mlx_lm.convert --hf-path THUDM/LongWriter-llama3.1-8b --mlx-path noguchis/mlx-LongWriter-llama3.1-8b --quantize --q-bits 8 --dtype bfloat16

Safetensors

Model size

2B params

Tensor type

F16

U32

Inference Providers NEW

This model isn't deployed by any Inference Provider. 🙋 Ask for provider support