mlx_lm.convert
One command turns a bf16 Hugging Face repo into a quantized MLX folder.
When to use it
Use this when the community hasn't published the quant you need — convert the original weights yourself at any bit width and group size.
One command turns a bf16 Hugging Face repo into a quantized MLX folder.
Use this when the community hasn't published the quant you need — convert the original weights yourself at any bit width and group size.