train_End2end.py

"""
@File : train_End2end.py
@Author: Dong Wang
@Date : 2020/5/8
"""

import numpy as np
from core.End2endNet import End2endNet
from core.End2endNet import train
from tensorflow.keras.models import load_model

## File paths
rootPath = 'D:/data/IIS/'
Subsystem1FolderPath = 'Dataset Subsystem 1/'
conditions = ["open_palm", "open_dorsal", "fist_palm", "fist_dorsal",
              "three_fingers_palm", "three_fingers_dorsal"]
fileType = '.webm'
height, width, channel = 240, 320, 3

# X_train = np.load(rootPath + Subsystem1FolderPath +
#                   'npy/X_train' + str(height) + '_' + str(width) + '_' + '.npy')
# y_train = np.load(rootPath + Subsystem1FolderPath +
#                   'npy/End2endY_trainCategorical'+str(height)+'_'+str(width)+'_'+'.npy')
#
# pretrained_weights='./saved_models/End2endWeights-best.hdf5'
# model = End2endNet(pretrained_weights='./saved_models/End2endWeights-best.hdf5')
#
# train(model=model, x=X_train, y=y_train, epochs=10, batch_size=8, validation_split=0.25)

## For test

X_test = np.load(rootPath + Subsystem1FolderPath +
                 'npy/X_test' + str(height) + '_' + str(width) + '_' + '.npy')
y_test = np.load(rootPath + Subsystem1FolderPath +
                 'npy/End2endY_testCategorical' + str(height) + '_' + str(width) + '_' + '.npy')
trained_model = load_model('./saved_models/End2endWeights-best.hdf5')
trained_model.evaluate(X_test, y_test)