TonyNemo / UBAR-MultiWOZ

AAAI 2021: "UBAR: Towards Fully End-to-End Task-Oriented Dialog System with GPT-2"
96 stars 25 forks source link

DST code #10

Open aritraraut opened 2 years ago

aritraraut commented 2 years ago

what I understood is the train_DST.py file is to train the model to make it able to generate the belief states, but I am not getting the difference between the codes of train.py and the train_DST.py. I am unable to understand how it is correctly getting the context and the label for train_DST file.