Skip to content

Pull requests: NVIDIA-NeMo/RL

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
Assignee
Filter by who’s assigned
Assigned to nobody Loading
Sort

Pull requests list

fix(grpo): skip unused logprob computations (rebases #2177, integrates #2174 + #2178) CI:Lfast Runs a fast test suite and re-use nightly `main` container (but sync dependencies to PRs version)
#2443 opened May 8, 2026 by jinglinglingling Loading…
Mxin/moe mamba sft Documentation Improvements or additions to documentation
#2442 opened May 8, 2026 by mxinO Contributor Draft
4 tasks
feat: data plane transfer queue integration
#2439 opened May 7, 2026 by ZhiyuLi-Nvidia Contributor Draft
4 tasks
Dynamo Nemo-RL K8s integration
#2429 opened May 6, 2026 by jthomson04 Contributor Draft
4 tasks
[WIP] don't review Documentation Improvements or additions to documentation
#2420 opened May 6, 2026 by shuyixiong Contributor Draft
4 tasks
feat: Auto research skill community-request
#2419 opened May 6, 2026 by vinhngx Contributor Loading…
fix: handle non-contiguous tensors in IPC weight refit community-request waiting-on-maintainers Waiting on maintainers to respond
#2418 opened May 5, 2026 by jlcanta Loading…
3 of 4 tasks
Support Megatron + SGLang community-request
#2416 opened May 5, 2026 by pengdurice Contributor Draft
2 of 4 tasks
[WIP] New refit integration branch
#2413 opened May 5, 2026 by youngeunkwon0405 Contributor Draft
4 tasks
ci: add maintainers
#2409 opened May 5, 2026 by thomasdhc Contributor Loading…
4 tasks
docs: add bump-dependency skill CI:docs Run doctest Documentation Improvements or additions to documentation
#2402 opened May 5, 2026 by ko3n1g Contributor Loading…
3 of 4 tasks
ci: Major refactor of release-workflows CI:docs Run doctest CI Relating to CI
#2397 opened May 5, 2026 by ko3n1g Contributor Loading…
2 of 3 tasks
fix(grpo): clearer error when overlong_filtering finds no truncated field community-request waiting-on-maintainers Waiting on maintainers to respond
#2395 opened May 4, 2026 by lonexreb Loading…
3 of 4 tasks
fix(nemo_gym): hard-fail when rollouts returned < requested community-request waiting-on-maintainers Waiting on maintainers to respond
#2394 opened May 4, 2026 by lonexreb Loading…
3 of 5 tasks
fix(workers): extend offload guard to v2 + Megatron paths community-request waiting-on-maintainers Waiting on maintainers to respond
#2393 opened May 4, 2026 by lonexreb Loading…
3 of 5 tasks
fix(dtensor): clear error when train/get_logprobs/score run offloaded community-request waiting-on-maintainers Waiting on maintainers to respond
#2392 opened May 4, 2026 by lonexreb Loading…
3 of 4 tasks
feat(megatron): expose checkpoint parallelism and RNG knobs community-request waiting-on-maintainers Waiting on maintainers to respond
#2391 opened May 4, 2026 by lonexreb Loading…
3 of 4 tasks
feat(megatron): expose hardcoded infrastructure params to user config community-request waiting-on-maintainers Waiting on maintainers to respond
#2390 opened May 4, 2026 by lonexreb Loading…
3 of 5 tasks
docs: add author field to README citation BibTeX community-request waiting-on-maintainers Waiting on maintainers to respond
#2389 opened May 4, 2026 by lonexreb Loading…
2 of 3 tasks
fix: bump accelerate floor to 1.13.0 for transformers 5.3.0 compat community-request waiting-on-maintainers Waiting on maintainers to respond
#2388 opened May 4, 2026 by lonexreb Loading…
2 of 4 tasks
test: cover converter CLI entry points community-request waiting-on-maintainers Waiting on maintainers to respond
#2387 opened May 4, 2026 by lonexreb Loading…
3 of 6 tasks
chore: Upgrade vLLM to 0.20.0 CI:L1 Run doctests, unit tests, and functional tests
#2384 opened May 1, 2026 by kajalj22 Contributor Draft
1 of 2 tasks
fix: configure port ranges to avoid TOCTOU port contention
#2380 opened May 1, 2026 by terrykong Collaborator Loading…
3 tasks
ProTip! no:milestone will show everything without a milestone.