Noisy/broken output

[yongbinkwon]

I have previously used this network just fine, but recently when trying to use it, it sometimes produces output that doesn't make sense. I have attached one such image here:

The output is inconsistent based on the input (same image outputs different results), and on some runs the network seems to work well. This seems to affect the first images of a batch the most. Any one else encountered this kind of issue? I'm using an RTX2080TI (over ssh) and cuda 11.2

[Hangz-nju-cuhk]

It might be caused by the inconsistency of pytorch and cuda versions with neural renderer.