[worker] fix: correctly determine is_vlm_model if sp > 1 (#3282)

### What does this PR do?

> Add **concise** overview of what this PR aims to achieve or
accomplish. Reference related GitHub issues and PRs that help with the
review.

Address 2nd issue in
#3281 (comment)

Currently, if we use ulysses sp, we rely on `multi_modal_inputs` to
check if it's a multi-modal model, but this can go wrong when we set
`data.return_multi_modal_inputs=False`, as that field won't exist even
if it's the VLM model.

As a result, it would be a reliable way to check by seeing if
`vision_config` field is in `self.actor_module.config` referring to
https://github.com/volcengine/verl/blob/1985eb14ff3e28985087804f939b5ea3a26e32cf/verl/workers/fsdp_workers.py#L317-L320

### Checklist Before Starting

- [X] Search for similar PRs. Paste at least one query link here: ...
- [X] 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
```

### 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.

- [X] Read the [Contribute
Guide](https://github.com/volcengine/verl/blob/main/CONTRIBUTING.md).
- [X] 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`
- [X] Add / Update [the
documentation](https://github.com/volcengine/verl/tree/main/docs).
- [X] 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: ...
- [X] 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).)

Signed-off-by: Hollow Man <[email protected]>

Author: HollowMan6

verl/workers/actor/dp_actor.py

@@ -150,7 +150,9 @@ def _forward_micro_batch(
 
                 # pad and slice the inputs if sp > 1
                 if self.use_ulysses_sp:
-                    is_vlm_model = "multi_modal_inputs" in micro_batch.keys()
+                    is_vlm_model = hasattr(
+                        getattr(self.actor_module, "module", self.actor_module).config, "vision_config"
+                    )
                     if is_vlm_model:
                         # vlm model's inputs will be sliced after embedding
                         input_ids_rmpad, position_ids_rmpad, pad_size = ulysses_pad(

verl/workers/engine/fsdp/transformer_impl.py

@@ -758,7 +758,7 @@ def prepare_model_inputs(self, micro_batch: TensorDict):
 
             # pad and slice the inputs if sp > 1
             if self.use_ulysses_sp:
-                is_vlm_model = "multi_modal_inputs" in micro_batch.keys()
+                is_vlm_model = hasattr(getattr(self.module, "module", self.module).config, "vision_config")
                 if is_vlm_model:
                     # vlm model's inputs will be sliced after embedding
                     input_ids_rmpad, position_ids_rmpad, pad_size = ulysses_pad(