Fix AutoAWQ -> MLX scales and bias dtype mismatch. Fixes #33.#38
Merged
Conversation
Closed
liang2kl
approved these changes
Apr 23, 2026
Member
liang2kl
left a comment
liang2kl
left a comment
There was a problem hiding this comment.
Good catch. This should be the change in #19 that breaks the inference (as also pointed out in ml-explore/mlx#3434).
Closed
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
2 participants
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
_convert_awq_linearemittedfp32scales andfp16biases, butmx.quantized_matmulrequires matching dtypes for the pair. This mismatch produced NaN logits on Apple Silicon, causing generation to stream token 0 ("!") for AutoAWQ paroquant checkpoints.Text generation works fine after applying this 2 line fix.
