Repository navigation
Test script to verify new models compatibility with RL pipeline #2
Description
Is your feature request related to a problem? Please describe.
Different libraries/environments for training vs generation can lead to unexpected errors which are hard to debug since RL is robust enough that just observing rewards isn't a good enough check.
Describe the solution you'd like
We have already added a "Adding New Models" guide which explains the importance of tracking logprob errors. The guide should be accompanied by a script that the user is expected to run every time they bring in a new model with RL pipeline which essentially tests if training and inference frameworks for a new model are compatible.
Activity
terrykong commentedon Aug 12, 2025
closing since we're moving towards standalone bite-sized diagnostic scripts https://docs.nvidia.com/nemo/rl/latest/adding-new-models.html#model-diagnostics
Merge pull request #2 from NVIDIA-NeMo/support_mcore_vpp
feat: add Qwen3.5-35B GRPO release test (NVIDIA-NeMo#2)
feat: add Qwen3.5-35B GRPO release test (#2)
feat: add Qwen3.5-35B GRPO release test (#2)
refactor(data-plane): consolidate producer codec, caller mints keys
refactor(data-plane): consolidate producer codec, caller mints keys
refactor(data-plane): consolidate producer codec, caller mints keys
refactor(data-plane): consolidate producer codec, caller mints keys
refactor(data-plane): consolidate producer codec, caller mints keys
refactor(xtoken): gold-loss sample_mask + config-driven vocab sizing
refactor(xtoken): gold-loss sample_mask + config-driven vocab sizing
feat(ft-experiment): finalize 4-setup fault-tolerance experiment configs
docs: enumerate all per-attempt W&B runs for controls in ft_experimen…
Merge pull request #2 from sharonyu-115/shuangy/nvfp4-vllm-side-quant
fix: address review feedback on inner TP-worker nsys profiling