About the dataset

[nicosouth]

hello!

i read your paper and github. i notice that you train the pre-trained model.
but i don't know how much data did you use in the pre-train.
can you provide any further information?

thank you!