Skip to content

Latest commit

 

History

History
29 lines (18 loc) · 792 Bytes

README.md

File metadata and controls

29 lines (18 loc) · 792 Bytes

Introduction

Code for the paper, DPTDR: Deep Prompt Tuning for Dense Passage Retrieval. We mainly develped our code based on coCondenser and P-tuning v2.

Installation

For environment, please run sh install_env.sh in a clean conda environment of python>=3.7. Then just run pip install -e ..

Reproduction for MS-MARCO

Please refer to examples/msmarco-passage-ranking.

Reproduction for Natural Questions

Please refer to examples/natural-questions.

References

@article{tang2022dptdr,
  title={DPTDR: Deep Prompt Tuning for Dense Passage Retrieval},
  author={Tang, Zhengyang and Wang, Benyou and Yao, Ting},
  journal={arXiv preprint arXiv:2208.11503},
  year={2022}
}