Skip to content

Pull requests: NVIDIA-NeMo/RL

New pull request New
129 Open 1,150 Closed
129 Open 1,150 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

feat: support top-p top-k in grpo CI:Lfast Runs a fast test suite and re-use nightly `main` container (but sync dependencies to PRs version)
#2053 opened Mar 3, 2026 by yuki-97 • Draft
5
feat: Add YaRN rope scaling support on Magatron-Bridge
#2052 opened Mar 3, 2026 by RayenTian • Draft
4 tasks
3
docs: add prerequisites, troubleshooting, and build verification for GRPO quickstart community-request documentation Improvements or additions to documentation
#2051 opened Mar 3, 2026 by brluobt Loading…
3 tasks
1
10
feat: caching gym venvs and apptainer for super branch documentation Improvements or additions to documentation
#2050 opened Mar 3, 2026 by terrykong Loading…
4 tasks
ci: Allow cancelling of unit tests CI:docs Run doctest CI Relating to CI
#2045 opened Mar 2, 2026 by chtruong814 Loading…
4 tasks
2
chore: bumpup Megatron-Bridge submodule to main CI:L2 Run doctests, unit tests, functional tests, and convergence tests Run CICD
#2039 opened Mar 1, 2026 by ZhiyuLi-Nvidia Loading…
4 tasks
4
fp8 refit opt
#2037 opened Feb 28, 2026 by Jianbing-D • Draft
4 tasks
1
feat: Add chunked linear ce loss function from hidden states community-request
#2036 opened Feb 27, 2026 by pengdurice Loading…
3 of 4 tasks
9
chore: address deprecation warning for using a non-tuple sequence for multidimensional indexing CI:L1 Run doctests, unit tests, and functional tests
#2032 opened Feb 27, 2026 by ananthsub • Draft
1 of 4 tasks
feat: basic ppo training implementation
#2027 opened Feb 26, 2026 by hXl3s • Draft
4 tasks
1
3
feat: Dynamo router support
#2023 opened Feb 25, 2026 by jthomson04 • Draft
4 tasks
8
feat: split validation statistics by task name CI:L1 Run doctests, unit tests, and functional tests documentation Improvements or additions to documentation
#2019 opened Feb 24, 2026 by yuki-97 Loading…
19
feat: adding wandb table log feature, showing concrete test samples community-request documentation Improvements or additions to documentation
#2018 opened Feb 24, 2026 by vinhngx Loading…
4 tasks
4
ci: Test GB200 runner CI:L1 Run doctests, unit tests, and functional tests CI Relating to CI
#2017 opened Feb 24, 2026 by chtruong814 • Draft
4 tasks
fix: use smoothed reward metric for VLM GRPO CLEVR convergence tests CI:docs Run doctest CI:L0 Run doctests and unit tests
#2015 opened Feb 23, 2026 by aroshanghias-nvd Loading…
3 tasks
1
2
Megatron refit
#2014 opened Feb 23, 2026 by wdykas Loading…
4 tasks
feat: support async checkpoint for megatron policy worker in non-colocated generation
#2007 opened Feb 23, 2026 by ananthsub • Draft
4 tasks
perf: Update moe_token_dispatcher_type default to alltoall CI:L2 Run doctests, unit tests, functional tests, and convergence tests
#2004 opened Feb 22, 2026 by parthmannan Loading…
4 tasks
1
perf: shard concat overhead community-request
#2002 opened Feb 21, 2026 by pjo256 Loading…
3 of 4 tasks
1
3
feat: prefetch gym venvs CI:L0 Run doctests and unit tests documentation Improvements or additions to documentation
#2000 opened Feb 21, 2026 by terrykong Loading…
4 tasks
14
refit latest update based on ahmads refactor
#1993 opened Feb 19, 2026 by shanmugamr1992 Loading…
4 tasks
24
Add is_async to CheckpointingConfig TypedDict community-request
#1991 opened Feb 19, 2026 by dmvevents Loading…
3 tasks
2
Gdpo
#1986 opened Feb 18, 2026 by nbasyl Loading…
4 tasks
@nbasyl
78
ci: Remove environments CI Relating to CI
#1981 opened Feb 18, 2026 by ko3n1g • Draft
4 tasks
3
feat: expose ability to configure port ranges
#1976 opened Feb 17, 2026 by ananthsub • Draft
4 tasks
ProTip! Filter pull requests by the default branch with base:main.