from CNN import CNN from LoadData import LoadData load_train = LoadData() data_train = load_train.data_train train_generator, x_train, x_valid, y_train, y_valid = load_train.loadDataTrain( ) input_shape = (load_train.img_rows, load_train.img_cols, 1) epochs = 100 lr = 0.0002 batch_size = 32 cnn = CNN(input_shape, len(train_generator.class_indices), lr) model = cnn.ConvNetModel() # history = model.fit(x_train, y_train, # batch_size=batch_size, # epochs=epochs, # verbose=1, # validation_data=(x_valid, y_valid)) model.fit(x_train, y_valid, batch_size=batch_size, epochs=epochs, verbose=1, validation_data=(x_valid, y_valid)) evaluation = model.evaluate(x_valid, y_valid, batch_size=batch_size, verbose=1) print('loss : %.2f, accuracy : %.2f' % (evaluation[0], evaluation[1]))