explainableml / acvc Goto Github PK

Hello, thank you for your contribution to the Single Source Domain Generalization Institute. I have a question want to ask you, you can compare the precision of the method of M - ADA is how to get, I'm not found in the code for the implementation of this method.

Hello Ilke, thanks for you great work of single source domain generalization "Attention Consistency on Visual Corruptions for Single-Source Domain Generalization". I'm currently doing some new experiments on the same area, but I have some questions about the experiment data.

1. How do you obtain the domain generalization accuracy results of PACS dataset of the baseline model ? Did you use the best generlization result during training epochs or the average accuracy?
2. When testing on the PACS dataset, do you test the whole subset of for example art painting or do you test on the test split of the art painting subset ?
   I'm looking forward to your timely reply.

Hello, I tried to restore ME-ADA method on PACS data set, but I failed. So I would like to ask if you can provide the code to implement ME-ADA on PACS.

Thank you for sharing your marvelous work!

I just want to know what the difference between DomainNet and FullDomainNet. Can I change FullDomainNet into DomainNet? And how?

Thank you again!