mlx - Browse /v0.32.3 at SourceForge.net

The interactive file manager requires Javascript. Please enable it or use sftp or scp.
You may still browse the files here.

Name	Modified	Size	InfoDownloads / Week
Parent folder
README.md	2026-09-28	11.5 kB	0
v0.32.3 source code.tar.gz	2026-09-28	4.5 MB	1
v0.32.3 source code.zip	2026-09-28	4.9 MB	0
Totals: 3 Items		9.4 MB	1

What's Changed

Fix scan and sort for a zero-size axis by @eyupcanakman in https://github.com/ml-explore/mlx/pull/4340
Add a correction parameter in std and var by @prady0t in https://github.com/ml-explore/mlx/pull/4348
Fix tests/run.py stuck in macOS CI by @zcbenz in https://github.com/ml-explore/mlx/pull/4410
Extension Example Updated by @jagrit06 in https://github.com/ml-explore/mlx/pull/4412
Fix sorted gather_qmm NAX row overflow above 32K by @PhilipJohnBasile in https://github.com/ml-explore/mlx/pull/3922
Fix deadlock caused by mx.clear_streams() holding GIL by @aleroot in https://github.com/ml-explore/mlx/pull/4413
Fix integer pow zeroing a whole SIMD vector on a negative exponent by @ayaangazali in https://github.com/ml-explore/mlx/pull/4354
Make concurrency cap on load adaptative so that I/O scales with the machine by @aleroot in https://github.com/ml-explore/mlx/pull/4408
Fix pad with an axes subset and negative axes by @kapellirohith in https://github.com/ml-explore/mlx/pull/4364
Use sdpa_vector_2pass_1_gqa for GQA size 12 and 16 by @dudududukim in https://github.com/ml-explore/mlx/pull/4380
python: Make axis of put_along_axis optional by @ayaangazali in https://github.com/ml-explore/mlx/pull/4360
python: Make axis default to -1 in take_along_axis by @aaishwarymishra in https://github.com/ml-explore/mlx/pull/4368
Propagate NaN in arg reductions by @atirna in https://github.com/ml-explore/mlx/pull/4291
Fix mx.from_fp8 E4M3FN NaN decode with branchless carry by @saud5150 in https://github.com/ml-explore/mlx/pull/4376
Update rules for PR limitation bypass list by @zcbenz in https://github.com/ml-explore/mlx/pull/4415
[CUDA] Use native events for GPU fence waits by @strayberry in https://github.com/ml-explore/mlx/pull/4401
Validate GGUF tensor dimensions by @roshaninfordham in https://github.com/ml-explore/mlx/pull/4378
Fix crash on ellipsis indexing with too many trailing indices by @Adityaj0 in https://github.com/ml-explore/mlx/pull/4396
[CUDA] Cholesky via cuSOLVER by @sashko-zakharchuk in https://github.com/ml-explore/mlx/pull/4208
python: Fix setitem with negative index preceded by None by @Adityaj0 in https://github.com/ml-explore/mlx/pull/4414
Update nanobind to 3.0.1 by @XXXXRT666 in https://github.com/ml-explore/mlx/pull/4417
chore: Use normalize_axis_index in scan ops by @Adityaj0 in https://github.com/ml-explore/mlx/pull/4383
python: Add explicit index and bytes support by @aaishwarymishra in https://github.com/ml-explore/mlx/pull/4388
docs: fix typo recived -> received by @vaibhav8a in https://github.com/ml-explore/mlx/pull/4424
Fix Metal grid sizing for strided scans by @TheDarkchip in https://github.com/ml-explore/mlx/pull/4420
Fix SGD and Adafactor weight decay mutating caller arrays by @2sumtech in https://github.com/ml-explore/mlx/pull/4426
Widen float16/bfloat16 to float32 in cpu reduce by @Ved235 in https://github.com/ml-explore/mlx/pull/4387
sdpa_vector_2pass_1_gqa kernel batch offset for K/V in gqa decode by @dudududukim in https://github.com/ml-explore/mlx/pull/4431
Make head-dim-256 prefill with array mask run fused NAX kernel by @dwijenpatel in https://github.com/ml-explore/mlx/pull/4416
python: Make unstack return tuple instead of list by @JasonHonKL in https://github.com/ml-explore/mlx/pull/4448
python: Make iter() throw TypeError for 0-dim array by @simeetnayan81 in https://github.com/ml-explore/mlx/pull/4425
Fix save and save_safetensors corrupting a lazily loaded source file by @Cdimoy in https://github.com/ml-explore/mlx/pull/4434
[Metal] Use hypot for complex ops by @JasonHonKL in https://github.com/ml-explore/mlx/pull/4345
Add get_array_buffer_size to query buffer size for arrays by @aleroot in https://github.com/ml-explore/mlx/pull/4436
Add M5 ultra tunings for non-quantized matmuls by @jagrit06 in https://github.com/ml-explore/mlx/pull/4447
[CUDA] Fix completion worker busy loop by @strayberry in https://github.com/ml-explore/mlx/pull/4452
Fix non transposed affine qmm dispatch logic by @RohanGautam in https://github.com/ml-explore/mlx/pull/4392
Fix sorted gather_qmm on ragged K by @erwinzhang7 in https://github.com/ml-explore/mlx/pull/4009
Don't restore a thread-affine stream when StreamContext dies on another thread by @michalk8 in https://github.com/ml-explore/mlx/pull/4462
Fix Pad vjp with an axes subset and negative axes by @nileshpatil6 in https://github.com/ml-explore/mlx/pull/4441
