We will make the code and data public in this repository after legal and user privacy review.
Feb 22nd, 2024: Our project has undergone numerous evaluations by AntGroup, and recently, we decided to test and refine our method on some public LLMs. We aim to release it on GitHub by the end of March.
April 8th, 2024:
Sorry about that. We are currently planning to make some new code adjustments. We will complete and make them public as soon as possible. Those interested in efficient LLMs can first check out our other repository on PainlessInferenceAcceleration: https://github.com/alipay/PainlessInferenceAcceleration.