Can you provide the pre-trained attributes prediction model to me?

genforce / interfacegan

[CVPR 2020] Interpreting the Latent Space of GANs for Semantic Face Editing

https://genforce.github.io/interfacegan/

MIT License

1.51k stars 281 forks source link

Can you provide the pre-trained attributes prediction model to me? #72

Open wty-ustc opened 3 years ago

wty-ustc commented 3 years ago

In your paper, you claimed This model is trained with multi-task losses to simultaneously predict smile, age, gender, eyeglasses, as well as the 5-point facial landmarks. Can you provide me with this model? Thank you very much!

GaParmar commented 3 years ago

I would also find this very useful

ShenYujun commented 3 years ago

Please try models here.

ucalyptus2 commented 3 years ago

@ShenYujun these were for celebahq, anything for ffhq? Will both work the same ways since both have similar brilliant facial capabilities?

lizaitang commented 2 years ago

Please try models here.

Thanks, I used this classfiers to test the result, but the prediction logits seems different each time, have you met the same issue?

federicoromeo commented 1 year ago

Can anyone provide a code snippet to make those classifiers work? I followed stylegan's and even stylespace's (see my issue for details) but i can't make those work!