Xinyang Li

About metric

5

Hi, can you share the script to reproduce your great metrics in your paper. Also, i want to compare my own model with yours, so it will help a lot...

About the bad results of stargan in your paper.

1

I found that in your paper in page.6, the results of stargan is really poor. I reproduced stargan in celeba-hq and got bad results too. But in stargan paper, it...

In your original paper, the attribute code is mostly exchanged to guide the translation, but in your code, you add using random noise to guide the translation like MUNIT, why?...

Why 4x4 Conv in this network, and where this idea come from

4

In discriminator, style encoder and content encoder, i find 4x4 conv filters. Where this idea came from or i missed something.

Are there some differences between datasets of ELEGANT and StarGAN?

1

I found that your method emphasized that you aligned the landmarks of faces for your training, so is it different from StarGAN in this aspect?

Voxel renderer

3

Will pytorch3d support voxel renderer in the future?

i wonder a better way to make this kind of attn-base generator convergence could you please share some train tricks while you train this architecture ? : )

Can nerfacc support batch-wise ops?

2

In my case, the nerf network may contains B different scenes. Can nerfacc support ray_marching with rays_o, rays_d (shape: B, N, 3) and sigma_fn (B, N, 3 -> B, N,...

Hello. I found that in your paper, the reconstruction loss is defined as L1-norm but in your code it is L2-norm. Why?