Search results

refactor: leverage MCore TransformerConfig.__post_init__ validation for config fields set in _apply_performance_config

Status: Open.

#2291 In NVIDIA-NeMo/RL;

· terrykong opened on Apr 19, 2026

Track Qwen3.5-related issues

Status: Open.

#2281 In NVIDIA-NeMo/RL;

· zpqiu opened on Apr 17, 2026

Remove penguin from comment

bugSomething isn't workingSomething isn't working

Status: Open.

#2273 In NVIDIA-NeMo/RL;

· cmunley1 opened on Apr 16, 2026

train/gen_kl_error persistently high (0.15–1.75) during GRPO training with NeMo Gym VLM support

bugSomething isn't workingSomething isn't working

Status: Open.

#2260 In NVIDIA-NeMo/RL;

· shashank3959 opened on Apr 13, 2026

Make functional tests cover examples/converters/*

enhancementNew feature or requestNew feature or request

Status: Open.

#2259 In NVIDIA-NeMo/RL;

· yuki-97 opened on Apr 13, 2026

Adding the support for MiniMax-M2.7

enhancementNew feature or requestNew feature or request

Status: Open.

#2251 In NVIDIA-NeMo/RL;

· sharonyu-115 opened on Apr 11, 2026

RFC: GPU Time-Sharing Support for NeMo RL Async GRPO

community-request

enhancementNew feature or requestNew feature or request

needs-follow-upIssue needs follow-upIssue needs follow-up

Status: Open.

#2244 In NVIDIA-NeMo/RL;

· taoluo opened on Apr 10, 2026

Expose hardcoded Megatron infrastructure parameters to user config

community-request

needs-follow-upIssue needs follow-upIssue needs follow-up

Status: Open.

#2229 In NVIDIA-NeMo/RL;

· nic-nvidia opened on Apr 8, 2026

save_megatron_model deadlocks during HF-to-Megatron checkpoint conversion (fully_parallel_save)

community-request

needs-follow-upIssue needs follow-upIssue needs follow-up

Status: Open.

#2225 In NVIDIA-NeMo/RL;

· nic-nvidia opened on Apr 7, 2026

transformers 5.3.0 requires accelerate >= 1.1.0

bugSomething isn't workingSomething isn't working

community-request

needs-follow-upIssue needs follow-upIssue needs follow-up

Status: Open.

#2221 In NVIDIA-NeMo/RL;

· bogdansalyp opened on Apr 6, 2026

Make only_unmask_final flag configurable for SFT.

community-request

enhancementNew feature or requestNew feature or request

good first issueGood for newcomersGood for newcomers

needs-follow-upIssue needs follow-upIssue needs follow-up

Status: Open.

#2219 In NVIDIA-NeMo/RL;

· dhineshkumar-r opened on Apr 6, 2026

Adding support for Gemma4 model family

enhancementNew feature or requestNew feature or request

Status: Open.

#2212 In NVIDIA-NeMo/RL;

· sharonyu-115 opened on Apr 4, 2026