Skip to content

Pull requests: ml-explore/mlx-swift-lm

New pull request New
45 Open 199 Closed
45 Open 199 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

Gemma4Text: load conversions that omit redundant KV-shared-layer K/V projections
#350 opened Jun 13, 2026 by jfranknichols Contributor Loading…
Expose ModelTypeRegistry.contains(_:) for pre-load support checks
#349 opened Jun 13, 2026 by jfranknichols Contributor Loading…
Add TranslateGemma support (Gemma 3 translation template)
#348 opened Jun 13, 2026 by beshkenadze Loading…
Add Hunyuan dense V1 (hunyuan_v1_dense): Hunyuan-MT-7B and Hy-MT2-7B
#347 opened Jun 13, 2026 by beshkenadze Loading…
Gemma 3: chunked prompt prefill, skip lm_head on prompt positions
#346 opened Jun 13, 2026 by beshkenadze Loading…
Qwen2-VL: implement M-RoPE in the language model (mirror of #239)
#345 opened Jun 13, 2026 by mnmly Loading…
fix MLXVLM prepare(): honor windowSize for chunked prefill (FastVLM / Gemma3 / Qwen2VL / LFM2VL / Pixtral / Mistral3)
#344 opened Jun 12, 2026 by john-rocky Contributor Loading…
1
Gemma 4: add audio and video multimodal support
#343 opened Jun 12, 2026 by atlascodesai Loading…
4 tasks done
Fix Gemma 4 QAT load: KV-shared layers carry no k_proj/v_proj/k_norm
#342 opened Jun 12, 2026 by Flo5k5 Loading…
1 of 4 tasks
remove references per author request
#341 opened Jun 11, 2026 by davidkoski Collaborator Loading…
3 of 4 tasks
Raise swift-syntax floor to 602.0.0 for prebuilt artifact resolution
#340 opened Jun 11, 2026 by roryford Loading…
2
Use MaterializedArray for Sendable conformance
#335 opened Jun 9, 2026 by davidkoski Collaborator Loading…
2 of 4 tasks
14
Add MLXFoundationModels: an MLX-backed FoundationModels LanguageModel
#334 opened Jun 9, 2026 by ctymoszek Loading…
11
Fix Gemma4 QAT (E-series) load: KV-shared layers have no k_proj/v_proj/k_norm
#330 opened Jun 5, 2026 by Flo5k5 Loading…
Add variance-normalized KV cache
#329 opened Jun 5, 2026 by aleroot Contributor Loading…
3 of 4 tasks
Add Gemma 4 12B unified
#327 opened Jun 3, 2026 by adrgrondin Contributor Loading…
3 tasks done
6
fix speculative decode tests
#326 opened Jun 3, 2026 by davidkoski Collaborator Loading…
3 of 4 tasks
1
1
Preserve aspect ratio in Gemma4 image preprocessing
#325 opened May 31, 2026 by Penrose0v0 Loading…
Improve Qwen3.5 recurrent cache handling
#323 opened May 30, 2026 by aleroot Contributor Loading…
4 tasks done
5
Add support for Qwen3-VL-MoE models
#322 opened May 30, 2026 by lucasnewman Loading…
4 tasks done
Make MLXLLM compilable on Linux
#321 opened May 29, 2026 by gmondada Loading…
2 of 4 tasks
1
Fix Gemma4 E-series load: make per_layer_model_projection quantizable
#320 opened May 28, 2026 by nyteshade Loading…
Add Swift model conversion API
#318 opened May 28, 2026 by aleroot Contributor Loading…
3 of 4 tasks
1
Add speculative decoding telemetry and memory gating
#314 opened May 27, 2026 by aleroot Contributor Loading…
4 tasks done
Add Nemotron Labs Diffusion
#310 opened May 22, 2026 by smdesai Contributor Loading…
2 of 4 tasks
11
ProTip! Updated in the last three days: updated:>2026-06-11.