Awesome

Dataset

Please download datasets to the corresponding directory under "data"

E-commerce https://drive.google.com/file/d/154J-neBo20ABtSmJDvm7DK0eTuieAuvw/view?usp=sharing.

Ubuntu https://www.dropbox.com/s/2fdn26rj6h9bpvl/ubuntudata.zip?dl=0

Douban https://www.dropbox.com/s/90t0qtji9ow20ca/DoubanConversaionCorpus.zip?dl=0&file_subpath=%2FDoubanConversaionCorpus

Our own dataset for segmentation is under DATASET directory

Source Code

• prepare data

  generate cutlist.txt

  <code>python segmentation_BERTCLS.py --datapath=data/xxx/xxx.txt</code>

  gather segmented data: data/xxx/xxxseg.txt:

  set interval = 2 for train.txt, interval = 10 for test.txt set corresponding datafile and dataset in data_process.py

  <code>python data_process.py</code>

• train

  <code>python run_TSbert_v3.py --task=alime --do_train --train_batch_size=20 --learning_rate=2e-5</code>

  The data will be saved in data/alime/input_cache_v3

  model will be saved in data/alime/model_save_v3, training log will also be saved in log.txt

• eval

  <code>python run_TSbert_v3.py --task=xxx</code>

  You can also load our trained model for testing https://drive.google.com/drive/folders/1_sRSmwlaAK_TPaVYYNhao81rXUW92z98?usp=sharing

Environment:

we use pre-trained BERT of pytorch version from https://github.com/huggingface/transformers

torch>=1.0.0

package: tqdm, boto3, requests, regex, sacremoses, openpyxl, numpy, sentencepiece

Reference

If you use this code please cite our paper:

@article{xu2020topic,
  title={Topic-aware multi-turn dialogue modeling},
  author={Xu, Yi and Zhao, Hai and Zhang, Zhuosheng},
  journal={arXiv preprint arXiv:2009.12539},
  year={2020}
}