Add custom processor for easier inference and mlx-vlm compatibility ea79a1f JettChenT commited on 6 days ago