| license: mit | |
| tags: | |
| - openvino | |
| - int4 | |
| This is an INT4 quantized version of the `Phi-3-mini-128k-instruct` model. The Python packages used in creating this model are as follows: | |
| ``` | |
| openvino==2024.3.0.dev20240528 | |
| openvino-nightly==2024.3.0.dev20240528 | |
| openvino-tokenizers==2024.3.0.0.dev20240528 | |
| optimum==1.19.2 | |
| optimum-intel==1.17.0.dev0+aefabf0 | |
| nncf==2.11.0.dev0+90a7f0d5 | |
| torch==2.3.0+cu121 | |
| transformers==4.40.2 | |
| ``` | |
| This quantized model is created using the following command: | |
| ``` | |
| optimum-cli export openvino -m "microsoft/Phi-3-mini-128k-instruct" --task text-generation-with-past --weight-format int4 --group-size 128 --trust-remote-code ./Phi-3-mini-128k-instruct-ov-int4 | |
| ``` | |
| For more details, run the following command from your Python environment: `optimum-cli export openvino --help` | |
| INFO:nncf:Statistics of the bitwidth distribution: | |
| | Num bits (N) | % all parameters (layers) | % ratio-defining parameters (layers) | | |
| |----------------|-----------------------------|----------------------------------------| | |
| | 8 | 25% (27 / 130) | 21% (25 / 128) | | |
| | 4 | 75% (103 / 130) | 79% (103 / 128) | |