Skip to content

Pull requests: waybarrios/vllm-mlx

New pull request New
17 Open 411 Closed
17 Open 411 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

fix(text-model-from-vlm): realize private lazy arrays before the model leaves the build thread
#614 opened Jun 12, 2026 by ursk Contributor Loading…
1
fix(ssd-cache): preserve original bfloat16 dtype across quantized spill
#612 opened Jun 12, 2026 by CBribiescas Contributor Loading…
Keep assistant tool_calls and tool messages on the MLLM chat path
#611 opened Jun 12, 2026 by waybarrios Owner Loading…
1
1
fix(engine): stop MLLM text route at the model's full config EOS set
#610 opened Jun 11, 2026 by ursk Contributor Loading…
3
Guard --mllm against continuous batching (silent empty output)
#601 opened Jun 6, 2026 by eejd Contributor Loading…
1
fix(step): add Step3p5 parser support
#598 opened Jun 6, 2026 by Thump604 Collaborator Loading…
@janhilgard
fix(qwen3-xml): parse bare <function=> without <tool_call> wrapper
#597 opened Jun 6, 2026 by CBribiescas Contributor Loading…
feat(mllm): auto-extract audio from video_url on omni models
#591 opened Jun 3, 2026 by txdadlab Loading…
1
fix(gpt-oss): route harmony prompts through openai-harmony (refs #568)
#581 opened May 25, 2026 by CBribiescas Contributor Loading…
1
2
fix(llama-tool-parser): recognize Llama 3.1+ / 3.3 tool-call formats
#580 opened May 25, 2026 by CBribiescas Contributor Loading…
1
Add SimpleEngine prefix trie cache
#574 opened May 24, 2026 by Thump604 Collaborator Loading…
@waybarrios @janhilgard
1
fix(gpt-oss): plumb harmony tool calls all the way through to response
#562 opened May 22, 2026 by CBribiescas Contributor Loading…
3 tasks
1
Reduce EngineCore idle polling
#552 opened May 20, 2026 by Thump604 Collaborator Loading…
@waybarrios @janhilgard
1
Keep MLLM media stream on owner thread
#551 opened May 19, 2026 by Thump604 Collaborator Loading…
@waybarrios @janhilgard
1
Keep VLM TextModel generation on owner thread
#543 opened May 17, 2026 by Thump604 Collaborator Loading…
@waybarrios @janhilgard
fix: Qwen tool streaming recovery
#497 opened May 4, 2026 by kylejeske Loading…
15
perf: O(1) tool lookup in ToolExecutor via lazily-cached name index optimization
#449 opened Apr 26, 2026 by clickbrain Contributor Loading…
@waybarrios
6
ProTip! Follow long discussions with comments:>50.