Add custom processor for easier inference and mlx-vlm compatibility 36be018 JettChenT commited on 5 days ago