bobig/FuseO1-R1-QwQ-SkyT1-Flash-32B-Q8

Quant made with the latest mlx-lm

This model is very good, my 2nd favorite for unpluged coding on Macs. SpecDec works with this draft model DeepScaleR-1.5B-Preview-Q8 but the acceptance rate is only 61%.

The Model bobig/FuseO1-R1-QwQ-SkyT1-Flash-32B-Q8 was converted to MLX format from FuseAI/FuseO1-DeepSeekR1-QwQ-SkyT1-Flash-32B-Preview using mlx-lm version 0.21.4.

Use with mlx

pip install mlx-lm

from mlx_lm import load, generate

model, tokenizer = load("bobig/FuseO1-R1-QwQ-SkyT1-Flash-32B-Q8")

prompt = "hello"

if tokenizer.chat_template is not None:
    messages = [{"role": "user", "content": prompt}]
    prompt = tokenizer.apply_chat_template(
        messages, add_generation_prompt=True
    )

response = generate(model, tokenizer, prompt=prompt, verbose=True)