about "ResClassifier" in the framework about afn HOT 3 CLOSED

Hello,

1. Our ResClassifier architectures are similar to them in MCD_DA for different datasets.
2. About the setting of our experiments, please contact the first author Ruijia Xu.

from afn.

Hello,

(1) Actually, as you can find in Figure 3(d), we have conducted an ablation study about the "Sensitivity of Embedding Size", you can check there.
(2) As said by Jihan, we followed MCD_DA and applied two fully-connected layers, each has 1000 neurons, as the bottleneck in the classifier on VisDA2017. Considering small sample size of other
benchmarks, we reduced bottleneck to one fully-connected layer.

Thanks again for your interest.

from afn.

OK. Thanks for your reply.

from afn.