Issues · ml-explore/mlx · GitHub

Why not implement this in Pytorch?
#12 · ToniRV opened on Dec 6, 2023
29
Common Installation Problems
#200 · awni opened on Dec 17, 2023
11

Labels Milestones New issue

[Feature Request] Support weightless RMSNorm (for FlashNorm weight folding trick)

#3465

· NilsGraf opened

on Apr 28, 2026

Metal: buffer destroyed mid-flight under concurrent custom-kernel workloads (untracked hazard mode + unretained CB)

#3461

· TheTom opened

on Apr 28, 2026

[BUG] Openclaw Tool calls not parsed: model outputs raw <tool_call> tokens instead of tool_calls field

#3460

· davidjotte opened

on Apr 27, 2026

[BUG] mx.clear_cache() causes SIGSEGV when streaming decoder holds state across calls (0.31.2 regression)

#3450

· what-name opened

on Apr 25, 2026

init(backend="any") selects ring singleton instead of JACCL when launched via mlx.launch --backend jaccl

#3442

· machiabeli opened

on Apr 22, 2026

[BUG] mlx-cpu backend has a hard dependncy on a deprecated version of gcc

#3430

· Evanev7 opened

on Apr 20, 2026

FSDP saving state improvement

#3401

· nastya236 opened

on Apr 12, 2026

SIGSEGV / hang in mx.fast.scaled_dot_product_attention under vmap with GQA/MQA shapes (n_heads != n_kv_heads)

#3383

· ishaansingh22 opened

on Apr 7, 2026

[BUG] mx.fast.metal_kernel: use-after-free when multiple custom kernels compose in lazy graph

#3347

· Ziqiao-git opened

on Mar 31, 2026

Add support for computing matrix determinants

#3335

· jessegrabowski opened

on Mar 30, 2026

Kernels for scipy.special statistical functions

#3330

· jessegrabowski opened

on Mar 29, 2026

SIGSEGV in Compiled::eval_gpu when auto-fused kernel receives lazy/mmap'd array with null MTLBuffer

#3329

· FHult opened

on Mar 29, 2026