Skip to content

Pull requests: huggingface/trl

New pull request New
101 Open 2,787 Closed
101 Open 2,787 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

Remove custom get_train/eval_dataloader from OnlineDPO
#5291 opened Mar 16, 2026 by albertvillanova Loading…
1
Remove TrainingArguments import from experimental trainers
#5290 opened Mar 16, 2026 by albertvillanova Loading…
1
docs: clarify PPO entropy metrics in PPO trainer docs
#5289 opened Mar 14, 2026 by biefan Loading…
1
perf(trl): set a timeout on vllm client HTTP calls
#5288 opened Mar 14, 2026 by tejasae-afk Loading…
Add reference to DeepSeekMath in accuracy_reward docstring
#5287 opened Mar 13, 2026 by qgallouedec Loading…
5 tasks
2
Prevent corruption of DPO VLM training if "keep_end" truncation_mode
#5286 opened Mar 13, 2026 by albertvillanova Loading…
1
3
Support max_length in DPO VLM training
#5284 opened Mar 13, 2026 by albertvillanova Loading…
1
6
Fix accuracy_reward crash when called from non-main thread
#5281 opened Mar 13, 2026 by qgallouedec Loading…
4
Add Cursor rules from AGENTS.md
#5280 opened Mar 12, 2026 by qgallouedec Loading…
7
Add Nemotron 3 to tests via tiny model
#5278 opened Mar 12, 2026 by sergiopaniego Loading…
5 tasks
8
Centralize AI agent templates in .ai
#5268 opened Mar 10, 2026 by qgallouedec Loading…
5
Introduce backend rollout-completions interface and decouple OpenEnv helper from vLLM internals
#5256 opened Mar 10, 2026 by rycerzes Loading…
1
2
async streaming grpo w prefetch
#5250 opened Mar 9, 2026 by winglian Loading…
5 tasks
6
batch params together in weight sync and async update the weights
#5249 opened Mar 9, 2026 by winglian Loading…
5 tasks
5
Introduce minimal generation backend interface for GRPO and RLOO trainers
#5244 opened Mar 8, 2026 by rycerzes Loading…
1
6
feat: log raw importance ratios and fraction of truncation/masking in vLLM importance sampling correction
#5243 opened Mar 8, 2026 by muupan Loading…
1 of 5 tasks
1
4
[GRPO] Fix re-tokenization bug in tool-calling loop by concatenating token IDs
#5242 opened Mar 7, 2026 by qgallouedec Loading…
2
28
Update openenv examples to use environment_factory
#5235 opened Mar 6, 2026 by sergiopaniego Loading…
8 tasks
10
Allow reward functions to log extra columns and scalar metrics
#5233 opened Mar 6, 2026 by manueldeprada Loading…
19
vLLM Server Sync via LoRA Adapter Reload (avoid merge + full weight sync) for GRPO
#5188 opened Feb 26, 2026 by lfranceschetti Loading…
2
1
Simplify NeMo Gym user experience
#5156 opened Feb 24, 2026 by cmunley1 Loading…
3
DPO padding-free
#5141 opened Feb 21, 2026 by qgallouedec • Draft
5 tasks
6
[GKD] Buffer Implementation for Distillation Trainer
#5137 opened Feb 20, 2026 by cmpatino Loading…
3 tasks done
32
MGPO feature addition
#5126 opened Feb 19, 2026 by damoonsh Loading…
2 of 5 tasks
wip(neuron): add neuron integration for SFT
#5125 opened Feb 18, 2026 by michaelbenayoun • Draft
ProTip! Adding no:label will show everything without a label.