metadata
datasets:
- cerebras/SlimPajama-627B
language:
- en
This is the pretrained model described in paper "Layer-Condensed KV Cache for Efficient Inference of Large Language Models", Appendix C.1. To access this model, please email me (wuhy1_AT_shanghaitech_DOT_edu_DOT_cn) your username and what you intend to use the model weights for. Before we make the model fully public, please do not redistribute it on the Internet.
To use the model, please manually clone the repository and then load the model from the local folder. Unfortunately, the AutoModel.from_pretrained function will not work, as I have not uploaded the custom codes yet.
See more details at our github repo: https://github.com/whyNLP/LCKV