dataset input

[kobic8]

Hello,
its not clear from the Readme file, not the paper from where do i download the images dataset, and the required format. My first phase is to try to replicate your results, and after that i want to try to train this on my own data set.
Thanks

[valentyn1boreiko]

Hello, its not clear from the Readme file, not the paper from where do i download the images dataset, and the required format. My first phase is to try to replicate your results, and after that i want to try to train this on my own data set. Thanks

Dear @kobic8, thank you for your question!

In the paper we rely on the pre-trained ImageNet models and do not require ImageNet locally. You can specify any images and any targets to generate counterfactuals with the pre-trained models. Please see the reply in Issue 8 for more details.

If you have any further questions, please feel free to reach out!

Best,
Valentyn