lixiaotong97/DSU

About the experiment results

Baymine opened this issue · 1 comments

Did you use the best results (show in the output log) as your final results? If not, would you please tell me how did you get the results reported in the paper? Thank you.

We use the results of the last checkpoint by default. Considering the dataset is relatively small and thus suffers fluctuations, such as PACS, we show the results with the best performances.
There can be some fluctuations in performances due to the different environments, the relative improvement should be significant regardless of these.

Best of wishes.