Skip to content

Latest commit

 

History

History
51 lines (42 loc) · 1.64 KB

README.md

File metadata and controls

51 lines (42 loc) · 1.64 KB

EnTeRED

This figure is generated by Midjourney, the right is backbone large model, and the left is knowledge lighweight model, the bridge is to the fusion strategy, and the river is to the semantic gap between different probility distribution.

EnTeRED Model

Introduction

This is the code for the paper Ered: Enhanced Text Representations with Entities and Descriptions.

Installation

pip install -r requirements.txt
python 3.9.9 cuda 11.6, torch 1.13.1

Contents

It includes the following contents:
├── baseline: bert, roberta
│ ├── run_finetune.py
│ └── command.sh
├── baseline: kadapter
│ ├── subdirectory1
│ ├── run_finetune_openentity_adapter.py
│ └── run_finetune_figer_adapter.py
│ ├── run_finetune_fewrel_adapter.py
│ └── run_finetune_tacrec_adapter.py
├── baseline: kadapter
│ ├── run_finetune_KT-attn.py
│ │ ├── command.sh
├── our model: DistilBERT as K
│ ├── run_KFormers.py
│ └── command_xxx.sh
└── our model: ChatGPT as K
├── run_entered_chatgpt_embedding.py

└── sst.sh

Contact Info

@misc{zhao2022ered,
title={Ered: Enhanced Text Representations with Entities and Descriptions},
author={Qinghua Zhao and Shuai Ma and Yuxuan Lei},
year={2022},
eprint={2208.08954},
archivePrefix={arXiv},
primaryClass={cs.CL}
}