lwgkzl / meddg Goto Github PK

TKS

我想自己线下评估一下模型，发现KD_Metric中需要new_cy_bii.pk，但这个文件没有公开，希望能公开一下，谢谢！

You have mentioned in your dataset that "First, a part of the dataset is manually annotated, and then an automatic program is designed to label the rest of the dataset." and then you have said "We then develop a program to label the rest conversation automatically. The program consists of regex rules for each item to cover all human-labeled data, and apply these rules to label other dialogues."

Can you please share the script for these regex rules?

Hi,

I am looking for English version dataset, however, from the google drive, I only saw Chinese version. Could you please release the English version as well?

Thank you

In the topic prediction folder, there are several scripts like baseline.py bert.py, baseline_five_class.py and bert_graph.py. Can you explain what each file is doing ?

Also, if you could tell how can we get the predictions of topic and classes for the test set from this script?