This repo contains the official implementation for the ACL 2023 paper: DiffuEmp: A Conditional Diffusion Language Model with Multi-Grained Control for Empathetic Response Generation

The main contribution of this paper is threefold: (1) We introduce explicit multi-grained control signals to solve the monotonous empathy problem, and convert the empathetic response generation into a controllable setting. (2) We propose DiffusEmp, a novel diffusion model-based framework, to unify the utilization of dialogue context and control signals, achieve elaborate control with a specific masking strategy, and integrate an emotion-enhanced matching method to produce diverse responses for a given context. (3) Experimental results show that our method outperforms competitive baselines in generating informative and empathetic responses.

pip install -r requirements.txt 

EmpatheticDialogue dataset comprises 24,850 open-domain multi-turn conversations between two interlocutors. To download the EmpatheticDialogues dataset:

wget https://dl.fbaipublicfiles.com/parlai/empatheticdialogues/empatheticdialogues.tar.gz

The frames_list.txt can be downloaded from the following link and placed in the same folder as the dataset. https://drive.google.com/file/d/1odTzaymJkfguF7Wl0vvcIAJwCVgtZXyl/view?usp=sharing

cd scripts
bash train.sh

cd scripts
bash run_decode.sh

cd scripts
python eval_seq2seq.py