- 12小时的单个消费级GPU上训练,性能就可以达到了ChatGPT的97%
- 4bit训练,但是性能与16bit相当
这个数据有点离谱。。
Tim Dettmers: QLoRA: 4-bit finetuning of LLMs is here! With it comes Guanaco, a chatbot on a single GPU, achieving 99% ChatGPT performance on the Vicuna benchmark:
Paper:
Code+Demo:
Samples:
Colab:
https://arxiv.org/abs/2305.14314 https://github.com/artidoro/qlora https://colab.research.google.com/drive/1kK6xasHiav9nhiRUJjPMZb4fAED4qRHb?usp=sharing https://colab.research.google.com/drive/17XEqL1JcmVWjHkT-WczdYkJlNINacwG7?usp=sharing