[moe] Make capacity_factor configurable and add sweep script#4056
Closed
claude[bot] wants to merge 1 commit intomainfrom
Closed
[moe] Make capacity_factor configurable and add sweep script#4056claude[bot] wants to merge 1 commit intomainfrom
claude[bot] wants to merge 1 commit intomainfrom
Conversation
…eep script
Add capacity_factor field to GrugModelConfig (default 1.25, matching the
existing hardcoded value) so it can be varied in experiment sweeps. Add
sweep_capacity_factor.py to sweep over {1.0, 1.125, 1.25, 1.5, 2.0}.
Fixes #4017
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
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
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Add capacity_factor as a field on GrugModelConfig (default 1.25, preserving the existing hardcoded constant) and pass it through MoEMLP to moe_mlp() instead of using the module-level constant. Add sweep_capacity_factor.py that runs the trial model at capacity factors {1.0, 1.125, 1.25, 1.5, 2.0} to determine whether the default masks avoidable overflow or throughput loss.
Fixes #4017