search

deu30303 / RUC

144 stars 15 forks source link

Final model selection #15

Closed qinghezeng closed 2 years ago

qinghezeng commented 2 years ago

Hi, thank you again for sharing the code!

I wonder which model did you use for the evaluation in the paper, Model 1 or Model 2 or some kind of ensembling (for example prediction3 in function test_ruc in lib/protocols.py)? And how did you select the best model? By selecting the best accuracy of prediction3 on training or validation set?

Thank you so much!

deu30303 commented 2 years ago

As I remember, I use ensembling results for evaluation. Because validation set is not possible for unsupervised learning setting, previous papers report both the last and best accuracy for comparison.