OSSForks/pytorch
1
0
Fork 0
mirror of https://github.com/zebrajr/pytorch.git synced 2025-12-06 12:20:52 +01:00
Code Issues Actions 38 Packages Projects Releases Wiki Activity

Add inference_mode hint message to use eval with inference. (#163619)

Browse Source 
Fixes #162923

## Test Result

### Before

<img width="985" height="889" alt="image" src="https://github.com/user-attachments/assets/41de5cfa-7b25-4ba4-ade8-a6df745dcb30" />

### After

<img width="913" height="977" alt="image" src="https://github.com/user-attachments/assets/b6c06860-8db3-4b5d-9d46-31ece01fb04d" />

Pull Request resolved: https://github.com/pytorch/pytorch/pull/163619
Approved by: https://github.com/jbschlosser
This commit is contained in:
zeshengzong 2025-09-24 20:07:11 +00:00 committed by PyTorch MergeBot
parent 0dce2afd44
commit 90a282504e
1 changed files with 6 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

6
torch/autograd/grad_mode.py
View File

@ -237,6 +237,12 @@ class inference_mode(_DecoratorContextManager):
Unlike some other mechanisms that locally enable or disable grad,
entering inference_mode also disables :ref:`forward-mode AD <forward-mode-ad>`.
.. warning::
`inference_mode` does NOT automatically set the model to evaluation mode.
For proper inference behavior (e.g., disabling dropout, using running statistics
in batch normalization), you must explicitly set your model to evaluation mode using
`model.eval()` in addition to using this context manager.
Args:
mode (bool or function): Either a boolean flag to enable or disable
inference mode, or a Python function to decorate with inference