README.md

---
license: mit
base_model: MiniMaxAI/MiniMax-M2
base_model_relation: quantized
quantized_by: turboderp
tags:
- exl3
---

EXL3 quants of [MiniMax-M2](https://huggingface.co/MiniMaxAI/MiniMax-M2)

⚠️ Requires ExLlamaV3 v0.0.12 (or v0.0.11 `dev` branch)

Base bitrates:

[2.00 bits per weight](https://huggingface.co/turboderp/MiniMax-M2-exl3/tree/2.0bpw)    
[3.00 bits per weight](https://huggingface.co/turboderp/MiniMax-M2-exl3/tree/3.0bpw)    
[4.00 bits per weight](https://huggingface.co/turboderp/MiniMax-M2-exl3/tree/4.0bpw)    

Optimized:

[2.04 bits per weight](https://huggingface.co/turboderp/MiniMax-M2-exl3/tree/2.04bpw)    
[2.27 bits per weight](https://huggingface.co/turboderp/MiniMax-M2-exl3/tree/2.27bpw)    
[3.04 bits per weight](https://huggingface.co/turboderp/MiniMax-M2-exl3/tree/3.04bpw)    
[3.50 bits per weight](https://huggingface.co/turboderp/MiniMax-M2-exl3/tree/3.5bpw)    
[4.03 bits per weight](https://huggingface.co/turboderp/MiniMax-M2-exl3/tree/4.03bpw)    


.        | KL-div |  ppl  | HumanEval@1
---------|--------|-------|-------------
2.00 bpw | 0.400  | 10.92 | 80.5%
2.04 bpw | 0.297  | 10.23 | 87.1%
2.27 bpw | 0.252  |  9.78 | 88.4%
3.00 bpw | 0.141  |  8.99 | 87.8%
3.04 bpw | 0.117  |  8.73 | 87.2% 
3.50 bpw | 0.094  |  8.78 | 88.4%
4.00 bpw | 0.087  |  8.58 | 89.6%
4.03 bpw | 0.077  |  8.61 | 87.8%
original |     -  |  8.51 | 87.2%¹

¹ Unconfirmed

<table>
  <tr>
    <td align="center">
      <a href="https://huggingface.co/turboderp/MiniMax-M2-exl3/blob/main/2.0bpw.svg">
        <img src="2.0bpw.svg" alt="2.00 bpw" width="160">
      </a>
      <div>2.00 bpw</div>
    </td>
    <td align="center">
      <a href="https://huggingface.co/turboderp/MiniMax-M2-exl3/blob/main/2.04bpw.svg">
        <img src="2.04bpw.svg" alt="2.04 bpw" width="160">
      </a>
      <div>2.04 bpw</div>
    </td>
    <td align="center">
      <a href="https://huggingface.co/turboderp/MiniMax-M2-exl3/blob/main/2.27bpw.svg">
        <img src="2.27bpw.svg" alt="2.27 bpw" width="160">
      </a>
      <div>2.27 bpw</div>
    </td>
    <td align="center">
      <a href="https://huggingface.co/turboderp/MiniMax-M2-exl3/blob/main/3.0bpw.svg">
        <img src="3.0bpw.svg" alt="3.00 bpw" width="160">
      </a>
      <div>3.00 bpw</div>
    </td>
  </tr>
  <tr>
    <td align="center">
      <a href="https://huggingface.co/turboderp/MiniMax-M2-exl3/blob/main/3.04bpw.svg">
        <img src="3.04bpw.svg" alt="3.04 bpw" width="160">
      </a>
      <div>3.04 bpw</div>
    </td>
    <td align="center">
      <a href="https://huggingface.co/turboderp/MiniMax-M2-exl3/blob/main/3.5bpw.svg">
        <img src="3.5bpw.svg" alt="3.50 bpw" width="160">
      </a>
      <div>3.50 bpw</div>
    </td>
    <td align="center">
      <a href="https://huggingface.co/turboderp/MiniMax-M2-exl3/blob/main/4.0bpw.svg">
        <img src="4.0bpw.svg" alt="4.00 bpw" width="160">
      </a>
      <div>4.00 bpw</div>
    </td>
    <td align="center">
      <a href="https://huggingface.co/turboderp/MiniMax-M2-exl3/blob/main/4.03bpw.svg">
        <img src="4.03bpw.svg" alt="4.00 bpw" width="160">
      </a>
      <div>4.03 bpw</div>
    </td>
  </tr>  
  <tr>
    <td align="center">
      <a href="https://huggingface.co/turboderp/MiniMax-M2-exl3/blob/main/api.svg">
        <img src="api.svg" alt="API" width="160">
      </a>
      <div>API</div>
    </td>
  </tr>  
</table>