How to use ekryski/FastVLM-0.5B-4bit with MLX:
# Download the model from the Hub pip install huggingface_hub[hf_xet] hf download ekryski/FastVLM-0.5B-4bit --local-dir FastVLM-0.5B-4bit
