[BUG] `quantized_matmul` produces wrong results for GQA `expand_dims` broadcasting when M < `vector_limit`
May 4, 2026 · #3480
cpp
Difficulty: Medium
Labels
bug
Parent Repository
ml-explore/mlx
cpp repository
25,988 1,754
Labels
ml-explore/mlx
cpp repository
Sign in required
Authenticate to use favourites & bookmarks