Skip to content

【API test of cinn】modify api dropout dytostaic bug #68637

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.

Sign up for GitHub

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

Merged
merged 1 commit into from
Oct 12, 2024
Merged

【API test of cinn】modify api dropout dytostaic bug #68637

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
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: 0 additions & 3 deletions python/paddle/autograd/backward_utils.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -318,9 +318,6 @@ def _check_vjp_dynamic_shape(op, inputs):
for items in inputs:
for item in items:
if item.initialized() and -1 in item.shape:
warnings.warn(
f"[Prim] Decomp op does not support dynamic shape -1, but got shape {item.shape} in inputs of op {op.name()} . Prim will skip its vjp op."
)
return True


Expand Down
4 changes: 3 additions & 1 deletion python/paddle/nn/functional/common.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1274,7 +1274,9 @@ def get_attrs(prog, dropout_prob, is_test, seed):
dtype = x.dtype
keep_prob = 1 - p
if training:
if in_dynamic_or_pir_mode() and p == 1.0:
if in_dynamic_mode() and p == 1.0:
return paddle.scale(x, scale=0.0)
elif in_pir_mode() and isinstance(p, (float, int)) and p == 1.0:
return paddle.scale(x, scale=0.0)

scale_input = (
Expand Down