Fix integer div-by-zero crash in CPU backend (sort, scan, binary ops) by @MarcosAsh in https://github.com/ml-explore/mlx/pull/4442
Use 64-bit file seeks on Windows by @dhiltgen in https://github.com/ml-explore/mlx/pull/4456
Use precise::exp in Sigmoid so compiled and eager sigmoid agree by @pierre427 in https://github.com/ml-explore/mlx/pull/4461
Use NAX attention for long unmasked D72/D80 inputs by @wyanzhao in https://github.com/ml-explore/mlx/pull/4455
Add D512 support to Metal vector attention by @wyanzhao in https://github.com/ml-explore/mlx/pull/4459
Add a scoped_env utility for tests by @zcbenz in https://github.com/ml-explore/mlx/pull/4474
[Metal] global scale for qmm by @nastya236 in https://github.com/ml-explore/mlx/pull/4458
Fix re-entering the same stream context manager by @michalk8 in https://github.com/ml-explore/mlx/pull/4478
leak global CommandEncoder -- avoid cuda synchronize on process shutdown by @davidkoski in https://github.com/ml-explore/mlx/pull/4480
Update CONTRIBUTING.md by @zcbenz in https://github.com/ml-explore/mlx/pull/4469
python: fix bytes() on non-contiguous arrays by @axiom-of-choice in https://github.com/ml-explore/mlx/pull/4449
Break the sibling cycle when an array is released by assignment by @tudalex in https://github.com/ml-explore/mlx/pull/4453
Win: Use DXGI for accurate WDDM VRAM memory budget by @dhiltgen in https://github.com/ml-explore/mlx/pull/4457
Add type hints and docstrings for ALiBi layer by @Ritabanm in https://github.com/ml-explore/mlx/pull/4472
Fix fp quantized matmul corruption when the quantized dim is not a multiple of 32 by @kapellirohith in https://github.com/ml-explore/mlx/pull/3912
Fix CUDA test synchronization flake by @dhiltgen in https://github.com/ml-explore/mlx/pull/4490
Skip bypass list updates in forks by @XXXXRT666 in https://github.com/ml-explore/mlx/pull/4495
Load global scales in qmm_t kernels by @dhiltgen in https://github.com/ml-explore/mlx/pull/4483
Release the GIL when copying Metal DLPack inputs (mx - pytorch mps bridge) by @WindChimeRan in https://github.com/ml-explore/mlx/pull/4497
Use matrix kernels for global-scale gather_qqmm by @dhiltgen in https://github.com/ml-explore/mlx/pull/4481
Adding metal kernels for the gated delta nets. by @tpegolotti in https://github.com/ml-explore/mlx/pull/4020
Fix jaccl ring all_gather: direction 1 slice was not mirrored by @Drifter4242 in https://github.com/ml-explore/mlx/pull/4443
Skip Metal-only gated delta kernel tests on other backends(fix CI) by @aleroot in https://github.com/ml-explore/mlx/pull/4522
[CUDA] Add global scale support to gather_qmm by @dhiltgen in https://github.com/ml-explore/mlx/pull/4507
[BUG][Metal] Deadlock in fence by @nastya236 in https://github.com/ml-explore/mlx/pull/4552
Fix ordering of complex64 with NaNs by @louen in https://github.com/ml-explore/mlx/pull/4519
Improve metal memory usage for SDPA D256 by @dhiltgen in https://github.com/ml-explore/mlx/pull/4505
Add support for matrix transpose by @aaishwarymishra in https://github.com/ml-explore/mlx/pull/4402
Improve metal memory usage for SDPA D512 by @dhiltgen in https://github.com/ml-explore/mlx/pull/4487
Support shapeless compilation of scan operations by @keeeeenw in https://github.com/ml-explore/mlx/pull/4510
[Metal] Gather mm improvement by @nastya236 in https://github.com/ml-explore/mlx/pull/4567
restoring perf regression for upsample linear mode align_corners=False by @sp4s-s in https://github.com/ml-explore/mlx/pull/4500
Add fused Metal kernels for fast.cross_entropy by @zsun6 in https://github.com/ml-explore/mlx/pull/4520
Add missing defaults to tri, tril, triu, gather_mm signatures by @rishabhsai in https://github.com/ml-explore/mlx/pull/4492
Add Metal SDPA support for D96/V64 by @wyanzhao in https://github.com/ml-explore/mlx/pull/4499
Raise IndexError for out of bounds axes by @devangpratap in https://github.com/ml-explore/mlx/pull/4484
Fix floor_divide for integers by @aaishwarymishra in https://github.com/ml-explore/mlx/pull/4515
Fix cpu binary ops on large data by @Prudctual in https://github.com/ml-explore/mlx/pull/4517
Do not throw from CUDA destructors and avoid implicit default streams in eval/compile by @aleroot in https://github.com/ml-explore/mlx/pull/4514
[Metal] gather_qmm improvement by @nastya236 in https://github.com/ml-explore/mlx/pull/4572
Fix floating-point constant precision in compiled kernels by @Ryan11c in https://github.com/ml-explore/mlx/pull/4511
chore: Fix assertion in ReduceScatter::eval_cpu by @ronaldmannak in https://github.com/ml-explore/mlx/pull/4557
Leak thread local streams on exit in main thread by @zcbenz in https://github.com/ml-explore/mlx/pull/4576