Skip to content

Pull requests: hiyouga/LlamaFactory

New pull request New
82 Open 1,309 Closed
82 Open 1,309 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

V1 muon optimizer
#10618 opened Jul 1, 2026 by HelloWorldBeginner Contributor Loading…
2 tasks
3
fix(data): correct keyword argument sample_idx to batch_idx in collator
#10612 opened Jun 30, 2026 by GSCSD1 Contributor Loading…
1
[data] fix KeyError on unmapped role tag in openai converter main loop
#10606 opened Jun 24, 2026 by he-yufeng Loading…
2 tasks done
2
[data] fix KeyError on invalid role tag in ranking dataset converters
#10601 opened Jun 23, 2026 by he-yufeng Loading…
2
[data] fix MiniMax tool_extractor to return content when no tool call parses
#10599 opened Jun 23, 2026 by he-yufeng Loading…
2
refactor(v1): replace custom template system with apply_chat_template
#10598 opened Jun 22, 2026 by frozenleaves Collaborator • Draft
8
Add configurable profiler for v1 training
#10595 opened Jun 21, 2026 by addsubmuldiv Loading…
1 of 2 tasks
5
[data] fix LFM2 tool extractor dropping positional arguments
#10594 opened Jun 19, 2026 by he-yufeng Loading…
2 tasks done
2
fix(data): return content when llama3/mistral tool extractors find no calls
#10592 opened Jun 18, 2026 by he-yufeng Loading…
3
Fix HuggingfaceEngine input kwargs defaults
#10590 opened Jun 17, 2026 by lxingy3 • Draft
1
6
fix: add defensive fallback for Qwen VL mrope get_rope_index shape mismatch
#10589 opened Jun 17, 2026 by Ottohere-Mourn Loading…
3
fix(collator): pass batch index positionally in single-subseq mrope path
#10588 opened Jun 17, 2026 by he-yufeng Loading…
1
3
fix(data): wrap MiniMax-M2 tool calls in <minimax:tool_call> tags
#10586 opened Jun 16, 2026 by he-yufeng Loading…
2 tasks done
fix: place Qwen tools before system prompt
#10579 opened Jun 14, 2026 by he-yufeng Loading…
1
2
fix: merge ShareGPT tool observations
#10578 opened Jun 14, 2026 by he-yufeng Loading…
2
docs: update Chinese NPU docs link
#10577 opened Jun 14, 2026 by he-yufeng Loading…
1
fix: avoid mutable input kwargs defaults
#10576 opened Jun 14, 2026 by he-yufeng Loading…
1
3
docs: clarify reasoning-only template guidance
#10574 opened Jun 14, 2026 by he-yufeng Loading…
1
fix(data): stabilize mixed-modality SFT schema
#10571 opened Jun 13, 2026 by yangtao-hub Loading…
2 tasks done
2
10
fix(qwen): align tool-call training templates with native format
#10570 opened Jun 12, 2026 by MrBlue-1996 Loading…
2
5
[feat] add embedding init methods and freeze original embeddings
#10563 opened Jun 11, 2026 by ximinng Contributor Loading…
2 tasks done
7
[Device-Metrics]: Add support to dump device metrics in log xpu, cpu
#10554 opened Jun 8, 2026 by koppulac Loading…
2 tasks
3
[misc] improve diagnostics for missing processor on multi-modal templates
#10553 opened Jun 6, 2026 by gaurav0107 Loading…
1
5
Add LoRA fine-tuning examples for Qwen3 on Ascend
#10552 opened Jun 6, 2026 by boes129 Loading…
3
[chat] fix hf engine input kwargs defaults
#10545 opened Jun 2, 2026 by Ruazzm Loading…
2 tasks done
1
4
ProTip! no:milestone will show everything without a milestone.