You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Thank you very much for sharing your wonderful work.
I have two question: 1. When I finish running the experiment, I can see many log files. For example, when running 10 tasks on CIFAR100, there will be 10 log files. Should I consider the Acc@1 from the latest log as the FAA metric? However, this seems to be inconsistent with the results output in the console.
2. Secondly, it seems that I'm unable to run the code on 5 datasets. Could you please tell me the reason for this?
The text was updated successfully, but these errors were encountered:
Hello,
Thank you very much for sharing your wonderful work.
I have two question: 1. When I finish running the experiment, I can see many log files. For example, when running 10 tasks on CIFAR100, there will be 10 log files. Should I consider the Acc@1 from the latest log as the FAA metric? However, this seems to be inconsistent with the results output in the console.
2. Secondly, it seems that I'm unable to run the code on 5 datasets. Could you please tell me the reason for this?
The text was updated successfully, but these errors were encountered: