All issues

Package.swift exclude list breaks when the mlx pin moves past v0.31.2 (jaccl restructured)

#446

· GoodOlClint opened

on Jul 30, 2026

Please bump the vendored mlx-core so the batched single-token RoPE fix (ml-explore/mlx#3498) reaches mlx-swift

#441

· magicnight opened

on Jul 8, 2026

[BUG] Regression in 0.31.4: holding evalLock across tostring (#410) deadlocks/crashes when an MLXArray is stringified during an in-flight eval()

#428

· stovak opened

on Jun 21, 2026

[BUG] `conv1d` produces wrong results for long sequences on the iOS GPU (output > 32768 frames: early frames collapse to a constant)

#424

· brannala opened

on Jun 17, 2026

[BUG] quantizedMM silently upcasts to float32 due to dtype mismatch between activations and scales/biases

#420

· czkoko opened

on Jun 7, 2026

Discussion: Sendable MLXArray

#417

· davidkoski opened

on May 29, 2026

Allow overriding the mlx.metallib search path for plugin-bundle embedding

#415

· olilarkin opened

on May 28, 2026

[Bug] check_error() throws std::runtime_error from completion handler → uncaught SIGABRT on iPhone 15 Pro / iOS 26.4 (Qwen3 1.7B 4bit)

#407

· kitadesign opened

on May 15, 2026

Feature request: prebuilt binary xcframework for Cmlx to eliminate cold build cost

#406

· sam-fakhreddine opened

on May 14, 2026

[BUG] many mlx-swift functions incorrectly promote to float32, e.g. zeros, ones, etc.

#390

· davidkoski opened

on Apr 5, 2026

[BUG] Add Gemma 4 model architecture support (gemma4)

#389

· sftinc opened

on Apr 5, 2026

[BUG] better compile support

#386

· davidkoski opened

on Apr 3, 2026