fix: add rotary_percent to Megatron TransformerConfig patch so MiniMax-M2 hf->torch_dist conversion works#2190
Open
mvanhorn wants to merge 1 commit into
Open
Conversation
4 tasks
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Summary
Converting a MiniMax-M2 HF checkpoint to
torch_distformat no longer crashes withTransformerConfig got an unexpected keyword argument 'rotary_percent'; the bundled Megatron config patch now declares the field.Why this matters
Issue #2129 reports the crash when running the hf -> torch_dist conversion for MiniMax-M2. The conversion passes
rotary_percentto Megatron'sTransformerConfig, but that field is not present in the pinned Megatron-LM. slime already carries missingTransformerConfigfields in itsdocker/patch/*/megatron.patchfiles, so the field belongs there:Changes
rotary_percent: Optional[float] = Noneto theTransformerConfigblock in every versionedmegatron.patch(latest,v0.5.12.post1,v0.5.9,v0.5.7,v0.5.6,v0.5.5.post1,v0.5.0rc0-cu126) so the patch stays consistent across the image tags.The default
Nonepreserves existing behavior for models that do not set it.Testing
Each patch file is a unified diff applied during the Docker image build. Confirmed each modified hunk's header line counts stay consistent after the insertion so the patch still applies cleanly.
Fixes #2129