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

# usermma/VibeThinker-3B-mlx-5Bit

The Model [usermma/VibeThinker-3B-mlx-5Bit](https://huggingface.co/usermma/VibeThinker-3B-mlx-5Bit) was converted to MLX format from [WeiboAI/VibeThinker-3B](https://huggingface.co/WeiboAI/VibeThinker-3B) using mlx-lm version **0.31.2**.

## Use with mlx

```bash
pip install mlx-lm
```

```python
from mlx_lm import load, generate

model, tokenizer = load("usermma/VibeThinker-3B-mlx-5Bit")

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)
```