hongyin/

chat-llama2-70b-80k

I am pleased to introduce an English-Chinese conversation assistant designed to reduce the cost of inference. The original vocabulary lacked good support for Chinese, so the vocabulary was customized and expanded. It is trained based on the Llama2-chat-70B, with a unique vocabulary and 70 billion parameters.

Due to time constraints, I only performed a small amount of instruction fine-tuning, which resulted in shortcomings in the model's QA capabilities. I hope you can continue to fine-tune your instructions.

Losing fat is the only way to solve all problems.

Please cite if you find it helpful.

@misc{zhu2023metaaid,
      title={MetaAID 2.5: A Secure Framework for Developing Metaverse Applications via Large Language Models}, 
      author={Hongyin Zhu},
      year={2023},
      eprint={2312.14480},
      archivePrefix={arXiv},
      primaryClass={cs.CR}
}