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.

Sign up for GitHub

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

Jump to bottom

Tensor full of nan during the train of vrcnet #29

Open

acavalli1996 opened this issue Jun 2, 2022 · 0 comments

acavalli1996 commented Jun 2, 2022 •

edited

Hy,
thanks for your work!
During the training of the vrcnet, the following problem occurs:

Especially, the problem occurs whithin the function edge_preserving_sampling, which is called inside the forward of the SA_SKN_Res_encoder:

ds_features, p_idx, pn_idx, ds_points = edge_preserve_sampling(features, points, sample_num, k)

and the feature tensors become full nan tensors.
Is there a way to solve this problem? Did I do something wrong during the training?

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment