EmoTrans: Emotional Transition-based Model for Emotion Recognition in Conversation

If you appreciate our project, please consider giving us a star ⭐ on GitHub to stay updated with the latest developments.

🚀 Welcome to the repo of EmoTrans!

EmoTrans addresses the unimodal ERC task by modeling emotion transition feature, accepted by COLING2024.

Zhongquan Jian, Ante Wang, Jinsong Su, Junfeng Yao, Meihong Wang, Qingqiang Wu

⚡️ Install

The following instructions are for Linux installation. We would like to recommend the requirements as follows.

• Python == 3.9.16
• CUDA Version >= 11.7

1. Clone this repository and navigate to the emotrans folder

git clone [email protected]:jian-projects/emotrans.git
cd emotrans

2. Install Package

conda create -n emotrans python==3.9.16
conda activate emotrans
# pip install -r env.txt

🌈 How to train and inference

1. Specify the path of RoBERTa-large in global_var.py

2. run the script:

python run_emotrans.py

3. We have re-written the code and thus lose the checkpoints, the performance can be easily reproduction with our provided code and hyper-parameters.

Citation

If you find it is useful for your research and applications, please cite using this BibTeX:

@inproceedings{2024.erc.emotrans,
  author={Zhongquan Jian and Ante Wang and Jinsong Su and Junfeng Yao and Meihong Wang and Qingqiang Wu},
  title={EmoTrans: Emotional Transition-based Model for Emotion Recognition in Conversation},
  booktitle={Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation},
  pages={5723--5733},
  year= {2024},
  doi={https://aclanthology.org/2024.lrec-main.508},
}