The datasets are unreconstructable.

[TobeyYang]

First, thanks for your inspiring research. I am also major in the dialogue system and very interested in this work.
I ran the given scripts in StyleFusion to reconstruct the Reddit datasets and found some steps were missing and only some .tsv files are obtained in the end. It is hard to follow now. Could you please provide the train.num, vali.num and test.num and vocab.txt in your experiments instead of toy data, thank you.

[mattc95]

I have encountered the same issue above.