phi-3-vision
Phi-3, -3.5, and -4 for Mac: Locally-run Vision and Language Models for Apple Silicon
streamline the fine-tuning process for multimodal models: PaliGemma 2, Florence-2, and Qwen2.5-VL