJ1puhVkmSJM1gtnvsPgisAx4F/ClwPvDJqvoo8AfAq4G/Ay5I8pSFLFSSJEkzmy3YPRhYUVXnVdUFNI81WZbkdlV1S1W9B7gn8M/AWxa4VkmSJM1gtmB3GvCvSR6f5NE0z6o7v6qu27BDVd1UVW8H7r6AdUqSJGkWswW7pwPfBVYAbwBuBB43aMequn5eK5MkSdJIZpwV247MHdEukiRJmmAb9VuxkiRJmjwGO0mSpI4w2EmSJHWEwU6SJKkjDHaSJEkdYbCTJEnqCIOdJElSRxjsJEmSOsJgJ0mS1BEGO0mSpI6YyGCX5NAkNWS5eEibFUP2/7fNXL4kSdJYTGSwA94HLBqwHAdcOkO79w9oc/iCVipJkjQhthl3AYNU1Xrg2t51SclC72kAABB4SURBVLYFngq8bIam66vq2hm2S5IkddakjtgN8uc0QfTD4y5EkiRpEk1TsHse8J6qWjvuQiRJkibRVAS7JPcFHgK8c5Zdn5Dk0iSXJPlykme2l3AH9XlYklVJVq1evXrea5YkSdrcpiLY0YzWfaGqfjjDPh8CHgU8DHgC8CngrcDJg3auquOranlVLV+yZMl81ytJkrTZTeTkiV5JFtFMmjh0pv2q6nt9q85L8nPgvUleUVUzzaaVJEmaetMwYvc04BrgkxvR9tz2dff5K0eSJGkyTUOwey5wQvsIlFsl2XoObfcDbgC+vxCFSZIkTZKJDnZJ9gF+H3hX3/oHAdcleVH7fpckZyV5UpK7J7lzkucBxwKv89l2kiRpSzDp99g9Fzi1qi7rW78eWNcu0Fyq/RDwCmAZsDPwPeAFVXXi5ilVkiRpvCY62FXVXwxZfw6wS8/7W4C3tIskSdIWaaIvxUqSJGnuDHaSJEkdYbCTJEnqCIOdJElSRxjsJEmSOsJgJ0mS1BEGO0mSpI4w2EmSJHWEwU6SJKkjDHaSJEkdYbCTJEnqCIOdJElSRxjsJEmSOsJgJ0mS1BEGO0mSpI4w2EmSJHWEwU6SJKkjDHaSJEkdYbCTJEnqCIOdJElSRxjsJEmSOsJgJ0mS1BEGO0mSpI4w2EmSJHWEwU6SJKkjDHaSJEkdsc24C5AmydIjTht3CZ1z8bEHjrsESdpiOGInSZLUEQY7SZKkjjDYSZIkdYTBTpIkqSMMdpIkSR1hsJMkSeoIg50kSVJHGOwkSZI6wmAnSZLUERMb7JKcmKQGLEfM0ObhSc5JsjbJBUkO2Zw1S5IkjdPEBrvWMcCivuW4QTsm2RM4FVgJ3AU4Fnhfkn03S6WSJEljNum/FXtjVV07x31fCpxdVW9o35+UZH/gSOD0hShOkiRpkkz6iN0oDgBO6Vt3KvCwJNuPoR5JkqTNqhPBLslONJdfL+zbdCGwLbB0QJvDkqxKsmr16tULX6QkSdICm/Rg97wklyX5cZLPJHlSkgzY77fb1yv71m94v2t/g6o6vqqWV9XyJUuWzGfNkiRJYzHJwe6fgEcDDwUOAVYB/wEcPWDfDZ/jlr71N7evWy9EgZIkSZNkYidPVNXZPW8vAlYlWQscmWRFVa3r2X5N+7pzXzeL2terF6hMSZKkiTHJI3aDnAvswK8DGwBVdRXwS2BZ3/7LgKIJhpIkSZ02bcFuP+AnVbUmSf/l1S/RzIzt9SjgmyM8MkWSJGlqTWSwS7JXO1nisUn2SLIsyVHA4cArkzwIuC7Ji3qavQl4TJLnJlnc/urEsxh8T54kSVLnTOo9dhcC36D55Ym7AtvRXIY9qKpOTfIAYF27AFBV30pyMPB64K3ApcBhVfXJzV28JEnSOExksKuqtcBr2mXQ9nOAXQas/yRgkJMkSVukibwUK0mSpNEZ7CRJkjrCYCdJktQRBjtJkqSOMNhJkiR1xETOipUkSQtv6RGnjbuEzrn42APHenxH7CRJkjrCYCdJktQRBjtJkqSOMNhJkiR1hMFOkiSpIwx2kiRJHWGwkyRJ6giDnSRJUkcY7CRJkjrCYCdJktQRBjtJkqSOMNhJkiR1hMFOkiSpIwx2kiRJHWGwkyRJ6giDnSRJUkcY7CRJkjrCYCdJktQRBjtJkqSOMNhJkiR1hMFOkiSpIwx2kiRJHWGwkyRJ6giDnSRJUkcY7CRJkjrCYCdJktQRBjtJkqSOMNhJkiR1hMFOkiSpIyY22CXZJck