[feature] Skip creating the CPU grad tensor when training #821
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Outdated suggestions cannot be applied.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.
What does this PR do?
This issue came up when identifying memory consumption for CPU offload. We create a CPU grad tensor at the beginning of the FW pass for every parameter even during eval. By skipping this, we save memory roughly equal to that of the parameters.
I don't think this change should have any negative effects given that we reset params at the beginning of FW but let me know if I haven't thought of a corner case. Thanks!
Before submitting
PR review
Anyone in the community is free to review the PR once the tests have passed.
If we didn't discuss your PR in Github issues there's a high chance it will not be merged.