Repository navigation
Conversation
Keep integrated TE routing on full-width hidden states while dispatching narrower expert activations. Return independent input cotangents through RouterComputationInfo and expose the input in the Flax wrapper. Runtime validation is deferred at the author request. [skip ci] Signed-off-by: Pawel Gadzinski <pgadzinski@nvidia.com>
This was referenced Nov 6, 2026
Drop the prose paragraph duplicating the router_input docstring. Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com> Signed-off-by: Pawel Gadzinski <pgadzinski@nvidia.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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Description
Kimi-K3 computes routing from full-width hidden states while its routed experts process a down-projected latent. Add an optional, independently shaped
RouterComputationInfo.router_inputand expose it through the Flax MoEBlock wrapper. Existing calls continue routing on the expert input.This is a draft for code review; runtime validation is deferred. The base branch is a snapshot of NVIDIA main at
e29c5d691dc0e9b05f285dc19ddbf44abe26c33a, so this fork PR contains only the feature change.This is the first PR in the stack: #73 → #74. pggPL/maxtext#1 uses the head of #74.
Type of change
Changes
Checklist: