Search results

Upstreaming improvements from fast-moving forks

Status: Open.

#221 In ml-explore/mlx-swift-lm;

· ronaldmannak opened on Apr 17, 2026

TokenRing.loadPrompt crash: prompt.dim(0) returns batch dim for 2D tensors

Status: Open.

#220 In ml-explore/mlx-swift-lm;

· VincentGourbin opened on Apr 17, 2026

[BUG] Gemma 4 model execution failed

Status: Open.

#219 In ml-explore/mlx-swift-lm;

· MuMuY66 opened on Apr 17, 2026

[BUG] 3.31.3 release's upgrade notes are 404'd

Status: Open.

#217 In ml-explore/mlx-swift-lm;

· tgaff-mfw opened on Apr 17, 2026

RFC: response-format parsing for Harmony / channel-based models doesn't fit the current Generation enum

Status: Open.

#214 In ml-explore/mlx-swift-lm;

· ronaldmannak opened on Apr 14, 2026

SmolVLM2: bestFit() upscales small images, creating excessive tiles and degrading performance by 9x

Status: Open.

#208 In ml-explore/mlx-swift-lm;

· VladimirTalyzin opened on Apr 13, 2026

Feature Request: Add Support for Multimodal Gemma 4 Audio Encoder & Preprocessing

Status: Open.

#207 In ml-explore/mlx-swift-lm;

· ghchinoy opened on Apr 12, 2026

[BUG] Gemma 4 model support — community Swift port available

Status: Open.

#198 In ml-explore/mlx-swift-lm;

· yejingyang8963-byte opened on Apr 9, 2026

Gemma 4 model support — community Swift port available

Status: Open.

#197 In ml-explore/mlx-swift-lm;

· yejingyang8963-byte opened on Apr 9, 2026

VLM broadcast_shapes crash with Qwen3-VL and real images

Status: Open.

#191 In ml-explore/mlx-swift-lm;

· kiki830621 opened on Apr 7, 2026

[BUG] version 3 todos

Status: Open.

#189 In ml-explore/mlx-swift-lm;

· davidkoski opened on Apr 6, 2026

[BUG] UserInput init doesn't always set self.images

Status: Open.

#182 In ml-explore/mlx-swift-lm;

· davidkoski opened on Apr 3, 2026