Knowledge Distillation of Large Language Models
Paper
•
2306.08543
•
Published
•
21
teacher-gpt2-1.5B is a gpt2-xlarge (1.5B) model supervised fine-tuned on databricks-dolly-15k.
It is used as the teacher model for MiniLLM series.
@inproceedings{minillm,
title={MiniLLM: Knowledge Distillation of Large Language Models},
author={Gu, Yuxian and Dong, Li and Wei, Furu and Huang, Minlie},
booktitle={Proceedings of ICLR},
year={2024}
}
Base model
openai-community/gpt2-xl