Skip to content

Pull requests: huggingface/trl

New pull request New
77 Open 1,643 Closed
77 Open 1,643 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

[WIP] [SFT] SFT doc rewrite
#3619 opened Jun 18, 2025 by qgallouedec Loading…
5 tasks
1
ClearML logging of visualization in RewardTrainer evaluation
#3602 opened Jun 16, 2025 by ioverho Loading…
2 of 5 tasks
Fix: corrected fsdp in GRPO trainer
#3582 opened Jun 13, 2025 by tryumanshow Loading…
2 of 5 tasks
Check rewards shapes in RewardTrainer
#3577 opened Jun 13, 2025 by ioverho Loading…
4 tasks done
Chisquare regularized DPO
#3573 opened Jun 12, 2025 by asparius Loading…
3
Add entropy based filtering inside the GRPOTrainer.
#3563 opened Jun 10, 2025 by pramodith Loading…
4 of 5 tasks
1
15
🥳 new rloo
#3533 opened Jun 3, 2025 by shirinyamani Loading…
5 tasks
1
Push KTAE impl
#3518 opened May 30, 2025 by SamComber Loading…
5 tasks
1
intuit
#3513 opened May 29, 2025 by shirinyamani Loading…
5 tasks
1
🎀 New defaults: gradient_checkpointing=True
#3510 opened May 29, 2025 by qgallouedec Loading…
5 tasks
6
Add Bidirectional Knowledge Distillation Option to GKDTrainer
#3508 opened May 29, 2025 by shaischaudhry Loading…
3 of 5 tasks
@kashif
3
HF Doc Builder style
#3498 opened May 26, 2025 by qgallouedec • Draft
1
[GRPO] Adds SSR priorized replay buffer
#3496 opened May 26, 2025 by edbeeching Loading…
1
[GRPO] Pad per minibatch instead of per generation batch
#3495 opened May 26, 2025 by edbeeching • Draft
3 tasks
18
feat(grpo): validate gradient_accumulation_steps vs steps_per_generation for on-policy GRPO
#3493 opened May 25, 2025 by HarryHsing Loading…
3
[GKD] Use vllm for the student model
#3475 opened May 21, 2025 by kashif • Draft
5 tasks
4
Add support for CB with native transformers
#3471 opened May 20, 2025 by ArthurZucker Loading…
@qgallouedec
5
Allow an user to train from a local dataset
#3470 opened May 19, 2025 by gogo2464 Loading…
1 of 5 tasks
27
Add async tool-enabled vLLM server for GRPO training via OpenAI-compatible interface
#3469 opened May 19, 2025 by BjarniHaukur Loading…
5 tasks
1
1
add support for image inputs in GRPO
#3460 opened May 16, 2025 by hellopahe Loading…
1
@shirinyamani
64
[SFT] add warning if dataset's input_ids exceed max_length
#3449 opened May 15, 2025 by HERIUN Loading…
1 of 5 tasks
1
Fix logging docs
#3447 opened May 14, 2025 by xingyaoww • Draft
2 of 5 tasks
1
🛠️ quantization support for vllm generation
#3428 opened May 8, 2025 by shirinyamani Loading…
5 tasks
13
Reintroducing step method in ppo_trainer
#3410 opened May 3, 2025 by jskaf34 Loading…
2 of 5 tasks
1
@shirinyamani
4
Fix GRPO/DAPO/Dr.GRPO documentation: formula corrections and KL divergence clarification
#3395 opened Apr 30, 2025 by JenWei0312 Loading…
1 of 5 tasks
10
ProTip! Adding no:label will show everything without a label.