Skip to content

Pull requests: unslothai/unsloth-zoo

New pull request New
33 Open 298 Closed
33 Open 298 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

update compiler for XLMRobertaModel
#383 opened Dec 13, 2025 by electroglyph Loading…
4
[FIX] Qwen3 moe torch compile issue
#381 opened Dec 12, 2025 by Datta0 • Draft
1
FP8: Load model on-the-fly in vLLM
#380 opened Dec 11, 2025 by andrewor14 Loading…
9
resolve triton windows for intel
#378 opened Dec 11, 2025 by leizhenyuan Loading…
4
FIX: Ensure modules_to_save are merged
#369 opened Dec 2, 2025 by oKatanaaa Loading…
1
8
Rework vLLM memory requirement calculations
#366 opened Dec 1, 2025 by Datta0 Loading…
9
Add flashinfer toggle for vllm backend
#362 opened Nov 25, 2025 by noisycat3 Loading…
4
Chunk Across Batch and Context length for logprob calculations for grpo
#357 opened Nov 21, 2025 by pluesclues Loading…
33
Filelocks
#325 opened Oct 17, 2025 by mmathew23 • Draft
Update compiler.py: resolve AttributeError when loading Qwen3 models with LoRA adapters
#305 opened Sep 26, 2025 by MengAiDev Loading…
1
fix Gemma3Attention.forward patch fail on transformers versions > 4.55.4
#281 opened Sep 15, 2025 by rolandtannous Loading…
fix vram estimation when loading a 16bit model with load_in_4bit=True
#193 opened Jul 8, 2025 by yizhe-skyfall Loading…
1
vlm grpo UnslothEfficientGRPO
#188 opened Jul 2, 2025 by GAD-cell Loading…
1
Change UnslothEfficientGRPO to accept log probs instead of embeddings
#172 opened Jun 19, 2025 by zkpranav Loading…
24
Debug: num_items_in_batch on a different device from loss.
#147 opened May 23, 2025 by KimRass Loading…
2
Convert n_items type from torch.Tensor to int
#139 opened May 15, 2025 by MoonRainy21 Loading…
14
Fix LoRA hot swapping
#116 opened Apr 17, 2025 by pbarker Loading…
2
fix double bos
#106 opened Mar 28, 2025 by tamewild Loading…
2
num_proc=os.cpu_count() when counting unused tokens
#104 opened Mar 27, 2025 by IsaacBreen Loading…
5
saving utils check both model.base_model and base_model.model
#95 opened Mar 22, 2025 by toranb Loading…
VLM Data Collator - Make text & image mixing work efficiently
#89 opened Mar 21, 2025 by mmathew23 Loading…
5
Add support for multiple instruction patterns in train_on_responses_only
#75 opened Mar 14, 2025 by marcelodiaz558 Loading…
2
Fix Protobuf<4.0 is deprecated, relax dependency
#68 opened Mar 10, 2025 by hoh Loading…
4
Fix VLM train_on_response_only
#60 opened Mar 1, 2025 by coding-famer Loading…
1
37
Fixed getting supported LoRA modules from vLLM.
#58 opened Feb 27, 2025 by LagPixelLOL Loading…
ProTip! Mix and match filters to narrow down what you’re looking for.