Thank you very much for your great work on Adaface. You have also documented the repository very well. I am very grateful for that.
I am trying to fine-tune Adaface with my own data set and then to evaluate it on the YLFW data set.
I have downloaded the dataset faces_webface_112 dataset from Insightface and converted it as you suggested here in order to see, if I could better understand, how the test dataset should be structured and whether I need for the evaluation of Adaface a kind of 'pair list' with images pairs to be compared.
I still can't understand exactly how I should properly structure my dataset folders for evaluation on YLFW.
Could you please give me some more clarity on this?
Thank you very much for your great work on Adaface. You have also documented the repository very well. I am very grateful for that.
I am trying to fine-tune Adaface with my own data set and then to evaluate it on the YLFW data set.
I have downloaded the dataset faces_webface_112 dataset from Insightface and converted it as you suggested here in order to see, if I could better understand, how the test dataset should be structured and whether I need for the evaluation of Adaface a kind of 'pair list' with images pairs to be compared.
I still can't understand exactly how I should properly structure my dataset folders for evaluation on YLFW.
Could you please give me some more clarity on this?
Many thanks in advance!
Best regards