def evaluate():
print('Loading model...')
model = load_model('')
_, _, test_X, _, _, test_Y = split_frame()
Y_original = test_Y
Y_predict = np.array(model.predict(test_X))
assert Y_original.shape == Y_predict.shape
print Y_predict.shape
def process():
train_X, val_X, test_X, train_Y, val_Y, test_Y = split_frame()
f = open('frame_content_data', 'wb')
pickle.dump(train_X, f)
pickle.dump(val_X, f)
pickle.dump(test_X, f)
pickle.dump(train_Y, f)
pickle.dump(val_Y, f)
pickle.dump(test_Y, f)
f.close()
def main():
print('Loading model...')
train_X, val_X, test_X, train_Y, val_Y, test_Y = split_frame()
print('Building model...')
model = Sequential()
model.add(Dense(1024, activation='relu', input_shape=(20, )))
model.add(Dropout(0.5))
model.add(Dense(1024, activation='relu'))
model.add(Dropout(0.5))
model.add(Dense(1024, activation='relu'))
model.add(Dropout(0.5))
model.add(Dense(1024, activation='relu'))
model.add(Dense(30))
model.add(Dense(1024, activation='relu'))
model.add(Dropout(0.5))
model.add(Dense(1024, activation='relu'))
model.add(Dropout(0.5))
model.add(Dense(1024, activation='relu'))
model.add(Dense(30))
model.add(Dense(1024, activation='relu'))
model.add(Dropout(0.5))
model.add(Dense(1024, activation='relu'))
model.add(Dropout(0.5))
model.add(Dense(1024, activation='relu'))
model.add(Dense(30))
print('Compiling model...')
model.compile(loss='mean_squared_error',
optimizer='adam',
metrics=['accuracy'],
sample_weight_mode=None)
print(model.summary())
print('Training model...')
model.fit(train_X,
train_Y,
batch_size=128,
epochs=15,
validation_data=(val_X, val_Y),
class_weight='auto',
callbacks=[
EarlyStopping(monitor='val_loss', patience=2, verbose=0),
ModelCheckpoint(filepath='model1_weights.{epoch:02d}.hdf5',
monitor='val_loss',
save_best_only=True,
verbose=0)
])
score, acc = model.evaluate(test_X, test_Y, batch_size=128)
print('Test score:', score)
print('Test accuracy:', acc)
def process():
mfcc, vad = split_frame()
f = open('mfcc_vad_3', 'wb')
pickle.dump(mfcc, f)
pickle.dump(vad, f)
f.close()
