File size: 910 Bytes
7579181
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
---
library_name: transformers
tags:
- aqlm
- llama
- facebook
- meta
- llama-3
- conversational
- text-generation-inference
base_model: meta-llama/Llama-3.2-1B-Instruct
---

Official [AQLM](https://arxiv.org/abs/2401.06118) quantization of [meta-llama/Llama-3.2-1B
](https://huggingface.co/meta-llama/Llama-3.2-1B-Instruct) finetuned with [PV-Tuning](https://arxiv.org/abs/2405.14852).

For this quantization, we used 2 codebooks of 8 bits and groupsize of 8.

Results: 
| Model | Quantization | MMLU (5-shot) | ArcC   | ArcE   | Hellaswag | PiQA   | Winogrande | Model size, Gb |
|-------|--------------|---------------|--------|--------|-----------|--------|------------|----------------|
| meta-llama/Llama-3.2-1B-Instruct | fp16         | 0.4551 | 0.3584	| 0.6856 | 0.4520 | 0.7432 | 0.5951 | 2.5            |
|       | 2x8g8        | 0.2886 | 0.2918 | 0.6065 | 0.3990 | 0.7057	| 0.5746 | 0.8           |