search

XingangPan / GAN2Shape

Code for GAN2Shape (ICLR2021 oral)
https://arxiv.org/abs/2011.00844
MIT License
571 stars 104 forks source link

How to do the prediction? #19

Open 30qwq opened 3 years ago

30qwq commented 3 years ago

The code seems to train an input image and evaluate the reconstruction result one by one. I wonder that can the model do the prediction of reconstructing 3D objects of untrained images? If so, how to do that?

Thanks.

XingangPan commented 3 years ago

@30qwq Our approach by design is an online approach, i.e., it trains on each testing image. You just need to perform GAN-inversion to get the latent code, and then apply our approach.