license, datasets, language, metrics, base_model, pipeline_tag

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

MiniLLM-gpt2-120M

paper | code

MiniLLM-gpt2-120M is a gpt2-base (120M) model distilled from gpt2-xlarge (1.5B) on databricks-dolly-15k.

Note: MiniLLM requires a SFT model for initilization to perform the PPO optimization.

Evaluation

We ask GPT-4 to give scores for the generated responses of MiniLLM. The prompts are taken from databricks-dolly-15k (test set), self-instruct, and vicuna

Baseline Models

• SFT w/o KD
• 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}
}