Fast in-loop MCQA #281
Merged
Fast in-loop MCQA #281
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
epwalsh
pushed a commit
that referenced
this pull request
May 27, 2025
Incorporate this one-line PR: allenai/OLMo-in-loop-evals#12 TL;DR: #281 made in-loop RC and BPB slower, this fixes that bug. **The RC/BPB in-loop evals run with `ai2-olmo-eval~=0.8.0` are correct evals, just slower.**
TianhuaTao
pushed a commit
that referenced
this pull request
May 28, 2025
Bump in-loop evals to v0.8.1. This will add "fast" MCQA, which performs MC tasks in 1 forward pass instead of 4 forward passes. (We extract the A/B/C/D logits from a single pass). allenai/OLMo-in-loop-evals#8 This will make the MC tasks 4x faster, and produces the same numbers. Also, added Java, Rust and C++ translated MBPP BPB.
TianhuaTao
pushed a commit
that referenced
this pull request
May 28, 2025
Incorporate this one-line PR: allenai/OLMo-in-loop-evals#12 TL;DR: #281 made in-loop RC and BPB slower, this fixes that bug. **The RC/BPB in-loop evals run with `ai2-olmo-eval~=0.8.0` are correct evals, just slower.**
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Outdated suggestions cannot be applied.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.
Bump in-loop evals to v0.8.1. This will add "fast" MCQA, which performs MC tasks in 1 forward pass instead of 4 forward passes. (We extract the A/B/C/D logits from a single pass).
allenai/OLMo-in-loop-evals#8
This will make the MC tasks 4x faster, and produces the same numbers.
Also, added Java, Rust and C++ translated MBPP BPB.