|
--- |
|
library_name: transformers |
|
tags: |
|
- llama |
|
- facebook |
|
- meta |
|
- llama-3 |
|
- conversational |
|
- text-generation-inference |
|
--- |
|
|
|
An official quantization of [meta-llama/Meta-Llama-3-8B](https://huggingface.co/meta-llama/Meta-Llama-3-8B) 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 16 weights. |
|
|
|
__The 1x16g16 models require aqlm inference library v1.1.6 or newer:__ |
|
|
|
`pip install aqlm[gpu,cpu]>=1.1.6` |
|
|
|
|
|
Note that a large portion of this model are the 16-bit embeddings/logits matrices. You can significantly reduce the model footprint by quantizing these matrices, e.g. using `bitsandbytes` LLM.int8 or NF4 formats. This does not require additional training. |
|
|
|
|
|
| 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 (this) | 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 | 1x16g8 | 4.57 | 21.9 | [Link](https://huggingface.co/ISTA-DASLab/Meta-Llama-3-70B-AQLM-PV-2Bit-1x16)| |
|
|
|
|
|
|
|
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. |
|
|
