Datasets:

ArXiv:
License:

Why processed English data is relatively small in size?

#9
by Xueren - opened

Why the processed English MedDialog only has 603 (482 + 60 + 61) dialogues. In the paper(https://aclanthology.org/2020.emnlp-main.743.pdf) Table 2, there are 257,332 dialogues in English dataset. What kind of processing is done in the processed dataset?

And I found the link to get raw English dataset is invalid, https://drive.google.com/drive/folders/1g29ssimdZ6JzTST6Y8g6h-ogUNReBtJD?usp=sharing. Can authors provide the link to download the English dataset? Thanks!

Sign up or log in to comment