Questions about Office31 dataset partition

[hehai131]

Thanks for your inspiring work in OSDA and your excellent code that help me much!

But when I reimplement your job in Office31 that something confused me that how did you split the Office31 dataset. Is that the first 10 categories as shared known and the last 11-31 categories as unknow in target dataset, as well as source only keep the first 10 categories in alphabetical order?

Thank you a lot!