Fix fp quantized matmul corruption when the quantized dim is not a multiple of 32 - #3912

Fix fp quantized matmul corruption when the quantized dim is not a multiple of 32#3912

kapellirohith wants to merge 1 commit into

ml-explore:mainml-explore/mlx:mainfrom

kapellirohith:fp-qmm-tail-fixkapellirohith/mlx:fp-qmm-tail-fixCopy head branch name to clipboard