DiffusPoll: Conditional Text Diffusion Model for Poll Generation

This repository provides the official implementation of the following paper:

DiffusPoll: Conditional Text Diffusion Model for Poll Generation. (ACL 2024)

1. Datasets

Datasets included in the ./datasets/ folder. We make our dataset from WeiboPolls.

*-key, *-topic means the ablations with different data after by the attribute extractor.

2. Requirements

pip install -r requirements.txt 

3. Training

The training script is launched in the scripts folder.

cd scripts
bash train.sh

Arguments explanation:

• --dataset: WeiboPolls datasets, mentioned above
• --div_loss: whether use the diversity loss
• --mask: whether use the task-specific mask strategy

4. Inference

You need to modify the path to model_dir, which is obtained in the training stage.

cd scripts
bash infer.sh

5. Evaluate

You need to specify the folder of decoded texts. This folder should contain the decoded files from the same model but sampling with different random seeds where |S|=10 .

cd scripts 
python eval_seq2seq.py --folder ../{your-path-to-outputs} --tokenizer char --mbr

Acknowledgement

DiffusPoll benifits from Diffuseq. We are grateful to the authors for work open-source.