[model] fix: stuck issue with mixed text-image data

[HollowMan6]

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.

On top of #3496

Possibly resolve #3665

This PR continue addressing the hanging issue as mentioned here: #3315 (comment)

It should be pretty safe and optional to do model.training checks even the model is in inference mode, as the value of inputs_embeds won't get modified anyway.

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

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.
• 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. (If not accessible, please try the Feishu group (飞书群).)

[gemini-code-assist]

Code Review

This pull request aims to fix a hanging issue with mixed text-image data by ensuring the vision part of the model is always exercised, even for text-only batches. This is achieved by removing the model.training condition, which triggers a dummy vision forward pass during inference as well. While this is a valid strategy to prevent deadlocks in distributed environments, it introduces a performance overhead for non-distributed inference. My review suggests refining this condition to apply the workaround only when in a distributed context, thus preserving performance for single-device inference scenarios.

[HollowMan6]

cc: @hiyouga @wlhgtc

[wlhgtc]

I encountered this problem as well: the model hangs in compute_log_prob, cause it was set to eval mode .
Great job!