Hello, I see that your work points out that you can train without using 3D priors, but how do you train the pix2pix network, how to train the discriminator without a deformation map and position map, if you only train the generator, I tried it in my own code, and there will be a series of problems
Hello, I see that your work points out that you can train without using 3D priors, but how do you train the pix2pix network, how to train the discriminator without a deformation map and position map, if you only train the generator, I tried it in my own code, and there will be a series of problems