[model] fix: fix temp dtype (#4813)

### What does this PR do?

- As title. Prevent temperature to be int.

### 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`, `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`.

---------

Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>

Author: vermouth1992

verl/workers/engine/fsdp/transformer_impl.py

@@ -736,6 +736,7 @@ def prepare_model_inputs(self, micro_batch: TensorDict):
         if not isinstance(temperature, torch.Tensor):
             temperature = torch.tensor([temperature] * input_ids.shape[0], device=input_ids.device)
 
+        temperature = temperature.to(torch.float32)
         assert temperature.shape[0] == input_ids.shape[0]
 
         # args used to get outputs
@@ -876,7 +877,7 @@ def prepare_model_outputs(self, output, output_args, micro_batch: TensorDict):
                 entropy_rmpad = output.entropy.squeeze(0)  # (total_nnz,)
             else:
                 logits_rmpad = output.logits.squeeze(0)  # (total_nnz, vocab_size)
-                logits_rmpad.div_(temperature_rmpad.unsqueeze(-1))
+                logits_rmpad.div_(temperature_rmpad.clamp(min=1e-8).unsqueeze(-1).to(logits_rmpad.dtype))
 
                 # if use_sp: ((total_nnz / sp) + pad) ; if not use_sp: (batch, seqlen)
                 inplace_backward = True
@@ -935,7 +936,7 @@ def prepare_model_outputs(self, output, output_args, micro_batch: TensorDict):
                 logits = output.logits  # (bsz, response_length, vocab_size)
                 temperature = output_args["temperature"]  # (bsz,)
                 temperature = temperature.unsqueeze(-1).unsqueeze(-1)
-                logits.div_(temperature)
+                logits.div_(temperature.clamp(min=1e-8).to(logits.dtype))
 
                 if calculate_entropy:
                     if not self.engine_config.entropy_checkpointing:

verl/workers/engine/megatron/transformer_impl.py

@@ -619,6 +619,7 @@ def forward_step(self, batch_iter: Iterator[TensorDict], model, postprocess_micr
         if not isinstance(temperature, torch.Tensor):
             temperature = torch.tensor([temperature] * input_ids.shape[0], device=input_ids.device)
 
+        temperature = temperature.to(torch.float32)
         assert temperature.shape[0] == input_ids.shape[0]
         temperature = verl_F.expand_as_nested(temperature, input_ids)  # (bsz, j1)
 
@@ -639,7 +640,7 @@ def logits_processor(logits, label, temperature):
             # avoid non-positive temperature such as padding
             temperature[temperature <= 0] = 1e-8
             assert torch.all(temperature > 0).item(), f"temperature tensor must be positive. Got {temperature}"
-            logits.div_(temperature.unsqueeze(dim=-1))
+            logits.div_(temperature.unsqueeze(dim=-1).to(logits.dtype))
             ret = {}
             if calculate_entropy:
                 logits_bak = logits.clone()