Skip to content

Pull requests: NVIDIA-NeMo/RL

New pull request New
72 Open 771 Closed
72 Open 771 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

fix: fix mcore train_iters in grpo CI:L1 Run doctests, unit tests, and functional tests
#1383 opened Oct 17, 2025 by yuki-97 Loading…
1
chore: use pydantic for yaml test validation documentation Improvements or additions to documentation r0.4.0
#1382 opened Oct 17, 2025 by terrykong Loading…
17
feat: add capability to update weights inflight during generation CI:L1 Run doctests, unit tests, and functional tests
#1381 opened Oct 16, 2025 by parthchadha Loading…
4 tasks
3
feat: Overlap param iteration and broadcast in non-colocated refit CI:L1 Run doctests, unit tests, and functional tests Performance Related to improving performance
#1379 opened Oct 16, 2025 by youngeunkwon0405 • Draft
4 tasks
@youngeunkwon0405
4
feat: additional validation losses for preference data documentation Improvements or additions to documentation
#1367 opened Oct 15, 2025 by jveronvialard • Draft
4 tasks
feat: Printing worker id to node.bundle_idx mapping
#1366 opened Oct 15, 2025 by youngeunkwon0405 • Draft
4 tasks
@youngeunkwon0405
3
Update RL to use megatron-bridge tot
#1358 opened Oct 14, 2025 by yaoyu-33 Loading…
4 tasks
5
feat: GSPO-token
#1357 opened Oct 14, 2025 by pjin-nvidia • Draft
4 tasks
feat: add kl penalty k1, k2 CI:L1 Run doctests, unit tests, and functional tests
#1349 opened Oct 13, 2025 by yuki-97 • Draft
5
feat: support truncated importance sampling CI:L1 Run doctests, unit tests, and functional tests
#1348 opened Oct 13, 2025 by yuki-97 Loading…
11
fix: Use custom chat template also in VLM's r0.4.0
#1344 opened Oct 11, 2025 by jseppanen Loading…
4
fix: Fix policy worker placement when using unified placement group CI:L1 Run doctests, unit tests, and functional tests r0.4.0
#1341 opened Oct 11, 2025 by guyueh1 Loading…
4 tasks
2
@guyueh1
29
Docs: Refactor Home Page and New About Section documentation Improvements or additions to documentation r0.4.0
#1338 opened Oct 10, 2025 by jgerh Loading…
@jgerh
10
feat: Integrate Penguin env
#1336 opened Oct 10, 2025 by bxyu-nvidia Loading…
4 tasks
21
chore: big version bump
#1334 opened Oct 10, 2025 by terrykong • Draft
3
17
fix: Add helpful error message if prepare_for_*() not called community-request
#1333 opened Oct 10, 2025 by Aniketsy Loading…
9
feat: add Megatron support for on-policy distillation CI:L1 Run doctests, unit tests, and functional tests r0.4.0
#1324 opened Oct 9, 2025 by zpqiu Loading…
4 tasks done
1
46
feat: Onboard perf recipes in tests
#1322 opened Oct 8, 2025 by guyueh1 • Draft
4 tasks
1
feat: Add debugger flag that can be turned on via RAY_DEBUG=legacy CI:L0 Run doctests and unit tests documentation Improvements or additions to documentation
#1312 opened Oct 8, 2025 by guyueh1 Loading…
4 tasks
@jgerh
20
feat: [Draft Do Not merge] Kitchen interface
#1310 opened Oct 8, 2025 by guyueh1 • Draft
4 tasks
fix: more robust fp8 rollout metric check CI:L0 Run doctests and unit tests r0.4.0
#1307 opened Oct 8, 2025 by terrykong Loading…
4 tasks
6
User/joyang/megatron cmp
#1298 opened Oct 7, 2025 by joyang-nv • Draft
4 tasks
3
fix: Pass policy.logprob_chunk_size to loss functions
#1295 opened Oct 7, 2025 by pjin-nvidia • Draft
4 tasks
3
fix: support arbitrary values for checkpointing.metric_name
#1291 opened Oct 6, 2025 by ashors1 Loading…
4 tasks
1
14
feat: Generation sampling params
#1290 opened Oct 6, 2025 by pjin-nvidia • Draft
4 tasks
ProTip! Filter pull requests by the default branch with base:main.