RuntimeError: Sizes of tensors must match except in dimension 0. Got 0 and 300 (The offending index is 0)

[Hezey]

When I was training my own data set with DN-DETR, I encountered the following problems, how do I solve them? Who can tell me, please? @HaoZhang534 @SangbumChoi @FengLi-ust @LYMDLUT

Traceback (most recent call last): File "D:\JetBrains\Pycharm_Project\Learn_Deep_Learning\DN-DETR\main.py", line 443, in main(args) File "D:\JetBrains\Pycharm_Project\Learn_Deep_Learning\DN-DETR\main.py", line 369, in main train_stats = train_one_epoch( File "D:\JetBrains\Pycharm_Project\Learn_Deep_Learning\DN-DETR\engine.py", line 53, in train_one_epoch outputs, mask_dict = model(samples, dn_args=dn_args) File "D:\Anaconda3\envs\DN_DETR\lib\site-packages\torch\nn\modules\module.py", line 889, in _call_impl result = self.forward(*input, *kwargs) File "D:\JetBrains\Pycharm_Project\Learn_Deep_Learning\DN-DETR\models\dn_dab_deformable_detr\dab_deformable_detr.py", line 206, in forward prepare_for_dn(dn_args, tgt_all_embed, refanchor, src.size(0), self.training, self.num_queries, self.num_classes, File "D:\JetBrains\Pycharm_Project\Learn_Deep_Learning\DN-DETR\models\dn_dab_deformable_detr\dn_components.py", line 70, in prepare_for_dn tgt = torch.cat([tgt_weight, indicator0], dim=1) + label_enc.weight[0][0]torch.tensor(0).cuda() RuntimeError: Sizes of tensors must match except in dimension 0. Got 0 and 300 (The offending index is 0)

[SangbumChoi]

I think it is the problem either your dataset wasn't load properly or lack of defining num_classes.

[Hezey]

I think it is the problem either your dataset wasn't load properly or lack of defining num_classes.

I've already set num_classes. There are some errors in torch.cat. How can I fix the code to solve the problem.

[SangbumChoi]

try to print the individual shape of each variable and debug it. Like i just said it seems to appear that tgt_weight is abnormal shape which is related to dataset preparation. Or did you modified self.two_stage and self.use_dab constraint?

[Hezey]

I didn't modified self.two_stage and self.use_dab. I'm thinking how can I modify and add a few lines of code to solve the problem