Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

add clear_grad for amp sample code #3455

Merged
merged 1 commit into from
Apr 25, 2021
Merged
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
3 changes: 3 additions & 0 deletions doc/paddle/api/paddle/amp/GradScaler_cn.rst
Original file line number Diff line number Diff line change
Expand Up @@ -46,6 +46,7 @@ GradScaler用于动态图模式下的"自动混合精度"的训练。它控制lo
scaled = scaler.scale(loss) # scale the loss
scaled.backward() # do backward
scaler.minimize(optimizer, scaled) # update parameters
optimizer.clear_grad()


.. py:function:: scale(var)
Expand Down Expand Up @@ -76,6 +77,7 @@ GradScaler用于动态图模式下的"自动混合精度"的训练。它控制lo
scaled = scaler.scale(loss) # scale the loss
scaled.backward() # do backward
scaler.minimize(optimizer, scaled) # update parameters
optimizer.clear_grad()

.. py:function:: minimize(optimizer, *args, **kwargs)

Expand Down Expand Up @@ -106,6 +108,7 @@ GradScaler用于动态图模式下的"自动混合精度"的训练。它控制lo
scaled = scaler.scale(loss) # scale the loss
scaled.backward() # do backward
scaler.minimize(optimizer, scaled) # update parameters
optimizer.clear_grad()



Expand Down