File size: 4,077 Bytes
0604be8
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
2dcccf9
6f0854f
0604be8
9093597
e46e040
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
0604be8
266efde
0604be8
5ebfd6b
0604be8
88d2178
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
---
license: cc-by-nc-4.0
language:
- en
quantized_by: TheMelonGod
pipeline_tag: text-generation
tags:
- quantized
- safetensors
- exllamav2
- mistral
base_model:
- Sao10K/MN-12B-Lyra-v4
base_model_relation: quantized
---
**Orignal Model by:** [Sao10K](https://huggingface.co/Sao10K)  
**Orignal Model:** [MN-12B-Lyra-v4](https://huggingface.co/Sao10K/MN-12B-Lyra-v4)  
For more information about the model, I highly recommend checking out the original model page and the creator while you're at it.

**ExLlamaV2 Quantizations:**  
**8.0bpw**:  [8hb](https://huggingface.co/TheMelonGod/MN-12B-Lyra-v4-exl2/tree/8hb-8.0bpw)   | [6hb](https://huggingface.co/TheMelonGod/MN-12B-Lyra-v4-exl2/tree/6hb-8.0bpw)  
**7.5bpw**:  [8hb](https://huggingface.co/TheMelonGod/MN-12B-Lyra-v4-exl2/tree/8hb-7.5bpw)   | [6hb](https://huggingface.co/TheMelonGod/MN-12B-Lyra-v4-exl2/tree/6hb-7.5bpw)  
**7.0bpw**:  [8hb](https://huggingface.co/TheMelonGod/MN-12B-Lyra-v4-exl2/tree/8hb-7.0bpw)   | [6hb](https://huggingface.co/TheMelonGod/MN-12B-Lyra-v4-exl2/tree/6hb-7.0bpw)  
**6.5bpw**:  [8hb](https://huggingface.co/TheMelonGod/MN-12B-Lyra-v4-exl2/tree/8hb-6.5bpw)   | [6hb](https://huggingface.co/TheMelonGod/MN-12B-Lyra-v4-exl2/tree/6hb-6.5bpw)  
**6.0bpw**:  [8hb](https://huggingface.co/TheMelonGod/MN-12B-Lyra-v4-exl2/tree/8hb-6.0bpw)   | [6hb](https://huggingface.co/TheMelonGod/MN-12B-Lyra-v4-exl2/tree/6hb-6.0bpw)  
**5.5bpw**:  [8hb](https://huggingface.co/TheMelonGod/MN-12B-Lyra-v4-exl2/tree/8hb-5.5bpw)   | [6hb](https://huggingface.co/TheMelonGod/MN-12B-Lyra-v4-exl2/tree/6hb-5.5bpw)  
**5.0bpw**:  [8hb](https://huggingface.co/TheMelonGod/MN-12B-Lyra-v4-exl2/tree/8hb-5.0bpw)   | [6hb](https://huggingface.co/TheMelonGod/MN-12B-Lyra-v4-exl2/tree/6hb-5.0bpw)  
**4.5bpw**:  [8hb](https://huggingface.co/TheMelonGod/MN-12B-Lyra-v4-exl2/tree/8hb-4.5bpw)   | [6hb](https://huggingface.co/TheMelonGod/MN-12B-Lyra-v4-exl2/tree/6hb-4.5bpw)  
**4.25bpw**: [8hb](https://huggingface.co/TheMelonGod/MN-12B-Lyra-v4-exl2/tree/8hb-4.25bpw) | [6hb](https://huggingface.co/TheMelonGod/MN-12B-Lyra-v4-exl2/tree/6hb-4.25bpw)  
**4.0bpw**:  [8hb](https://huggingface.co/TheMelonGod/MN-12B-Lyra-v4-exl2/tree/8hb-4.0bpw)   | [6hb](https://huggingface.co/TheMelonGod/MN-12B-Lyra-v4-exl2/tree/6hb-4.0bpw)  
**3.75bpw**: [8hb](https://huggingface.co/TheMelonGod/MN-12B-Lyra-v4-exl2/tree/8hb-3.75bpw) | [6hb](https://huggingface.co/TheMelonGod/MN-12B-Lyra-v4-exl2/tree/6hb-3.75bpw)  
**3.5bpw**:  [8hb](https://huggingface.co/TheMelonGod/MN-12B-Lyra-v4-exl2/tree/8hb-3.5bpw)   | [6hb](https://huggingface.co/TheMelonGod/MN-12B-Lyra-v4-exl2/tree/6hb-3.5bpw)  
**3.0bpw**:  [8hb](https://huggingface.co/TheMelonGod/MN-12B-Lyra-v4-exl2/tree/8hb-3.0bpw)   | [6hb](https://huggingface.co/TheMelonGod/MN-12B-Lyra-v4-exl2/tree/6hb-3.0bpw)  
**2.75bpw**: [8hb](https://huggingface.co/TheMelonGod/MN-12B-Lyra-v4-exl2/tree/8hb-2.75bpw) | [6hb](https://huggingface.co/TheMelonGod/MN-12B-Lyra-v4-exl2/tree/6hb-2.75bpw)  
**2.5bpw**:  [8hb](https://huggingface.co/TheMelonGod/MN-12B-Lyra-v4-exl2/tree/8hb-2.5bpw)   | [6hb](https://huggingface.co/TheMelonGod/MN-12B-Lyra-v4-exl2/tree/6hb-2.5bpw)  
**2.25bpw**: [8hb](https://huggingface.co/TheMelonGod/MN-12B-Lyra-v4-exl2/tree/8hb-2.25bpw) | [6hb](https://huggingface.co/TheMelonGod/MN-12B-Lyra-v4-exl2/tree/6hb-2.25bpw)  
**2.0bpw**:  [8hb](https://huggingface.co/TheMelonGod/MN-12B-Lyra-v4-exl2/tree/8hb-2.0bpw)   | [6hb](https://huggingface.co/TheMelonGod/MN-12B-Lyra-v4-exl2/tree/6hb-2.0bpw)  

[Measurement File](https://huggingface.co/TheMelonGod/MN-12B-Lyra-v4-exl2/blob/main/MN-12B-Lyra-v4-measurement.json) _(Default/built-in calibration dataset was used)_

If you need a specific model quantized or particular bits per weight, please let me know. I’m happy to help.

Your feedback and suggestions are always welcome! They help me improve and make quantizations better for everyone.

Special thanks to [turboderp](https://huggingface.co/turboderp) for developing the tools that made these quantizations possible. Your contributions are greatly appreciated!