Skip to content

Pull requests: ml-explore/mlx

New pull request New
37 Open 2,347 Closed
37 Open 2,347 Closed
Author
Filter by author
Loading
Label
Filter by label
Loading
Use alt + click/return to exclude labels
or ⇧ + click/return for logical OR
Projects
Filter by project
Loading
Milestones
Filter by milestone
Loading
Reviews
Filter by reviews
No reviews Review required Approved review Changes requested
Assignee
Filter by who’s assigned
Assigned to nobody Loading
Sort
Sort by
Newest Oldest Most commented Least commented Recently updated Least recently updated Best match
Most reactions
👍 👎 😄 🎉 😕 ❤️ 🚀 👀

Pull requests list

Fix int64 type cast error when loading GGUF metadata arrays
#3823 opened Jul 9, 2026 by danlee2002 Contributor Loading…
4 tasks done
3
Make index autodiff errors explicitly recommend stop_gradient
#3820 opened Jul 8, 2026 by dogukanveziroglu Contributor Loading…
5
Document default values in normalization layer docstrings
#3819 opened Jul 8, 2026 by Pablosinyores Contributor Loading…
1
Refactor wheel building script
#3818 opened Jul 8, 2026 by zcbenz Collaborator Loading…
Fix BatchNorm running variance estimator
#3817 opened Jul 8, 2026 by ishtihoss Contributor Loading…
3
Fix shapeless matmul with dynamic batch dimensions
#3813 opened Jul 7, 2026 by varshneydevansh Contributor Loading…
3 of 4 tasks
1
3
Add builds for free-threaded python
#3812 opened Jul 7, 2026 by zcbenz Collaborator Loading…
1
[CUDA][Improvement] Fast path for gather
#3794 opened Jul 3, 2026 by nastya236 Collaborator Loading…
3
Raise qmv batch limit for large matrices on M5-class GPUs
#3791 opened Jul 2, 2026 by pierre427 Loading…
4
Decompose small kernel-depth 3D convs into 2D convs
#3785 opened Jul 1, 2026 by katlun-lgtm Contributor Loading…
1
5
[WIP] [CUDA] fsdp
#3768 opened Jun 25, 2026 by nastya236 Collaborator • Draft
Add gather_qqmm
#3757 opened Jun 24, 2026 by zcbenz Collaborator Loading…
Propagate CPU errors to events
#3742 opened Jun 22, 2026 by zcbenz Collaborator Loading…
8
Add antialias support to nn.Upsample for linear and cubic modes
#3677 opened Jun 14, 2026 by lyonsno Contributor Loading…
Fix state corruption when a primitive throws during eval
#3675 opened Jun 13, 2026 by WindChimeRan Loading…
4 tasks done
2
Fused SDPA full-attention for head_dim=192/256 (revival of #3293)
#3660 opened Jun 11, 2026 by hojin12312 Loading…
3 of 4 tasks
1
Add stft and istft to mlx.core.fft
#3639 opened Jun 8, 2026 by eyupcanakman Contributor Loading…
4 tasks done
6
Add reflect and symmetric padding modes to mx.pad
#3608 opened May 30, 2026 by katlun-lgtm Contributor Loading…
4
Implement batched matmul for large 1D dot products
#3580 opened May 22, 2026 by Ved235 Loading…
2 of 4 tasks
24
feat(fast): add sliding-window SDPA kernel path
#3552 opened May 15, 2026 by rabbitson87 Loading…
9
Fix JACCL GID selection on Apple Thunderbolt RDMA (errno 22 RTR fail)
#3468 opened Apr 29, 2026 by danielkristofik Loading…
3 of 4 tasks
1
Fix SDPA vmap with GQA/MQA shapes (n_heads != n_kv_heads)
#3385 opened Apr 8, 2026 by Brooooooklyn Contributor Loading…
4 tasks done
1
[Metal] Fused Flash Attention backward (VJP) kernels
#3241 opened Mar 11, 2026 by Brooooooklyn Contributor Loading…
1
Add bias support to QQLinear
#3215 opened Mar 6, 2026 by mdepree Loading…
4 tasks done
7
Add all_to_all collective primitive
#3164 opened Feb 24, 2026 by 0xDaizz Loading…
2 of 4 tasks
1
ProTip! Adding no:label will show everything without a label.