Skip to content

Pull requests: NVIDIA-NeMo/RL

New pull request New
287 Open 1,712 Closed
287 Open 1,712 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

Add Nano v3 async GRPO config
#2973 opened Jun 28, 2026 by snowmanwwg Contributor Loading…
1
feat: NCCL-Xfer refit merge PR (FFN-only support version) CI:L1 Run doctests, unit tests, and functional tests Performance Related to improving performance
#2971 opened Jun 27, 2026 by youngeunkwon0405 Contributor Loading…
4 tasks
@youngeunkwon0405
3
ci: Bump Megatron-Bridge to 2a3f64b CI:L1 Run doctests, unit tests, and functional tests
#2969 opened Jun 27, 2026 by svcnvidia-nemo-ci Contributor Loading…
3
Ashors/super nightlies2
#2968 opened Jun 27, 2026 by ashors1 Contributor Loading…
4 tasks
1
fix(grpo_sync): skip refit for colocated MegatronGeneration CI:Lfast Runs a fast test suite and re-use nightly `main` container (but sync dependencies to PRs version) r0.7.0
#2967 opened Jun 26, 2026 by ZhiyuLi-Nvidia Contributor Loading…
2 tasks done
2
fix: Add missing variables in mopd nightly CI:Lfast Runs a fast test suite and re-use nightly `main` container (but sync dependencies to PRs version)
#2966 opened Jun 26, 2026 by yfw Contributor Loading…
4 tasks
2
fix: tune small scale super configs for h100 CI:L0 Run doctests and unit tests super-v3
#2965 opened Jun 26, 2026 by macandro96 Contributor Loading…
4 tasks
2
Fix: resolve the issue regarding HybridEP with sequence packing illegal memory address issue
#2964 opened Jun 26, 2026 by seonjinn Contributor Loading…
4 tasks
@seonjinn
4
fix: allow router replay trace fallback composition CI:Lfast Runs a fast test suite and re-use nightly `main` container (but sync dependencies to PRs version) r0.7.0
#2963 opened Jun 26, 2026 by zyzhou5 Contributor Loading…
4
fix: sanitize non-finite logprobs in vllm async worker community-request
#2962 opened Jun 26, 2026 by yuchenwang3 Loading…
1
fix: bump prometheus-fastapi-instrumentator>=8.0.2 for fastapi>=0.137 compat CI:L1 Run doctests, unit tests, and functional tests r0.7.0
#2960 opened Jun 26, 2026 by kajalj22 Contributor Loading…
4
fix(vllm): reset MoE weights to 3D before refit on Blackwell
#2959 opened Jun 26, 2026 by ashors1 Contributor • Draft
4 tasks
1
1
feat(vlm): route VLM GRPO through TQ trainer when data_plane.enabled CI:L1 Run doctests, unit tests, and functional tests
#2957 opened Jun 26, 2026 by ZhiyuLi-Nvidia Contributor Loading…
1 task done
2
test(xtoken): add >=2-teacher nightly for cross-tokenizer distillation CI:L0 Run doctests and unit tests Documentation Improvements or additions to documentation
#2952 opened Jun 26, 2026 by avenkateshha Contributor Loading…
4
feat: Add configurable vLLM thinking token budget community-request Documentation Improvements or additions to documentation waiting-on-maintainers Waiting on maintainers to respond
#2947 opened Jun 26, 2026 by kota-row Loading…
4 tasks done
1
feat: auto-detect CPUS_PER_WORKER from Slurm in ray.sub CI:docs Run doctest Documentation Improvements or additions to documentation
#2943 opened Jun 25, 2026 by terrykong Collaborator Loading…
3
feat(algorithms): AReaL decoupled-PPO async training on the SingleController
#2940 opened Jun 25, 2026 by haitian-nvidia • Draft
3 of 4 tasks
1
draft: gym yield rollouts Documentation Improvements or additions to documentation
#2939 opened Jun 25, 2026 by yfw Contributor • Draft
4 tasks
1
feat: support MTP inference for nemotron super
#2938 opened Jun 25, 2026 by yfw Contributor • Draft
4 tasks
1
feat(data_plane): bump TransferQueue to v0.1.8 for mooncake cpu rdma backend
#2935 opened Jun 25, 2026 by ZhiyuLi-Nvidia Contributor Loading…
4 tasks
2
[Don't merge] feat(grpo): add NRL_GEN_BENCHMARK_SKIP_TRAINING for async generation
#2930 opened Jun 25, 2026 by RayenTian Contributor • Draft
4 tasks
1
refactor(eval): route AIME eval through the response dataset registry CI:L1 Run doctests, unit tests, and functional tests
#2928 opened Jun 25, 2026 by NolenLiang Contributor Loading…
2
chore(docker): remove royalty-obligating codec libs from release image
#2922 opened Jun 24, 2026 by kajalj22 Contributor Loading…
2 tasks
1
fix: serialize uv installs to avoid nvidia-cutlass-dsl install race CI:Lfast Runs a fast test suite and re-use nightly `main` container (but sync dependencies to PRs version)
#2921 opened Jun 24, 2026 by youngeunkwon0405 Contributor Loading…
4 tasks
@youngeunkwon0405
5
perf: batch worker port discovery CI:Lfast Runs a fast test suite and re-use nightly `main` container (but sync dependencies to PRs version)
#2920 opened Jun 24, 2026 by macandro96 Contributor Loading…
4 tasks
3
ProTip! Adding no:label will show everything without a label.