Multi forward MCH eviction fix #2836
Open
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
facebook-github-bot
added
the
CLA Signed
|
This pull request was exported from Phabricator. Differential Revision: D71491003 |
aliafzal
force-pushed
the
export-D71491003
branch
from
29ea307 to
141e513
Compare
aliafzal
added a commit
to aliafzal/torchrec
that referenced
this pull request
Mar 20, 2025
Summary: ## Issue: Direct tensor modification during training with multiple forward passes breaks PyTorch's autograd graph, causing "one of the variables needed for gradient computation has been modified by an inplace operation" runtime error. ## Solution: Use in-place updates with .data accessor to safely reinitialize evicted embeddings without invalidating gradient computation. Differential Revision: D71491003
|
This pull request was exported from Phabricator. Differential Revision: D71491003 |
aliafzal
added a commit
to aliafzal/torchrec
that referenced
this pull request
Mar 20, 2025
Summary: Pull Request resolved: pytorch#2836 ## Issue: Direct tensor modification during training with multiple forward passes breaks PyTorch's autograd graph, causing "one of the variables needed for gradient computation has been modified by an inplace operation" runtime error. ## Solution: Use in-place updates with .data accessor to safely reinitialize evicted embeddings without invalidating gradient computation. Differential Revision: D71491003
aliafzal
force-pushed
the
export-D71491003
branch
from
141e513 to
4a19949
Compare
Summary: ## Issue: Direct tensor modification during training with multiple forward passes breaks PyTorch's autograd graph, causing "one of the variables needed for gradient computation has been modified by an inplace operation" runtime error. ## Solution: Use in-place updates with .data accessor to safely reinitialize evicted embeddings without invalidating gradient computation. Differential Revision: D71491003
aliafzal
force-pushed
the
export-D71491003
branch
from
4a19949 to
5e4633e
Compare
|
This pull request was exported from Phabricator. Differential Revision: D71491003 |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Summary:
Issue:
Direct tensor modification during training with multiple forward passes breaks PyTorch's autograd graph, causing "one of the variables needed for gradient computation has been modified by an inplace operation" runtime error.
Solution:
Use in-place updates with .data accessor to safely reinitialize evicted embeddings without invalidating gradient computation.
Differential Revision: D71491003