problem

[cs19469]

swin_tiny_grpb None False Setting backbone: fragments_backbone

Segmentation fault

When i run the vaq.py, somenthing like this

[teowu]

Can you tell me your hardware and software versions? You are not the only one who meet the problem, and I am trying to solve this together for you.

[cs19469]

Thank you for your reply. my device is: Tesla K80, cuda 10.2, python3.8, I have no idea about this, but it works now!

[woshin]

there is something wrong with decord and torch packages.
crash at evaluator = DiViDeAddEvaluator(**opt["model"]["args"]).to(args.device)

simplest solution is import torch first instead of decord. it works for me, but I have not found the root cause)