Official AQLM quantization of google/gemma-2b.
For this quantization, we used 2 codebooks of 8 bits.
Results (0-shot acc):
| Model | Quantization | WinoGrande | PiQA | HellaSwag | ArcE | ArcC | Model size, Gb |
|---|---|---|---|---|---|---|---|
| gemma-2b | None | 0.6472 | 0.7715 | 0.5279 | 0.7403 | 0.4053 | 5.0 |
| 2x8 | 0.5801 | 0.6828 | 0.3891 | 0.5791 | 0.2534 | 1.6 |
To learn more about the inference, as well as the information on how to quantize models yourself, please refer to the official GitHub repo.