Commits

Breadcrumbs

History for

onmain

User selector

All users

All time

Commit History

Commits on Jan 26, 2026

• Fix flaky macOS test (#3063)

  

  Awni Hannun

  authored
  5bd99dd

  Copy full SHA for 5bd99dd

Commits on Jan 23, 2026

• Build and test python package on Windows CI (#3049)

  zcbenz

  authored
  fd27829

  Copy full SHA for fd27829

• Allow take on empty array when it makes sense (#3046)

  

  Awni Hannun

  authored
  9bac6f8

  Copy full SHA for 9bac6f8

Commits on Jan 21, 2026

• Use higher precision for linspace with double (#3029)

  

  Awni Hannun

  authored
  1d56dfd

  Copy full SHA for 1d56dfd

Commits on Jan 19, 2026

• Type Enhancement for Func Transforms and Bug Fix (#3003)

  Show description for 46d0fdc
  

  XXXXRT666

  and

  Awni Hannun

  authored
  46d0fdc

  Copy full SHA for 46d0fdc

• Fix python package install path in stubgen (#3009)

  zcbenz

  authored
  d96a2bd

  Copy full SHA for d96a2bd

Commits on Jan 18, 2026

• Fix sharding of quantized models with non-power-of-2 bits (#3006)

  kernelpool

  authored
  ca14d3d

  Copy full SHA for ca14d3d

Commits on Jan 17, 2026

• fix distributed all_to_sharded bias shard axis from -2 to -1 (#2987)

  gufengc

  authored
  d2bef3c

  Copy full SHA for d2bef3c

Commits on Jan 14, 2026

• Reverts changing the MLX_IBV_DEVICES to MLX_JACCL_DEVICES (#2999)

  angeloskath

  authored
  3fe7794

  Copy full SHA for 3fe7794

• patch bump for next release (#2991)

  

  Awni Hannun

  authored
  2469fc2

  Copy full SHA for 2469fc2

Commits on Jan 13, 2026

• Expose to/from fp8 in Python and don't auto-convert fp8 when loading from safetensors (#2985)

  

  Awni Hannun

  authored
  099dcc0

  Copy full SHA for 099dcc0

• Fix RandomBits::is_equivalent to include width (#2978)

  Show description for 4160ec1
  

  4 peopleauthored
  4160ec1

  Copy full SHA for 4160ec1

• replace MLX_IBV_COORDINATOR with MLX_JACCL_COORDINATOR (#2986)

  Evanev7

  authored
  a819779

  Copy full SHA for a819779

Commits on Jan 12, 2026

• fix doc (#2988)

  CC-Yeh

  authored
  7b1c469

  Copy full SHA for 7b1c469

• Add asarray to __array_namespace__ (#2966)

  Anri-Lombard

  authored
  edab937

  Copy full SHA for edab937

Commits on Jan 5, 2026

• QQ linear (#2931)

  nastya236

  authored
  5037317

  Copy full SHA for 5037317

• fix numpy dtype bug (#2960)

  

  Awni Hannun

  authored
  1680b6f

  Copy full SHA for 1680b6f

• Fix doc issues in `mlx.nn.init.he_normal` and `mlx.nn.hard_tanh` (#2968)

  Show description for 1df6c2a
  

  Redempt1onzzZZ

  and

  Awni Hannun

  authored
  1df6c2a

  Copy full SHA for 1df6c2a

Commits on Dec 24, 2025

• Fetch nanobind with cmake (#2949)

  zcbenz

  authored
  26dfe4f

  Copy full SHA for 26dfe4f

• [CUDA] Implement gather_mm_rhs (#2902)

  zcbenz

  authored
  1d21d0e

  Copy full SHA for 1d21d0e

Commits on Dec 23, 2025

• Metal/CPU nvfp4 and mxfp8 (#2946)

  

  Awni Hannun

  authored
  1eef1d1

  Copy full SHA for 1eef1d1

• Fixes in mlx.distributed_config (#2947)

  angeloskath

  authored
  9cfda1a

  Copy full SHA for 9cfda1a

Commits on Dec 21, 2025

• Fix for non row-contig scales (#2941)

  

  Awni Hannun

  authored
  c0fea26

  Copy full SHA for c0fea26

Commits on Dec 19, 2025

• Fix pid in local launch (#2936)

  angeloskath

  authored
  d9f4d8d

  Copy full SHA for d9f4d8d

• Set install rpath of python bindings with cmake (#2934)

  zcbenz

  authored
  fc19a08

  Copy full SHA for fc19a08

• Set rpath with cmake for CUDA build (#2932)

  zcbenz

  authored
  b2e2b19

  Copy full SHA for b2e2b19

Commits on Dec 17, 2025

• Thunderbolt RDMA communications backend (#2808)

  angeloskath

  authored
  9194ec2

  Copy full SHA for 9194ec2

Commits on Dec 16, 2025

• qqmm (#2789)

  Show description for 4cf5b29
  

  4 peopleauthored
  4cf5b29

  Copy full SHA for 4cf5b29

Commits on Dec 11, 2025

• Fix grad in place updates (#2899)

  

  Awni Hannun

  authored
  81dfe5f

  Copy full SHA for 81dfe5f

• Update nanobind pin to most recent version (#2896)

  ngoldbaum

  authored
  e1fee00

  Copy full SHA for e1fee00

• Fix input buffer donation in compile (#2897)

  CC-Yeh

  authored
  3c8ce9b

  Copy full SHA for 3c8ce9b

Commits on Dec 4, 2025

• Add a 2-pass col reduce for CUDA (#2863)

  angeloskath

  authored
  997cfc7

  Copy full SHA for 997cfc7

• Do a PyPi release for cuda on arm (#2866)

  

  Awni Hannun

  authored
  1fa8dc5

  Copy full SHA for 1fa8dc5

• fix compile copying (#2871)

  

  Awni Hannun

  authored
  a6d6717

  Copy full SHA for a6d6717

• Layer norm throws on dimension mismatch (#2870)

  

  Awni Hannun

  authored
  941cfe2

  Copy full SHA for 941cfe2