All issues

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

[BUG] DLPackCompatible protocol members are declared as attributes, so np.ndarray fails to match — mx.array(np_arr) no longer type-checks

#4371

· dgrauet opened

on Aug 21, 2026

[BUG] `mx.dequantize` produces wrong results on a non-contiguous slice (regression on 0.32.1)

#4370

· AirRunner opened

on Aug 21, 2026

Performance guidance + GPU extended-precision (double-double) support — benchmark data attached

#4363

· bahaehmimdi opened

on Aug 20, 2026

[BUG] `setitem` with a single Ellipsis fails

#4358

· ev-br opened

on Aug 19, 2026

Support fusing element-wise operations into reductions

#4355

· sakupan102 opened

on Aug 19, 2026

[BUG] Metal JIT build: qmv template instantiation passes 6 args to 5-param kernel templates (runtime abort)

#4350

· PowerBeef opened

on Aug 18, 2026

[BUG] `mx.abs` / `mx.sqrt` / `mx.log` on `complex64` are silently wrong on GPU across a third of float32's range

#4344

· JasonHonKL opened

on Aug 18, 2026

[Performance] Avoid dense zero work in stride-2 ConvTranspose3d on Metal

#4341

· ternaus opened

on Aug 18, 2026

[CUDA] gather_qmm ~24x slower than equal-FLOPs quantized_matmul on GB10 (sm_121) — makes quantized MoE prefill unusable

#4339

· NORDIT opened

on Aug 18, 2026

[BUG] CPU errors are not propagated through async_eval and abort the process

#4329

· ayaangazali opened

on Aug 17, 2026

[BUG][CUDA] CommandEncoder remains corrupted after graph-cache thrashing exception

#4326

· strayberry opened

on Aug 17, 2026

Rename `ddof` parameter in `std` and `var` to `correction`?

#4325

· prady0t opened

on Aug 17, 2026