[rollout]fix: vllm_rollout_spmd.py when return_raw_chat=True #2156

zyfzjsc988 · 2025-06-23T12:05:46Z

What does this PR do?

fix : batch size and the size of raw_prompt unmatching when setting data.return_raw_chat=True

fix bug when using data.return_raw_chat=True in GRPO algorithm with reward model:
File "/ossfs/workspace/repository/verl/verl/single_controller/ray/base.py", line 625, in func return getattr(self.worker_dict[key], name)(*args, **kwargs) File "/ossfs/workspace/repository/verl/verl/single_controller/base/decorator.py", line 534, in inner return func(*args, **kwargs) File "/ossfs/workspace/repository/verl/verl/workers/fsdp_workers.py", line 634, in generate_sequences output = self.rollout.generate_sequences(prompts=prompts) File "/ossfs/workspace/repository/verl/verl/utils/debug/performance.py", line 78, in f return self.log(decorated_function, *args, **kwargs) File "/ossfs/workspace/repository/verl/verl/utils/debug/performance.py", line 88, in log output = func(*args, **kwargs) File "/opt/conda/lib/python3.10/site-packages/torch/utils/_contextlib.py", line 116, in decorate_context return func(*args, **kwargs) File "/ossfs/workspace/repository/verl/verl/workers/rollout/vllm_rollout/vllm_rollout_spmd.py", line 346, in generate_sequences return DataProto(batch=batch, non_tensor_batch=non_tensor_batch) File "<string>", line 6, in __init__ File "/ossfs/workspace/repository/verl/verl/protocol.py", line 214, in __post_init__ self.check_consistency() File "/ossfs/workspace/repository/verl/verl/protocol.py", line 325, in check_consistency assert val.shape[0] == batch_size, f"key {key} length {len(val)} is not equal to batch size {batch_size}" AssertionError: key raw_prompt length 128 is not equal to batch size 640

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, sglang, vllm, rollout, trainer, ci, training_utils, recipe, hardware, deployment, ray, worker, single_controller, misc, perf, model, algo, env, tool, ckpt, doc, data
- 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

High-Level Design

Demonstrate the high-level design if this PR is complex.

Specific Changes

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.
Apply pre-commit checks: pre-commit install && pre-commit run --all-files --show-diff-on-failure --color=always
Add / Update the documentation.
Add unit or end-to-end test(s) to the CI workflow 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 in the verl Slack workspace.

when using `data.return_raw_chat=True` in GRPO algorithm with reward model error raise: File "/ossfs/workspace/repository/verl/verl/workers/rollout/vllm_rollout/vllm_rollout_spmd.py", line 346, in generate_sequences return DataProto(batch=batch, non_tensor_batch=non_tensor_batch) File "<string>", line 6, in __init__ File "/ossfs/workspace/repository/verl/verl/protocol.py", line 214, in __post_init__ self.check_consistency() File "/ossfs/workspace/repository/verl/verl/protocol.py", line 325, in check_consistency assert val.shape[0] == batch_size, f"key {key} length {len(val)} is not equal to batch size {batch_size}" AssertionError: key raw_prompt length 128 is not equal to batch size 640

…ine#2156) ### What does this PR do? > fix : batch size and the size of raw_prompt unmatching when setting `data.return_raw_chat=True` fix bug when using `data.return_raw_chat=True` in GRPO algorithm with reward model: ` File "/ossfs/workspace/repository/verl/verl/single_controller/ray/base.py", line 625, in func return getattr(self.worker_dict[key], name)(*args, **kwargs) File "/ossfs/workspace/repository/verl/verl/single_controller/base/decorator.py", line 534, in inner return func(*args, **kwargs) File "/ossfs/workspace/repository/verl/verl/workers/fsdp_workers.py", line 634, in generate_sequences output = self.rollout.generate_sequences(prompts=prompts) File "/ossfs/workspace/repository/verl/verl/utils/debug/performance.py", line 78, in f return self.log(decorated_function, *args, **kwargs) File "/ossfs/workspace/repository/verl/verl/utils/debug/performance.py", line 88, in log output = func(*args, **kwargs) File "/opt/conda/lib/python3.10/site-packages/torch/utils/_contextlib.py", line 116, in decorate_context return func(*args, **kwargs) File "/ossfs/workspace/repository/verl/verl/workers/rollout/vllm_rollout/vllm_rollout_spmd.py", line 346, in generate_sequences return DataProto(batch=batch, non_tensor_batch=non_tensor_batch) File "<string>", line 6, in __init__ File "/ossfs/workspace/repository/verl/verl/protocol.py", line 214, in __post_init__ self.check_consistency() File "/ossfs/workspace/repository/verl/verl/protocol.py", line 325, in check_consistency assert val.shape[0] == batch_size, f"key {key} length {len(val)} is not equal to batch size {batch_size}" AssertionError: key raw_prompt length 128 is not equal to batch size 640` ### 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`, `sglang`, `vllm`, `rollout`, `trainer`, `ci`, `training_utils`, `recipe`, `hardware`, `deployment`, `ray`, `worker`, `single_controller`, `misc`, `perf`, `model`, `algo`, `env`, `tool`, `ckpt`, `doc`, `data` - 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 ``` ### High-Level Design > Demonstrate the high-level design if this PR is complex. ### Specific Changes > 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?tab=readme-ov-file#contribution-guide). - [ ] Apply [pre-commit checks](https://github.com/volcengine/verl?tab=readme-ov-file#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).

zyfzjsc988 requested review from wuxibin89, PeterSH6 and chenhaiq as code owners June 23, 2025 12:05

vermouth1992 approved these changes Jun 23, 2025

View reviewed changes

vermouth1992 merged commit d69528f into volcengine:main Jun 23, 2025
1 check passed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

[rollout]fix: vllm_rollout_spmd.py when return_raw_chat=True #2156

[rollout]fix: vllm_rollout_spmd.py when return_raw_chat=True #2156

Uh oh!

zyfzjsc988 commented Jun 23, 2025

Uh oh!

Uh oh!

Uh oh!

[rollout]fix: vllm_rollout_spmd.py when return_raw_chat=True #2156

[rollout]fix: vllm_rollout_spmd.py when return_raw_chat=True #2156

Uh oh!

Conversation

zyfzjsc988 commented Jun 23, 2025

What does this PR do?

Checklist Before Starting

Test

API and Usage Example

High-Level Design

Specific Changes

Checklist Before Submitting

Uh oh!

Uh oh!

Uh oh!