for i in range(freeze_layers[phi]):
model.layers[i].trainable = False
if True:
#--------------------------------------------#
# BATCH_SIZE不要太小，不然训练效果很差
#--------------------------------------------#
BATCH_SIZE = 4
Lr = 1e-3
Init_Epoch = 0
Freeze_Epoch = 50
gen = Generator(bbox_util, BATCH_SIZE, lines[:num_train],
lines[num_train:],
(image_sizes[phi], image_sizes[phi]), NUM_CLASSES)
model.compile(loss={
'regression': smooth_l1(),
'classification': focal()
},
optimizer=keras.optimizers.Adam(Lr))
print('Train on {} samples, val on {} samples, with batch size {}.'.
format(num_train, num_val, BATCH_SIZE))
model.fit(gen.generate(True, eager=False),
steps_per_epoch=max(1, num_train // BATCH_SIZE),
validation_data=gen.generate(False, eager=False),
validation_steps=max(1, num_val // BATCH_SIZE),
epochs=Freeze_Epoch,
verbose=1,
initial_epoch=Init_Epoch,
callbacks=[logging, checkpoint, reduce_lr, early_stopping])
for i in range(freeze_layers[phi]):
epoch_size = num_train // BATCH_SIZE
epoch_size_val = num_val // BATCH_SIZE
lr_schedule = tf.keras.optimizers.schedules.ExponentialDecay(
initial_learning_rate=Lr,
decay_steps=epoch_size,
decay_rate=0.95,
staircase=True)
print('Train on {} samples, val on {} samples, with batch size {}.'.
format(num_train, num_val, BATCH_SIZE))
optimizer = tf.keras.optimizers.Adam(learning_rate=lr_schedule)
for epoch in range(Init_Epoch, Freeze_Epoch):
fit_one_epoch(model, focal(), smooth_l1(), optimizer, epoch,
epoch_size, epoch_size_val, gen, gen_val,
Freeze_Epoch, get_train_step_fn())
for i in range(freeze_layers[phi]):
model.layers[i].trainable = True
if True:
#--------------------------------------------#
# BATCH_SIZE不要太小，不然训练效果很差
#--------------------------------------------#
BATCH_SIZE = 4
Lr = 5e-5
Freeze_Epoch = 50
Epoch = 100