Kind inquiry about the dataset VGGFace2

[pxzhang94]

the link of VGGFace 2 (https://www.robots.ox.ac.uk/~vgg/data/vgg_face2/) is not available, is it possible there is another link we can access

[seymayucer]

hello, you can download many datasets from here: https://github.com/ZhaoJ9014/face.evoLVe.PyTorch, for 112x112 aligned images of VGGFace2 you can use this link: https://drive.google.com/file/d/1N7QEEQZPJ2s5Hs34urjseFwIoPVSmn4r/view

[pxzhang94]

Hi Seyma, Thanks for your kind reply, it is a very useful repo for my research. However, I'm studying on the fairness of DL models, I wonder if it's possible to get the dataset VGGFace2 1200. Thanks in advance! Have a good day! Zhang Peixin Ph.D Candidate College of Computer Science and Technology, Zhejiang University +86 18868106685  

…

------------------ 原始邮件 ------------------ 发件人: "seymayucer/VGG1200-Races" ***@***.***>; 发送时间: 2021年3月22日(星期一) 晚上6:26 ***@***.***>; ***@***.******@***.***>; 主题: Re: [seymayucer/VGG1200-Races] Kind inquiry about the dataset VGGFace2 (#1) hello, you can download many datasets from here: https://github.com/ZhaoJ9014/face.evoLVe.PyTorch, for 112x112 aligned images of VGGFace2 you can use this link: https://drive.google.com/file/d/1N7QEEQZPJ2s5Hs34urjseFwIoPVSmn4r/view — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub, or unsubscribe.

[seymayucer]

Hello, you can simply take images from the datasets by using VGGFace2_1200_Train_Subjects_Metadata.csv.But if you are asking for vggface2-1200-races then you can generate images using the pre-trained models.

[pxzhang94]

Sorry to bother you again, is it possible to share the trained face recognition models (ResNet 100) with batchnorm-dropout-fc-batchnorm-softmax before and after applying the race transfer? Zhang Peixin College of Computer Science and Technology, Zhejiang University  

…

------------------ 原始邮件 ------------------ 发件人: "seymayucer/VGG1200-Races" ***@***.***>; 发送时间: 2021年3月23日(星期二) 凌晨3:24 ***@***.***>; ***@***.******@***.***>; 主题: Re: [seymayucer/VGG1200-Races] Kind inquiry about the dataset VGGFace2 (#1) Hello, you can simply take images from the datasets by using VGGFace2_1200_Train_Subjects_Metadata.csv.But if you are asking for vggface2-1200-races then you can generate images using the pre-trained models. — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub, or unsubscribe.