File size: 1,600 Bytes
5801a94 f6ae235 5801a94 77e69e1 2573c50 f130fdc 5801a94 |
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 |
---
library_name: transformers
tags:
- llama
- facebook
- meta
- llama-3
- conversational
- text-generation-inference
---
An official quantization of [meta-llama/Meta-Llama-3-70B](https://huggingface.co/meta-llama/Meta-Llama-3-70B) using [PV-Tuning](https://arxiv.org/abs/2405.14852) on top of [AQLM](https://arxiv.org/abs/2401.06118) .
For this quantization, we used 1 codebook of 16 bits for groups of 8 weights.
| Model | AQLM scheme | WikiText 2 PPL | Model size, Gb | Hub link |
|------------|-------------|----------------|----------------|--------------------------------------------------------------------------|
| meta-llama/Meta-Llama-3-8B | 1x16g8 | 6.99 | 4.1 | [Link](https://huggingface.co/ISTA-DASLab/Meta-Llama-3-8B-AQLM-PV-2Bit-1x16) |
| meta-llama/Meta-Llama-3-8B | 1x16g16 | 9.43 | 3.9 | [Link](https://huggingface.co/ISTA-DASLab/Meta-Llama-3-8B-AQLM-PV-1Bit-1x16) |
| meta-llama/Meta-Llama-3-70B (this) | 1x16g8 | 4.57 | 21.9 | [Link](https://huggingface.co/ISTA-DASLab/Meta-Llama-3-70B-AQLM-PV-2Bit-1x16)|
The 1x16g16 (1-bit) models are on the way, as soon as we update the inference lib with their respective kernels.
To learn more about the inference, as well as the information on how to quantize models yourself, please refer to the [official GitHub repo](https://github.com/Vahe1994/AQLM).
The original code for PV-Tuning can be found in the [AQLM@pv-tuning](https://github.com/Vahe1994/AQLM/tree/pv-tuning) branch.
|