mlx_lm.generate --model mlx-community/OpenELM-270M-Instruct --prompt "Write a story about Einstein" --max-tokens 512 --temp=0.0 on M3Max64GB