yeongjoonJu / KoDialogT5

Modeling Dialogue in Korean, leveraging T5

Korean Dialogue Generation Model (T5)

I had tried to use KE-T5 as a backbone but I think the model weights in PyTorch have some problems. So I recommend utilizing Kolang-T5.

Requirements

• python 3.7 or 3.8 can be used

  pip install -r requirements.txt

  I used CUDA 11.3 version

Logger

• Wandb (Weights & Biases) is good logger which can view text logs

  

Datasets

Update soon

Train

bash scripts/pretrain_dialog.sh

Interact

python interact.py