metadata

license: apache-2.0
datasets:
  - databricks/databricks-dolly-15k
language:
  - en
metrics:
  - rouge
base_model:
  - openai-community/gpt2-large
pipeline_tag: text-generation

SFT-gpt2-760M

paper | code

SFT-gpt2-760M is a gpt2-large (760M) model supervised fine-tuned on databricks-dolly-15k.

It is used as a baseline for MiniLLM.

Other Baselines

KD
SeqKD

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}
}