Skip to content

Pull requests: ml-explore/mlx

New pull request New
67 Open 2,484 Closed
67 Open 2,484 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

Reject a negative max_norm in clip_grad_norm
#4058 opened Aug 7, 2026 by ayaangazali Contributor Loading…
4 tasks done
Build extensions with the interpreter running the build
#4057 opened Aug 7, 2026 by ayaangazali Contributor Loading…
3 of 4 tasks
Use the full row-contiguity check for GatherQMM's quantized inputs
#4056 opened Aug 7, 2026 by dudududukim Loading…
3 of 4 tasks
1
Fix hadamard_transform on the GPU for n = m with no power-of-2 factor
#4054 opened Aug 7, 2026 by axiom-of-choice Loading…
1
4
Fix Cholesky crash when empty matrix (#4032)
#4053 opened Aug 7, 2026 by FeliGame Loading…
4 tasks done
Request MSL 4.1 from the runtime compiler on macOS 27
#4052 opened Aug 7, 2026 by magnuslundstedt Loading…
Fix and enable non-transposed NAX qmm
#4051 opened Aug 7, 2026 by gordofreemo Loading…
4 tasks done
Add graph-based GPU Timer API
#4050 opened Aug 7, 2026 by XXXXRT666 Contributor Loading…
4 tasks done
Remove the per-op completion handler in gpu::eval
#4048 opened Aug 7, 2026 by kernelpool Contributor Loading…
3 of 4 tasks
Fix: propagate NaN when dealing with cummax and curmin
#4047 opened Aug 7, 2026 by JasonHonKL Contributor Loading…
4 tasks done
1
1
Reject negative dimensions in broadcast_to and random shapes
#4046 opened Aug 7, 2026 by ayaangazali Contributor Loading…
4 tasks done
1
Report bad host arguments as CLI errors instead of tracebacks
#4045 opened Aug 7, 2026 by ayaangazali Contributor Loading…
3 of 4 tasks
Fix concurrent Metal kernel cache lookup
#4043 opened Aug 7, 2026 by kitty-eu-org Loading…
Add mx.searchsorted with CPU, Metal and CUDA kernels await verification This pull request is non-trivial and requires a human expert to verify its correctness.
#4035 opened Aug 6, 2026 by erwinzhang7 Loading…
1
2
Fix empty matrix case in cholesky
#4033 opened Aug 6, 2026 by prady0t Loading…
4 tasks
1
2
Pick BM from rows per expert in gather_qmm_rhs_nax await verification This pull request is non-trivial and requires a human expert to verify its correctness.
#4023 opened Aug 5, 2026 by dwijenpatel Loading…
Adding metal kernels for the gated delta nets.
#4020 opened Aug 5, 2026 by tpegolotti Collaborator Loading…
docs: add a usage page on growing arrays
#4019 opened Aug 5, 2026 by apocryphx Contributor Loading…
3
Fix Metal FFT for sizes above 2**20
#4013 opened Aug 5, 2026 by adityasingh2400 Contributor Loading…
1
3
Keep randint samples inside [low, high) await verification This pull request is non-trivial and requires a human expert to verify its correctness.
#4012 opened Aug 5, 2026 by adityasingh2400 Contributor Loading…
1
1
Add Metal extension build support
#4004 opened Aug 4, 2026 by XXXXRT666 Contributor Loading…
4 tasks done
1
Increase ccache size in CI
#3999 opened Aug 4, 2026 by zcbenz Member Loading…
[CUDA] Fix finding cuda 13 headers in JIT compilation
#3995 opened Aug 4, 2026 by zcbenz Member Loading…
1
Enable half-precision complex Metal FFT kernels
#3981 opened Aug 2, 2026 by PhysicistJohn Contributor Loading…
Optimize large NVFP4 QMV on M5 Max await verification This pull request is non-trivial and requires a human expert to verify its correctness.
#3961 opened Aug 1, 2026 by dhiltgen Contributor Loading…
3 of 4 tasks
2
ProTip! Filter pull requests by the default branch with base:main.