[vllm] fix: apply moe weight loader patch for standard wight loading (#5234)

### What does this PR do?

After async rollout engine refactor, attempting to load a moe model
would fail with the following error because the weight loader patch
cannot apply correctly:

> AttributeError: 'Parameter' object has no attribute 'weight_loader'

And the [spmd implement in
v0.6.1](https://github.com/verl-project/verl/blob/release/v0.6.1/verl/workers/rollout/vllm_rollout/vllm_rollout_spmd.py#L647-L651)
as following:

```python
    async def update_weights(self, weights: Generator[tuple[str, torch.Tensor], None, None], **kwargs):
...
        else:
            from verl.utils.vllm.patch import patch_vllm_moe_model_weight_loader

            model = self.inference_engine.worker.model_runner.model
            patch_vllm_moe_model_weight_loader(model)
            model.load_weights(weights)
```

### Checklist Before Starting

- [ ] 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`, `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.

- [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`
- [ ] 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`.

Signed-off-by: zjchenn <zjchenn@gmail.com>

Author: zjchenn

verl/workers/rollout/vllm_rollout/utils.py

@@ -210,6 +210,14 @@ def update_weights_from_ipc(self, peft_config: dict = None, base_sync_done=False
             buffer, shm = rebuild_shared_memory(shm_name, shm_size, dtype=torch.uint8)
         socket.send(b"")
 
+        use_standard_weight_load = not (peft_config and base_sync_done) and not is_fp8_model(
+            self.model_runner.vllm_config
+        )
+
+        # Re-apply here because async IPC weight sync can happen long after init and lose MoE weight_loader attrs.
+        if use_standard_weight_load:
+            patch_vllm_moe_model_weight_loader(self.model_runner.model)
+
         # receive bucket and update weights
         while True:
             metadata = socket.recv_pyobj()