Skip to content

Comments

[sglang, doc] feat: add NPU GRPO training scripts for Qwen3-30B (Megaton/SGLang backends) and update doc#5060

Merged
wucong25 merged 1 commit intoverl-project:mainfrom
hustmf:sglang-pr
Jan 29, 2026
Merged

[sglang, doc] feat: add NPU GRPO training scripts for Qwen3-30B (Megaton/SGLang backends) and update doc#5060
wucong25 merged 1 commit intoverl-project:mainfrom
hustmf:sglang-pr

Conversation

@hustmf
Copy link
Contributor

@hustmf hustmf commented Jan 27, 2026

What does this PR do?

add NPU GRPO training scripts for Qwen3-30B (Megaton/SGLang backends) and update doc

Checklist Before Starting

  • Search for similar PRs. Paste at least one query link here: ...
  • Format the PR title as [{modules}] {type}: {description} (This will be checked by the CI)
    • {modules} include fsdp, megatron, veomni, sglang, vllm, rollout, trainer, ci, training_utils, recipe, hardware, deployment, ray, worker, single_controller, misc, perf, model, algo, env, tool, ckpt, doc, data, cfg, reward
    • If this PR involves multiple modules, separate them with , like [megatron, fsdp, doc]
    • {type} is in feat, fix, refactor, chore, test
    • If this PR breaks any API (CLI arguments, config, function signature, etc.), add [BREAKING] to the beginning of the title.
    • Example: [BREAKING][fsdp, megatron] feat: dynamic batching

Test

For changes that can not be tested by CI (e.g., algorithm implementation, new model support), validate by experiment(s) and show results like training curve plots, evaluation results, etc.

API and Usage Example

Demonstrate how the API changes if any, and provide usage example(s) if possible.

# Add code snippet or script demonstrating how to use this

Design & Code Changes

Demonstrate the high-level design if this PR is complex, and list the specific changes.

Checklist Before Submitting

Important

Please check all the following items before requesting a review, otherwise the reviewer might deprioritize this PR for review.

@hustmf hustmf changed the title [sglang, doc] feat: add NPU GRPO training scripts for Qwen2.5-32B/Qwen3-30B (Megaton/SGLang backends) and update doc [sglang, doc] feat: add NPU GRPO training scripts for Qwen3-30B (Megaton/SGLang backends) and update doc Jan 27, 2026
@hustmf
Copy link
Contributor Author

hustmf commented Jan 28, 2026

@wuxibin89 @FightingZhen This pr is ready

# Hardware Configuration
trainer.nnodes="${NNODES}"
trainer.n_gpus_per_node="${NPUS_PER_NODE}"
trainer.device='npu'
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Suggested change
trainer.device='npu'

verl now supports automatic device configuration, so no longer need to explicitly set trainer.device='npu'.

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

we will remove it

+===========+=================+
| Python | == 3.11 |
+-----------+-----------------+
| HDK | >= 25.3.RC1 |
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Is 25.3.RC1 necessary? HDK updating is really challenging in a real production environment.

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

resharding in sglang is developed based on ipc which is supported in new HDK


基础环境准备
+-----------+-----------------+
| software | version |
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Do it match daily image?

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

same as the DockerFile we provide

@wucong25 wucong25 merged commit c98cb8c into verl-project:main Jan 29, 2026
8 checks passed
RobotGF pushed a commit to RobotGF/verl that referenced this pull request Jan 30, 2026
…ton/SGLang backends) and update doc (verl-project#5060)

### What does this PR do?

add NPU GRPO training scripts for Qwen3-30B (Megaton/SGLang backends)
and update doc

### Checklist Before Starting

- [ ] Search for similar PRs. Paste at least one query link here: ...
- [ ] Format the PR title as `[{modules}] {type}: {description}` (This
will be checked by the CI)
- `{modules}` include `fsdp`, `megatron`, `veomni`, `sglang`, `vllm`,
`rollout`, `trainer`, `ci`, `training_utils`, `recipe`, `hardware`,
`deployment`, `ray`, `worker`, `single_controller`, `misc`, `perf`,
`model`, `algo`, `env`, `tool`, `ckpt`, `doc`, `data`, `cfg`, `reward`
- If this PR involves multiple modules, separate them with `,` like
`[megatron, fsdp, doc]`
  - `{type}` is in `feat`, `fix`, `refactor`, `chore`, `test`
- If this PR breaks any API (CLI arguments, config, function signature,
etc.), add `[BREAKING]` to the beginning of the title.
  - Example: `[BREAKING][fsdp, megatron] feat: dynamic batching`

### Test

> For changes that can not be tested by CI (e.g., algorithm
implementation, new model support), validate by experiment(s) and show
results like training curve plots, evaluation results, etc.

### API and Usage Example

> Demonstrate how the API changes if any, and provide usage example(s)
if possible.

```python
# Add code snippet or script demonstrating how to use this
```

### Design & Code Changes

> Demonstrate the high-level design if this PR is complex, and list the
specific changes.

### Checklist Before Submitting

> [!IMPORTANT]
> Please check all the following items before requesting a review,
otherwise the reviewer might deprioritize this PR for review.

- [ ] Read the [Contribute
Guide](https://github.com/volcengine/verl/blob/main/CONTRIBUTING.md).
- [ ] Apply [pre-commit
checks](https://github.com/volcengine/verl/blob/main/CONTRIBUTING.md#code-linting-and-formatting):
`pre-commit install && pre-commit run --all-files --show-diff-on-failure
--color=always`
- [ ] Add / Update [the
documentation](https://github.com/volcengine/verl/tree/main/docs).
- [ ] Add unit or end-to-end test(s) to [the CI
workflow](https://github.com/volcengine/verl/tree/main/.github/workflows)
to cover all the code. If not feasible, explain why: ...
- [ ] Once your PR is ready for CI, send a message in [the `ci-request`
channel](https://verl-project.slack.com/archives/C091TCESWB1) in [the
`verl` Slack
workspace](https://join.slack.com/t/verl-project/shared_invite/zt-3855yhg8g-CTkqXu~hKojPCmo7k_yXTQ).
(If not accessible, please try [the Feishu group
(飞书群)](https://applink.larkoffice.com/client/chat/chatter/add_by_link?link_token=772jd4f1-cd91-441e-a820-498c6614126a).)
- [ ] If your PR is related to the `recipe` submodule, please also
update the reference to the submodule commit via `git submodule update
--remote` or `cd recipe && git pull origin main`.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

4 participants