Skip to content

README.md

Latest commit

 

History

History
33 lines (23 loc) · 1.18 KB

README.md

File metadata and controls

33 lines (23 loc) · 1.18 KB

Introducing Graph Context into Language Models through Parameter-Efficient Fine-Tuning for Lexical Relation Mining

Efficiently tuning language models using an inherent knowledge graph for lexical relation mining tasks.

Jingwen Sun, Zhiyi Tian, Yu He, Jingwei Sun, and Guangzhong Sun

University of Science and Technology of China (USTC)

image

News

  • 02.16.2025 Full code has been initially released.
  • 05.16.2025 Paper has been accepted to the main conference of ACL 2025.

Getting Started

You can simply run the following scripts: train_lrc.py, train_causal.py, and train_le.py to reproduce experimental results. An example is as follows:

python train_le.py \
        --dataset_path ./dataset/le/splits \
        --dataset lexical \
        --plm_path meta-llama/Meta-Llama-3-8B \
        --peft FT \
        --epoch 30 \
        --lr 1e-5 \
        --lr_min 1e-6

The folder ./dataset/le contains the lexical entailment (LE) datasets. The lexical relation classification datasets can be accessed from huggingface.