Add a Mistral3 multimodal vision-language model to the library. The implementation should include:

Add a Mistral3 multimodal vision-language model to the library. The implementation should include:: a task in HF ML Tasksmith (Harbor dataset). Mistral3Config ( model type="mistral3" ) with sub-configs vision config (default: Pixtral) and text config (default: Mistral), and fields: image token…