def fit(self, train_x, train_y, validation_data_fit, train_loop_num, **kwargs):
val_x, val_y = validation_data_fit
# if train_loop_num == 1:
# patience = 2
# epochs = 3
# elif train_loop_num == 2:
# patience = 3
# epochs = 10
# elif train_loop_num < 10:
# patience = 4
# epochs = 16
# elif train_loop_num < 15:
# patience = 4
# epochs = 24
# else:
# patience = 8
# epochs = 32
epochs = 3
patience = 2
callbacks = [tf.keras.callbacks.EarlyStopping(monitor='val_loss', patience=patience)]
self._model.fit(train_x, ohe2cat(train_y),
epochs=epochs,
callbacks=callbacks,
validation_data=(val_x, ohe2cat(val_y)),
verbose=1, # Logs once per epoch.
batch_size=16,
shuffle=True)
def find_lr(data_index):
D = AutoSpeechDataset(
os.path.join(
r"/home/chengfeng/autospeech/data/data0{}".format(data_index),
'data0{}.data'.format(data_index)))
D.read_dataset()
metadata = D.get_metadata()
x_train, y_train = D.get_train()
my_model = CrnnModel()
x_train = my_model.preprocess_data(x_train)
log(f'x_train shape: {x_train.shape}; y_train shape: {y_train.shape}')
y_train = ohe2cat(y_train)
my_model.init_model(input_shape=x_train.shape[1:],
num_classes=metadata[CLASS_NUM])
lr_finder = LRFinder(my_model._model)
lr_finder.find(x_train,
y_train,
start_lr=0.0001,
end_lr=1,
batch_size=64,
epochs=200)
# Plot the loss, ignore 20 batches in the beginning and 5 in the end
lr_finder.plot_loss(n_skip_beginning=20, n_skip_end=5)
lr_finder.plot_loss_change(sma=20,
n_skip_beginning=20,
n_skip_end=5,
y_lim=(-0.01, 0.01))
def fit(self, train_x, train_y, validation_data_fit, train_loop_num,
**kwargs):
val_x, val_y = validation_data_fit
callbacks = [
tf.keras.callbacks.EarlyStopping(monitor='val_loss', patience=3)
]
epochs = 10 if train_loop_num == 1 else 30
self._model.fit(
train_x,
ohe2cat(train_y),
epochs=epochs,
callbacks=callbacks,
validation_data=(val_x, ohe2cat(val_y)),
verbose=1, # Logs once per epoch.
batch_size=32,
shuffle=True)
def fit(self, train_x, train_y, validation_data_fit, train_loop_num,
**kwargs):
val_x, val_y = validation_data_fit
# if train_loop_num == 1:
# patience = 2
# epochs = 3
# elif train_loop_num == 2:
# patience = 3
# epochs = 10
# elif train_loop_num < 10:
# patience = 4
# epochs = 16
# elif train_loop_num < 15:
# patience = 4
# epochs = 24
# else:
# patience = 8
# epochs = 32
patience = 2
# epochs = self.epoch_cnt + 3
epochs = 3
callbacks = [
tf.keras.callbacks.EarlyStopping(monitor='val_loss',
patience=patience)
]
self._model.fit(
train_x,
ohe2cat(train_y),
epochs=epochs,
callbacks=callbacks,
validation_data=(val_x, ohe2cat(val_y)),
# validation_split=0.2,
verbose=1, # Logs once per epoch.
batch_size=32,
shuffle=True,
# initial_epoch=self.epoch_cnt,
# use_multiprocessing=True
)
self.epoch_cnt += 3
def fit(self, train_x, train_y, validation_data_fit, train_loop_num, **kwargs):
val_x, val_y = validation_data_fit
epochs = 5
patience = 2
batch_size = 32
# over_batch = len(train_x) % batch_size
# append_idx = np.random.choice(np.arange(len(train_x)), size=batch_size-over_batch, replace=False)
# train_x = np.concatenate([train_x, train_x[append_idx]], axis=0)
# train_y = np.concatenate([train_y, train_y[append_idx]], axis=0)
callbacks = [tf.keras.callbacks.EarlyStopping(
monitor='val_loss', patience=patience)]
self._model.fit(train_x, ohe2cat(train_y),
epochs=epochs,
callbacks=callbacks,
validation_data=(val_x, ohe2cat(val_y)),
verbose=1, # Logs once per epoch.
batch_size=batch_size,
shuffle=True)
def fit(self, x_train, y_train, *args, **kwargs):
self._model.fit(x_train, ohe2cat(y_train))