未验证 提交 4247fac8 编写于 作者: T Tingquan Gao 提交者: GitHub

support Piecewise.learning_rate (#2899)

上级 657037c4
...@@ -403,10 +403,23 @@ class Piecewise(LRBase): ...@@ -403,10 +403,23 @@ class Piecewise(LRBase):
warmup_start_lr=0.0, warmup_start_lr=0.0,
last_epoch=-1, last_epoch=-1,
by_epoch=False, by_epoch=False,
learning_rate=None,
**kwargs): **kwargs):
if learning_rate:
decay_epochs = list(range(0, epochs, 30))[1:]
values = [
learning_rate * (0.1**i)
for i in range(len(decay_epochs) + 1)
]
logger.warning(
"When 'learning_rate' of Piecewise has beed set, "
"the learning rate scheduler would be set by the rule that lr decay 10 times every 30 epochs. "
f"So, the 'decay_epochs' and 'values' have been set to {decay_epochs} and {values} respectively."
)
super(Piecewise, super(Piecewise,
self).__init__(epochs, step_each_epoch, values[0], warmup_epoch, self).__init__(epochs, step_each_epoch, values[0], warmup_epoch,
warmup_start_lr, last_epoch, by_epoch) warmup_start_lr, last_epoch, by_epoch)
self.values = values self.values = values
self.boundaries_steps = [e * step_each_epoch for e in decay_epochs] self.boundaries_steps = [e * step_each_epoch for e in decay_epochs]
if self.by_epoch is True: if self.by_epoch is True:
......
Markdown is supported
0% .
You are about to add 0 people to the discussion. Proceed with caution.
先完成此消息的编辑!
想要评论请 注册