Add Gemma 3 test.

[pytorchbot]

Summary: Also, let multimodal runner accept image/audio data as floats too

Differential Revision: D84001548

[pytorch-bot]

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/14837

• 📄 Preview Python docs built from this PR

Note: Links to docs will display an error until the docs builds have been completed.

❗ 2 Active SEVs

There are 2 currently active SEVs. If your PR is affected, please view them below:

• Macos CI runners unavailable
• [ROCm][CI] Machines under the label linux.rocm.gpu.2 are undergoing maintenance.

❌ 1 Cancelled Job, 3 Pending

As of commit 8f29449 with merge base e0dda90 ():

CANCELLED JOB - The following job was cancelled. Please retry:

• pull / test-llava-runner-linux / linux-job (gh)
  ##[error]The operation was canceled.

This comment was automatically generated by Dr. CI and updates every 15 minutes.

[GregoryComer]

Rebasing to see if later picks resolve the build issue.

[GregoryComer]

@shoumikhin Do you have a recommendation to fix this build error? Thanks.

/Users/runner/work/executorch/executorch/pytorch/executorch/extension/llm/apple/ExecuTorchLLM/Exported/ExecuTorchLLMMultimodalRunner.mm:250:58: error: no matching constructor for initialization of 'llm::Audio'
          nativeInputs.emplace_back(llm::MultimodalInput(llm::Audio(
                                                         ^
In file included from /Users/runner/work/executorch/executorch/pytorch/executorch/extension/llm/apple/ExecuTorchLLM/Exported/ExecuTorchLLMMultimodalRunner.mm:13:
In file included from /Users/runner/work/executorch/executorch/pytorch/executorch/extension/llm/runner/multimodal_runner.h:24:
In file included from /Users/runner/work/executorch/executorch/pytorch/executorch/extension/llm/runner/irunner.h:18:
In file included from /Users/runner/work/executorch/executorch/pytorch/executorch/extension/llm/runner/stats.h:11:
In file included from /Users/runner/work/executorch/executorch/pytorch/executorch/extension/llm/runner/util.h:11:
In file included from /Users/runner/work/executorch/executorch/pytorch/executorch/extension/llm/runner/multimodal_prefiller.h:15:
In file included from /Users/runner/work/executorch/executorch/pytorch/executorch/extension/llm/runner/multimodal_input.h:14:
/Users/runner/work/executorch/executorch/pytorch/executorch/extension/llm/runner/audio.h:35:24: note: candidate constructor (the implicit copy constructor) not viable: requires 1 argument, but 4 were provided
struct ET_EXPERIMENTAL Audio {
                       ^
/Users/runner/work/executorch/executorch/pytorch/executorch/extension/llm/runner/audio.h:35:24: note: candidate constructor (the implicit move constructor) not viable: requires 1 argument, but 4 were provided
/Users/runner/work/executorch/executorch/pytorch/executorch/extension/llm/runner/audio.h:35:24: note: candidate constructor (the implicit default constructor) not viable: requires 0 arguments, but 4 were provided

[shoumikhin]

@GregoryComer it fails because of the lack of another cherry-pick
@larryliu0820 knows which one exactly