# Release Notes

[![View Source On Gitee](https://mindspore-website.obs.cn-north-4.myhuaweicloud.com/website-images/r2.3.q1/resource/_static/logo_source_en.svg)](https://gitee.com/mindspore/golden_stick/blob/r0.4/RELEASE.md)
## MindSpore Golden Stick 0.4.0 Release Notes

### Major Features and Improvements

* Added post-training weight quantization W8A16 algorithm `RoundToNearest`, which realizes the lossless compression parameters of Llama2 7B/13B/70B and Baichuan2 13B networks by over 40%.

### API Change

* Added `PTQConfig` to configure the post-training quantization algorithm.
* Added `PTQMode` enumeration class, which can be configured in 'PTQConfig', is used to distinguish between the two phases of the quantization algorithm: the quantization phase and the deployment phase.
* Added `BackendTarget` enumeration class, which can be configured in `PTQConfig`, to indicate the backend to which the quantized network will eventually be deployed. For example, 'BackendTarget.Ascend' indicates that it will eventually be deployed to the Ascend backend of MindSpore.

### Contributors

Thanks goes to these wonderful people:

zhuxiaoxion, hangangqiang

Contributions of any kind are welcome!