reduce_lr = ReduceLROnPlateau('val_loss',
factor=0.1,
patience=int(patience / 4),
verbose=1)
trained_models_path = base_path + '_mini_XCEPTION'
model_names = trained_models_path + '.{epoch:02d}-{val_acc:.2f}.hdf5'
model_checkpoint = ModelCheckpoint(model_names,
'val_loss',
verbose=1,
save_best_only=True)
callbacks = [model_checkpoint, csv_logger, early_stop, reduce_lr]
# loading dataset
faces, emotions = load_fer2013()
faces = preprocess_input(faces)
num_samples, num_classes = emotions.shape
xtrain, xtest, ytrain, ytest = train_test_split(faces,
emotions,
test_size=0.2,
shuffle=True)
model.fit_generator(data_generator.flow(xtrain, ytrain, batch_size),
steps_per_epoch=len(xtrain) / batch_size,
epochs=num_epochs,
verbose=1,
callbacks=callbacks,
validation_data=(xtest, ytest))
model.save("hdf.h5")
early_stop = EarlyStopping('val_loss', patience=patience)
reduce_lr = ReduceLROnPlateau('val_loss', factor=0.1,
patience=int(patience/4), verbose=1)
trained_models_path = base_path + '_mini_XCEPTION'
model_names = trained_models_path + '.{epoch:02d}-{val_acc:.2f}.hdf5'
model_checkpoint = ModelCheckpoint(model_names, 'val_loss', verbose=1,
save_best_only=True)
callbacks = [model_checkpoint, csv_logger, early_stop, reduce_lr]
# loading dataset
if os.path.isfile(base_data_path + "processed_data_without_pseudo.pickle"):
with open(base_data_path + "processed_data_without_pseudo.pickle", "rb") as f:
faces, emotions = pickle.load(f)
f.close()
else:
faces, emotions = load_fer2013(custom=True)
faces = preprocess_input(faces)
with open(base_data_path + "processed_data_without_pseudo.pickle", "wb") as f:
pickle.dump((faces, emotions), f)
f.close()
pseudo = False
if pseudo and os.path.isfile("./data/pseudo_data.pickle"):
with open("./data/pseudo_data.pickle", "rb") as f:
pseudoImages, pseudoLabels = pickle.load(f)
f.close()
pseudoImages = np.asarray(pseudoImages).reshape((len(pseudoImages), 48, 48, 1))
pseudoLabels = pd.get_dummies(pseudoLabels).as_matrix()
