Search results

Remove unused converter_type

community-request

Status: Open.

#749 In NVIDIA-NeMo/RL;

· yfw opened on Jul 25, 2025

Support NGC Pytorch containers

needs-design

Status: Open.

#741 In NVIDIA-NeMo/RL;

· terrykong opened on Jul 24, 2025

Add option for sequence padding with packed sequences

enhancementNew feature or requestNew feature or request

Status: Open.

Feature

#740 In NVIDIA-NeMo/RL;

· ahmadki opened on Jul 24, 2025

Improve sequence packing tests

testingRelated to testingRelated to testing

Status: Open.

Task

#739 In NVIDIA-NeMo/RL;

· ahmadki opened on Jul 24, 2025

NCCL error when using non-colocated generation on single node

bugSomething isn't workingSomething isn't working

Status: Open.

Bug

#722 In NVIDIA-NeMo/RL;

· yuki-666 opened on Jul 23, 2025

Reward Model Training on Megatron backend

Status: Open.

#720 In NVIDIA-NeMo/RL;

· terrykong opened on Jul 23, 2025

Sequence Packing (and Dynamic Batching) support in DPO

Status: Open.

#719 In NVIDIA-NeMo/RL;

· terrykong opened on Jul 23, 2025

Hang on Megatron backend with MoE models when EP>1 and DP>1

Status: Open.

#718 In NVIDIA-NeMo/RL;

· terrykong opened on Jul 23, 2025

Support Multi-epoch training in GRPO (and max_num_steps > max_steps_in_an_epoch)

Status: Open.

#711 In NVIDIA-NeMo/RL;

· terrykong opened on Jul 22, 2025

Break up grpo_train to be more composable

enhancementNew feature or requestNew feature or request

external

Status: Open.

#706 In NVIDIA-NeMo/RL;

· snowmanwwg opened on Jul 22, 2025

[meta] Before posting an issue

Status: Open.

#701 In NVIDIA-NeMo/RL;

· terrykong opened on Jul 21, 2025

[suggest] please refactor data module

external

Status: Open.

#688 In NVIDIA-NeMo/RL;

· tpoisonooo opened on Jul 18, 2025