pengbaolin / soloist

MIT License
77 stars 13 forks source link

Pretraining on Taskmaster #3

Open wise-east opened 2 years ago

wise-east commented 2 years ago

The paper mentions using Taskmaster as another dataset for pretraining. Is there code for preprocessing Taskmaster for training as well? I think there's only code for processing SGD.

mrghofrani commented 1 year ago

I also couldn't find any file associated with preprocessing of the Taskmaster dataset.