def test_split_with_larger_number(self):
sample_list = list(range(self.CHUNK_SIZE))
chunk, rest = utils.split(sample_list, self.CHUNK_SIZE + 1)
self.assertEqual(len(chunk), self.CHUNK_SIZE)
self.assertEqual(len(rest), 0)
def test_split(self):
sample_list = list(range(self.CHUNK_SIZE))
split_size = 10
chunk, rest = utils.split(sample_list, split_size)
self.assertEqual(len(chunk), split_size)
self.assertEqual(len(rest), self.CHUNK_SIZE - split_size)
def fit(self, X, n_epochs=50, batch_size=256):
indices_fracs = split(fracs=[0.9, 0.1], N=len(X), seed=0)
X_train, X_valid = X[indices_fracs[0]], X[indices_fracs[1]]
self.autoencoder.fit(X_train,
X_train,
epochs=n_epochs,
batch_size=batch_size,
shuffle=True,
validation_data=(X_valid, X_valid))
def fit_generator(self, X, n_epochs, batch_size=256):
indices_fracs = split(fracs=[0.9, 0.1], N=len(X), seed=0)
X_train, X_valid = X[indices_fracs[0]], X[indices_fracs[1]]
dataAugmentaion = ImageDataGenerator(fill_mode="nearest")
self.autoencoder.fit_generator(dataAugmentaion.flow(X_train,
X_train,
batch_size=32),
validation_data=(X_valid, X_valid),
steps_per_epoch=len(X_train) // 32,
epochs=10)
def fit(self, X, n_epochs=50, batch_size=256):
es = EarlyStopping(monitor='val_loss', verbose=1, patience=10)
indices_fracs = split(fracs=[0.9, 0.1], N=len(X), seed=0)
X_train, X_valid = X[indices_fracs[0]], X[indices_fracs[1]]
self.autoencoder.fit(X_train,
X_train,
epochs=n_epochs,
batch_size=batch_size,
shuffle=True,
validation_data=(X_valid, X_valid),
callbacks=[es])