search

KHB1698 / DC-GCT

20 stars 1 forks source link

The problem of multi frame training #3

Open Wzh-42 opened 7 months ago

Wzh-42 commented 7 months ago

Thanks for your work! when I train: python main.py --train --frames 9 or more The assertionError is as follows： “ Traceback (most recent call last): File "main.py", line 139, in loss = train(opt, actions, train_dataloader, model, optimizer, epoch) File "main.py", line 25, in train return step('train', opt, actions, train_loader, model, optimizer, epoch) File "main.py", line 48, in step loss = mpjpe_cal(output_3D, out_target) File "/root/autodl-tmp/DC-GCT/common/utils.py", line 21, in mpjpe_cal assert predicted.shape == target.shape AssertionError ”

How should I handle it？Thanks!

KHB1698 commented 7 months ago

Sorry, we do not provide multi-frame training for the time being. But for your problem, you can download our latest version again, we are using weighted_mpjpe loss function. You can also comment out the line 'assert predict.shape == target.shape', which might solve your confusion.