IAMM

Code and data for ACL 2024: An Iterative Associative Memory Model for Empathetic Response Generation.

Environment

CUDA Version==11.7
python==3.10.13
torch==2.0.1

Environment Installation

To run this code, you should：

1. Clone the project from github.

git clone https://github.com/zhouzhouyang520/IAMM.git

2. Enter the project root directory

cd IAMM/

3. Download the data, and put it into the project root directory. Baidu cloud link with Code: 8suu or Google cloud link

unzip ED.zip
mkdir data
mv ED data
mkdir vectors
mv glove.6B.300d.txt vectors

4. Install required packages

pip install -r requirements.txt

or

conda env create -f environment.yml

The project mainly referenced the code from CEM and ESCM. If there are issues during installation, you can try using the environments from these two projects.

Model Training

out_file=iamm.log 
nohup python main.py --model iamm --cuda --device_id 7 --pointer_gen --word_topk 5 --ctx_topk 15 --cs_topk 5 > $out_file 2>&1 &

Model Test

out_file=iamm_test.log 
nohup python main.py --model iamm --test --test_model_name IAMM_19_9496.3010_0.0268 --cuda --device_id 7 --pointer_gen --batch_size 48 --word_topk 5 --ctx_topk 15 --cs_topk 5 --hgnn_hidden 300 --hgnn_out 50 > $out_file 2>&1 &

If this work is helpful, please kindly cite as:

@inproceedings{
anonymous2024an,
title={An Iterative Associative Memory Model for Empathetic Response Generation},
author={Zhou Yang and Zhaochun Ren and Yufeng Wang and Haizhou Sun and Chao Chen and Xiaofei Zhu and Xiangwen Liao},
booktitle={The 62nd Annual Meeting of the Association for Computational Linguistics},
year={2024},
url={https://openreview.net/forum?id=VemvrNidxs}
}