How to use beshkenadze/cohere-transcribe-03-2026-mlx-4bit with MLX:
# Download the model from the Hub pip install huggingface_hub[hf_xet] huggingface-cli download --local-dir cohere-transcribe-03-2026-mlx-4bit beshkenadze/cohere-transcribe-03-2026-mlx-4bit
beshkenadze commited on 