-
Notifications
You must be signed in to change notification settings - Fork 406
Description
I've been using PyAV to decode videos with standard code functions. Currently, I'm testing my model with PyTorch, during which I need to read some videos and convert them into latent representations. However, I found that the process consistently hangs on certain specific videos. This issue is reproducible. Strangely, when I use a standalone Python CLI script to read the same video, it works fine.
I downgraded PyAV to version 12.0.0, and the problem disappeared. Could it be that recent updates have introduced a conflict between PyAV and PyTorch?
Unfortunately, I don't have the bandwidth at the moment to provide a minimal reproducible example. I'm not an expert in video encoding/decoding, so I'm hoping someone more familiar with the recent changes could help check if anything in the changelog might be related to this issue.