Fix HellaSwag #4982
Closed
Fix HellaSwag #4982
Conversation
This file contains hidden or 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
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Outdated suggestions cannot be applied.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.
HellaSwag is broken on current master (see #4980). It is related to KV cache handling.
Instead of trying to sort it out, I just changed to evaluating the full ending (so context + ending together) for all 4 endings.
The performance hit is surprisingly low: it runs 400 tasks in 55 seconds with a 7B model (
fp16, not quantized), versus 49 seconds before this change (where is the time going? The number of tokens being evaluated is at least two times more). The result for LLaMA-v2 is now 77.00, as we had before, versus 53.00 on master.Left the existing version commented out for now.