/JrksyQ1Jvp/kb2bYf0WSGrD82+asW5IkaVy2GXcBM1gJrAYOAi4GHg68L8nlVbVySJv3A8/pW3fTglUoSZI0QSY52L2kqr7b835lkqcDB9KEvkHWV9W1C1+aJEnS5JnYS7F9oW6DG4EdN3ctkiRJ02Big12/JHcA9gc+P+5aJEmSJtFUBLskAf6d5l67k2bY9QlJLk1ySZIvJ3lmkm2H9HlYklVJVq1evXoBqpYkSdq8Jvkeu14vBfYF9q6qdUP2+RDwKZoJF7vQjO69FXgE8Bf9O1fV8cDxAMuXL6/5L1nSQll6xGnjLqFTLj72wHGXIGmeTHywS/JY4Bjg/1XV+cP2q6rv9a06L8nPgfcmeUVVXbqQdUqSJI3bRF+KTXJf4IPA66rqAxvRxbnt6+7zV5UkSdJkmtgRuyS7A6cCp1bVigHbt66qm2fpZj/gBuD781+hJGkYL5fPPy+Zay4mcsQuyXbAJ4BfAn+TZOeeZackDwKuS/Kidv9dkpyV5ElJ7p7kzkmeBxxLM9rns+0kSVLnTeqI3R2Bvds/909ZvQR4IrCuXQCuoZk88QpgGbAz8D3gBVV14kIXK0mSNAkmMthV1cVAZtltl579bwHe0i6SJElbpIm8FCtJkqTRGewkSZI6wmAnSZLUEQY7SZKkjjDYSZIkdYTBTpIkqSMMdpIkSR1hsJMkSeoIg50kSVJHGOwkSZI6wmAnSZLUEQY7SZKkjjDYSZIkdYTBTpIkqSMMdpIkSR1hsJMkSeoIg50kSVJHGOwkSZI6wmAnSZLUEQY7SZKkjjDYSZIkdYTBTpIkqSMMdpIkSR1hsJMkSeoIg50kSVJHGOwkSZI6wmAnSZLUEQY7SZKkjjDYSZIkdYTBTpIkqSMMdpIkSR1hsJMkSeoIg50kSVJHGOwkSZI6wmAnSZLUEQY7SZKkjpjoYJfkkCQXJFmb5Jwk+82y/8Pb/da27Q7ZXLVKkiSN28QGuzbEvQ84BrgL8GHgtCR7Dtl/T+BUYGW7/7HA+5Lsu1kKliRJGrOJDXbAkcDKqjq5qlZX1THAOcBLhuz/UuDsqnpDu/9JwIfafiRJkjpvIoNdkh2AfYBT+jadChwwpNkBQ/Z/WJLt57dCSZKkyTORwQ5YBmwLXNi3/kLgLkl27F2ZZCeay6+D9t8WWLowZUqSJE2ObcZdwBC7tq9X9q2/EgiwC3BDz/rfnmH/3v5uleQw4LD27bVJ/mejq+2excCacRcxm7xh3BWM1VScI/A8MQXnaQs/R+B5mgZTcY5gs52nuw7bMKnBbsNI4i19629uX7fexP2pquOB4ze2wC5Lsqqqlo+7Dg3nOZoOnqfp4HmafJ6juZvUS7HXtK87961f1L5evYn7S5Ikdc6kBruLgKK5167XMmBNVV3Tu7KqrgJ+OWT/avuTJEnqtIkMdlV1NbCK286AfRTwRYAk/ZdXvzRk/29W1bULUWeHeYl68nmOpoPnaTp4niaf52iOUlXjrmGgJAcBHwCeBpwOPAU4Dtib5t7ArwJHVNVb2/0fCPwXcDjNw4wfDrwXeFJVfXJz1y9JkrS5TeSIHUBVfRx4Ic0vSPyUZgbrQVV1DrAeWNcuG/b/FnAw8Lx2/9cDhxnqJEnSlmJiR+wkSZI0mokdsdNtJTk9SfUsq5N8LMnd2u2HJlk/Q/tD+9r3Lqe2+6xI8sMh7b+Q5MSNqHuXJGe2x7nzqO2nyTSeoyTLk3w2yTVJfpXki0n+aJQ+ps20nac0npXkG0mua+v9zyT3GfGjT5VpO08D2h/SHmvFxvYxbtN+DrZEBrvpcxLNY1x+i+Zn17YHPp1k2zm2v7lt378cPP+lQpIHA99p69xSTM05SnIn4NPAZ4H7APejuZXhtCS/Nd/HmzBTc56Ae9PcmnIccDfgwcBNwGeT7LIAx5sk03SebpVkV+CNwA8W8jibyVSegy3VpD6gWMOt75nle0GSVwFnA/eaawebeZbwbsCzaX4p5Mub8bjjNE3n6Grgj6rq0g0rkrwUuBxYTjPbvKum6Tz9GHhQVa1t31+e5LnAL4D9gE9spjrGYZrOU683Ae8DHjKGY8+3aT0HWyRH7KbfhnDe/3NqE6GqTqmqT4+7jjGb2HNUVdf0hrrWje3rjv37d9wkn6cbekLdBju1rxNX7wKb2PO0QZJ9gX2B1423kgUzsecgyY5JXp7kvPb2kp8leXOStNvPSPL2vjZJckmSF7fvH5Dm9qG1SS5O8pz2NogVY/hIIzPYTbEki4FXASdX1WXjrke3NaXn6OnAdcCZ4y5kc5m285TkjsBbaJ73ecaYy9lspuE8JdkBeCfw/AFhfOpNwTm4J83l4sNpfk/1EJqnZRzSbn8/8OdJevPPnwB3Bj6YZDfgc8ClNLen/EXb/r6bpfp5YLCbPk9NsibJFcBqYFfgpSO037pt37/0P9x5VkmuHLL85ah9dczUnqMkd6cZZTi2qn416vGmzFSepyQX0dwHeSfgwKq6+TYddsu0nadXAedV1WdH7X+CTdM5+HZV/VlVfaWqrqiqrwFfA/643f5hYAnwpz3dPgk4vap+RvMP222Bv6qqH1XVWcCfM0X3iXuP3fT5GPBymlB+J+BFwHlJHj7H9jcD9x+wfs1G1DLsXzBXbERfXTKV5yjJ9sCHgHNonh/ZdVN5noCHAfcA/g74YpKHtD+r2FVTc56S/AHNPcV/sBF9T7KpOQdVVe2o4uOABwB70oS6n7bbf5nkszQjeGe0l2gPBla0/TwIOKuqrt/QcVX9IMn3N6LWsTDYTZ/re4a/L03yTeB7wD8An5pLB7MMn68H+n+ubYOtaWbizaWfLdm0nqN3A4uBx1bV0McXdMhUnqf2nsgN9V4KPJNmtmxXTdN5ehewoqp+Ppe6psjUnIMkDwFOoXkaw2dpRuvWA+nZ7f3APyY5nCb07Q58tN22GzDo/G1MCB0LL8VOufYyzPk0j0OYD7+k+R/2IItxNG5k03COkryG5l+4f1ZVl29CbVNrGs5Tr6q6Bvg+cPdN6WfaTPh5+mPg9b2XCGlmxR7Rvu/CDNlJPwdvA06rqn2q6vVVtRK4pm+fT9I8buUhNCN3p/WMev8S+O0B/U7NY4UMdlOuHUbeExj4cMeNcCawKMmf9B3nTu1xzpqn42wxJv0cJXkycBTwlKr6zjzVOHUm+TwluXeSbfrW7Qj8HvCjeah1akzyeQKW0TwL8v49yyrg33r+PPUm/BzsSfNb8hv6WEQzmeJW7WXWT9CEuoNpRvA2+A6wd5pJMBv6+J2236ngpdjps02Snds/704z8+c+NPc87AHQs73XrfcLDNm+vqrWVtX5SVYCH0nyDJpnFd2D5jLdN2j+paOZTc05SvJA4ETgNcBX+457Y1XdOLBhN0zNeQIOAh6X5HXAN4Hb09wHeQvwnhH6mUZTc56q6uL+dUnWAlcO2jZFpuYcAP8NPC3JF2ke2fQWBj+66f004W4tcFrP+ncBRwDvTnIkzcje22n+W7tlhDrGp6pcpmQBTgeqZ7mB5l86B7TbD+3b3rvcd5btp/YcZzuaG0l/BKwDLqP5j2PRJtS+b3ucO4/7e/Qc/Ua9K2Y43opxf5+ep1v72Z5mssS3af4iuobmL6Vl4/4uPU9z+gwrxv1dbinngCZwfpXmkU0/onlUyYeAE/v224bmQez/PqCP+9EEynU0ky5eSjNqePi4z8dclrQfQpIkSX3aJwb8DDi4qib+F5S8x06SJInmknGSdyV5UJLF7SNsTqQZufvqzK0ng8FOkiSpcSPNZeGP04zSfYHmcSkH1JQ8DNxLsZIkSR3hiJ0kSVJHGOwkSZI6wmAnSZLUEQY7SZKkjjDYSZIkdYTBTpIkqSP+PwlRnyIpK22KAAAAAElFTkSuQmCC\n","text/plain":["<Figure size 720x576 with 1 Axes>"]},"metadata":{"needs_background":"light"}}]},{"cell_type":"markdown","source":["# 추가 실험\n","- 이 모델에게 새로운 이미지를 주고, 제대로 학습이 이루어졌는지 확인해봅시다. \n","- 아 물론, COCO Dataset 이외의 데이터를 주게 되면, 실험 결과가 부정확할 확률이 높지만, 재미로 해봅시다. \n","- 아래 URL을 바꾸면 이미지를 바꿀 수 있습니다. "],"metadata":{"id":"s_chTFtW-ouv"}},{"cell_type":"code","source":[""],"metadata":{"id":"Ss_FcTDq-pO-"},"execution_count":null,"outputs":[]}]}