-
Notifications
You must be signed in to change notification settings - Fork 522
All issues
Issue creation is restricted in this repository
Issues
is:issue state:open
is:issue state:open
Search results
Split-API train step misses the TP grad all-reduce for replicated params (qk-layernorm / SP layernorms)
bugSomething isn't workingSomething isn't workingStatus: Open.#3716 In NVIDIA-NeMo/RL;- Status: Open.#3698 In NVIDIA-NeMo/RL;
[RFC] Make native LoRA adapter sync the default for RL refit and keep merged full-weight sync opt-in
enhancementNew feature or requestNew feature or requestStatus: Open.#3696 In NVIDIA-NeMo/RL;[Async PPO] cleanup tracking issue
enhancementNew feature or requestNew feature or requestStatus: Open.#3695 In NVIDIA-NeMo/RL;MXFP8 rollout support for Qwen3.5 (grouped-MoE expert refit)
enhancementNew feature or requestNew feature or requestwaiting-on-maintainersWaiting on maintainers to respondWaiting on maintainers to respondStatus: Open.#3694 In NVIDIA-NeMo/RL;Qwen3.8 27B RL validation and recipe
DocumentationImprovements or additions to documentationImprovements or additions to documentationenhancementNew feature or requestNew feature or requestStatus: Open.#3675 In NVIDIA-NeMo/RL;Verify evals on Papers with Code
DocumentationImprovements or additions to documentationImprovements or additions to documentationwaiting-on-maintainersWaiting on maintainers to respondWaiting on maintainers to respondStatus: Open.#3664 In NVIDIA-NeMo/RL;MCore refit backend nvshmem does not function
bugSomething isn't workingSomething isn't workingStatus: Open.#3646 In NVIDIA-NeMo/RL;[Bug] save_consolidated times out and truncates Gemma 4 31B safetensors with FSDP2/DTensor in v0.7.0
bugSomething isn't workingSomething isn't workingStatus: Open.#3624 In NVIDIA-NeMo/RL;docs: update stale config-conventions skill guidance
DocumentationImprovements or additions to documentationImprovements or additions to documentationStatus: Open.#3598 In NVIDIA-NeMo/RL;Anchor TimeoutChecker at SLURM_JOB_START_TIME so checkpoint_must_save_by budgets from allocation start
bugSomething isn't workingSomething isn't workingStatus: Open.#3596 In NVIDIA-NeMo/RL;SC checkpointing: persist partial rollout state so in-flight generations resume instead of restarting
bugSomething isn't workingSomething isn't workingDocumentationImprovements or additions to documentationImprovements or additions to documentationStatus: Open.#3594 In NVIDIA-NeMo/RL;
