search

yogeshbalaji / Generate_To_Adapt

Implementation of "Generate To Adapt: Aligning Domains using Generative Adversarial Networks"
https://arxiv.org/pdf/1704.01705.pdf
142 stars 33 forks source link

replicating results of GTA from mnist<->usps #11

Open stellaxu opened 5 years ago

stellaxu commented 5 years ago

Hi, from the github code, I did not found the code for transfer between mnist and usps. Would you please add this code part into the github? Currently, no matter what I have manipulated for these two datasets, the test accuracy is 70%+. May I know the reason for it? Thank you.

Dr-Zhou commented 4 years ago

Hi, from the github code, I did not found the code for transfer between mnist and usps. Would you please add this code part into the github? Currently, no matter what I have manipulated for these two datasets, the test accuracy is 70%+. May I know the reason for it? Thank you.

I meet the same question. Could you solve it?

lujingyang1029 commented 4 years ago

Hi, i got a another interesting problem. I am just running the default setup, svhn-> mnist, during the training, the val rate is around 90%, after eval process, the accuracy is around 75%. Not sure what happened.
Fyi, the epoch is set by default. Anybody has the same problem?