KD-OPT-1.3B
KD-OPT-1.3B is a model distilled from OPT-13B on databricks-dolly-15k with token-level forward KLD.
It is used as a baseline for MiniLLM.
Other Baselines
Citation
@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}
}
- Downloads last month
- 11
Model tree for MiniLLM/KD-OPT-1.3B
Base model
facebook/opt-1.3b