[Bugfix] DiffusionGemma: fix multimodal support

[mmastrac]

Purpose

DiffusionGemma reuses Gemma4ForConditionalGeneration._process_image_input for its image path. That method reads self._enable_mm_lora, which Gemma4's constructor sets. DiffusionGemma's doesn't, so any request with an image raises AttributeError in the encoder runner and takes the engine down.

Set the flag in DiffusionGemma's constructor the same way Gemma4 does.

Required for #57250 to support multimodal inputs.

Test plan

Hardware: 1 DGX Spark.

Send an image.

vllm serve nvidia/diffusiongemma-26B-A4B-it-NVFP4 --max-model-len 4096 --diffusion-config '{"canvas_length": 64}'

curl -s localhost:8000/v1/chat/completions -H 'content-type: application/json' -d '{
  "model": "nvidia/diffusiongemma-26B-A4B-it-NVFP4", "max_tokens": 32,
  "messages": [{"role": "user", "content": [
    {"type": "image_url", "image_url": {"url": "https://upload.wikimedia.org/wikipedia/commons/thumb/4/47/PNG_transparency_demonstration_1.png/280px-PNG_transparency_demonstration_1.png"}},
    {"type": "text", "text": "What is in the image?"}]}]}'

Test results

Without patch: 500 error (AttributeError)

With patch: a successful response

---

Essential Elements of an Effective PR Description Checklist

• The purpose of the PR, such as "Fix some issue (link existing issues this PR will resolve)".
• The test plan, such as providing test command.
• The test results, such as pasting the results comparison before and after, or e2e results
• (Optional) The necessary documentation update, such as updating supported_models.md and examples for a new model.

[claude]

Claude Code Review

This pull request is from a fork — automated review is disabled. A repository maintainer can comment @claude review to run a one-time review.

[LucasWilkinson]

LGTM thanks for splitting this out!

[LucasWilkinson]

/ci run

[LucasWilkinson]

/ci run

[github-actions]

✅ Triggered Buildkite CI #89902 for commit 55e28de3232d.

[github-actions]

✅ @mmastrac, CI is now available for this PR.

• /ci run starts upstream CI; /amd-ci run starts AMD CI only.
• Your branch must contain every commit currently on its upstream target branch. Merge or rebase onto the latest target branch, then rerun the command. Append --allow-stale to a run command to test an outdated branch at your own risk.
• /ci retry retries failed jobs in the CI build for the current PR head. If the current head has no CI build, it starts a new CI build for the current head containing only jobs that failed in the latest earlier CI build for this PR.
• /amd-ci retry retries failed jobs in AMD CI for the current PR head. Use /amd-ci run when the current head has no AMD CI build.
• /ci cancel cancels scheduled or running CI builds for this PR branch; /amd-ci cancel does the same for AMD CI only.

[github-actions]

✅ CI is already running for this commit: https://buildkite.com/vllm/ci/builds/89902

[mmastrac]

/ci run

[github-actions]

✅ Triggered Buildkite CI #90083 for commit 73e6401b8670.