未验证 提交 0e192063 编写于 作者: W wangguanzhong 提交者: GitHub

update lr for paddle 2.0 (#1564)

上级 aa834b2e
...@@ -61,7 +61,7 @@ class PiecewiseDecay(object): ...@@ -61,7 +61,7 @@ class PiecewiseDecay(object):
for i in self.gamma: for i in self.gamma:
value.append(base_lr * i) value.append(base_lr * i)
return optimizer.lr_scheduler.PiecewiseLR(boundary, value) return optimizer.lr.PiecewiseDecay(boundary, value)
@serializable @serializable
......
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