LoRA test: check that all the tensors are materialized. #1395
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
carmocca
reviewed
May 7, 2024
awaelchli
approved these changes
May 7, 2024
awaelchli
approved these changes
May 7, 2024
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Outdated suggestions cannot be applied.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.
Hi there 👋
As @awaelchli suggested, this PR extends the test for LoRA to make sure that all the tensors (parameters, buffers, attributes) are materialized.
With just a forward pass we depend on the config: if LoRA is applied to all matrices (Q, K and V), then we don't need
zero_padmethod during a forward pass and thus might not notice thatlora_indinside the method is a non-materialized tensor.