Gemma: LLM - Spanish response to the English prompt.

[psiddh]

Model URL: https://huggingface.co/pytorch/gemma-3-4b-it-HQQ-INT8-INT4/resolve/main/model.pte

Tokenizer URL: https://huggingface.co/pytorch/gemma-3-4b-it-HQQ-INT8-INT4/resolve/main/tokenizer.json

Prompt:
what is in this photo

Temperature: 0

Runner invocation:
(Use https://github.com/pytorch/executorch/blob/main/extension/llm/runner/multimodal_runner.cpp#L88)
"<start_of_turn>user\n<start_of_image>"
Image (896x896 normalized)
"what is in this photo<end_of_turn>\n<start_of_turn>model"

cc @larryliu0820 @mergennachin @cccclai @helunwencser @jackzhxng