Checklist Before Starting

• Search for similar PR(s).

What does this PR do?

Add an example for DeepSeek 671B GRPO

Specific Changes

• Need [Mcore] dpskv3 671B #1694
• Set torch._dynamo.config.suppress_errors = True at entrypoint, need a more formal workaround.

Additional Info.

• vllm as backend, sglang working in process ([Feature] Performance boost for VerlEngine's update_weights_from_tensor sgl-project/sglang#6762). Merged when both backends are ready.
• For DeepSeek-V3-0324 at gsm8k, the reward starts from 0.8 and saturated at around 0.95 using only 3 steps.
• Memory peaks around 120GB, consider using TP/ETP for a lower requirement. Would be lovely to add a config for cards with memory below 80GB.

Checklist Before Submitting

• Read the Contribute Guide.
• Apply pre-commit checks.
• Add [BREAKING] to the PR title if it breaks any API.
• Update the documentation about your changes in the docs.
• Add CI test(s) if necessary.