VibeThinker-3B-mlx-fp16/README.md

license, language, base_model, tags, pipeline_tag, library_name

license	language	base_model	tags	pipeline_tag	library_name
mit	en	WeiboAI/VibeThinker-3B	math code reasoning gpqa instruction-following mlx mlx-my-repo	text-generation	transformers

usermma/VibeThinker-3B-mlx-fp16

The Model usermma/VibeThinker-3B-mlx-fp16 was converted to MLX format from WeiboAI/VibeThinker-3B using mlx-lm version 0.31.2.

Use with mlx

pip install mlx-lm

from mlx_lm import load, generate

model, tokenizer = load("usermma/VibeThinker-3B-mlx-fp16")

prompt="hello"

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

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