Skip to content

Pull requests: huggingface/trl

New pull request New
107 Open 2,853 Closed
107 Open 2,853 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

Pass tools as None to apply_chat_template when it is an empty list
#5380 opened Mar 26, 2026 by rabinadk1 Loading…
2 of 5 tasks
1
6
Use datasets Json dtype to prevent insertion of None values
#5376 opened Mar 26, 2026 by albertvillanova Loading…
12
Remove truncation_mode from DPO
#5372 opened Mar 25, 2026 by albertvillanova Loading…
7
[vllm-serve] Add extra_llm_kwargs for passing additional arguments to vllm.LLM()
#5367 opened Mar 25, 2026 by jonahsamost Loading…
1 of 5 tasks
6
Remove deprecated TRACKIO_SPACE_ID env var from all scripts
#5365 opened Mar 24, 2026 by sergiopaniego Loading…
5 tasks
1
add more generaic device suppport for CI tests
#5357 opened Mar 24, 2026 by kaixuanliu Loading…
Enforce PR template for first-time contributors and document AI usage policy
#5356 opened Mar 24, 2026 by qgallouedec Loading…
3 of 8 tasks
10
Add chunked LM head for memory-efficient log-prob computation for AsyncGRPOTrainer
#5349 opened Mar 23, 2026 by AmineDiro Loading…
24
[Test] Fix *test_training_vlm_multi_image* by skipping vision params in assertion
#5341 opened Mar 22, 2026 by YangKai0616 Loading…
5
Fix Liger kernel crash with device_map="auto" on multi-GPU in GRPOTrainer
#5340 opened Mar 22, 2026 by YangKai0616 Loading…
7
Enable Tensor Parallelism in SFT script
#5331 opened Mar 21, 2026 by songhappy Loading…
2
Support multimodal tool responses in environment_factory for VLM training
#5323 opened Mar 20, 2026 by sergiopaniego Loading…
5 tasks
20
(5/5) async grpo metrics
#5322 opened Mar 20, 2026 by AmineDiro Loading…
6
(4/5) async grpo break out of generation loop (is_done)
#5321 opened Mar 20, 2026 by AmineDiro Loading…
6
(3/5) Cancel Stale inflight tasks
#5320 opened Mar 20, 2026 by AmineDiro Loading…
10
(1/5) Add callback to sync weights before training begins
#5319 opened Mar 20, 2026 by AmineDiro Loading…
4
(2/5) Refactor RolloutCompletion in Async Rollout Worker
#5318 opened Mar 20, 2026 by AmineDiro Loading…
4
Expand the list of attention implementations compatible with packing
#5316 opened Mar 19, 2026 by mariosasko Loading…
1 of 5 tasks
1
fix: fix a bug in vLLM weight synchronization when vllm_enable_sleep_mode=True 🩹 for patch
#5313 opened Mar 19, 2026 by muupan Loading…
1 of 5 tasks
1
6
Show conversations instead of decoded text in the completions table
#5309 opened Mar 19, 2026 by qgallouedec Loading…
10
Add support for logging extra columns in reward functions and update related tests
#5308 opened Mar 19, 2026 by qgallouedec Loading…
4
fix: skip ref adapter when peft config uses target_parameters
#5292 opened Mar 16, 2026 by gambletan Loading…
3 tasks
1
2
docs: clarify PPO entropy metrics in PPO trainer docs
#5289 opened Mar 14, 2026 by biefan Loading…
1
2
Add Nemotron 3 to tests via tiny model
#5278 opened Mar 12, 2026 by sergiopaniego Loading…
5 tasks
11
Introduce backend rollout-completions interface and decouple OpenEnv helper from vLLM internals
#5256 opened Mar 10, 2026 by rycerzes Loading…
1
2
ProTip! no:milestone will show everything without a milestone.