Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

reuslt #2

Open
Tiandishihua opened this issue Nov 22, 2023 · 1 comment
Open

reuslt #2

Tiandishihua opened this issue Nov 22, 2023 · 1 comment

Comments

@Tiandishihua
Copy link

Why did I generate poor mesh results using the weights you provided?

@X-zhangyang
Copy link
Collaborator

X-zhangyang commented Nov 29, 2023

You can easily obtain results similar to sample.
The unsatisfactory mesh quality may be attributed to the following reasons:

  1. The results in paper are obtained using Renderpeople as the training dataset, which is a high-quality commercial dataset. Due to the limitations of the dataset, we are unable to provide pre-trained weights using Renderpeople. The released pre-trained weights are trained on non-commercial datasets such as Thuman and custom-humans. The quality of the training dataset directly affects the final results of the model.
  2. Similar to the 2D generator, failures may occur when sampling from the Gaussian distribution. We provide an alternative pre-trained weights v2 that you can try for more cases.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants