yingqianwang / deoccnet Goto Github PK

when i tried to download compressed LFs in folders using Baidu , i have to register first and when i tried to register an error appeared saying that "Registration from overseas and Hong Kong, Macao and Taiwan regions is not currently supported."
and i want to try the training using LF_original_5 and 15.zip in folders

and i send to the authors and they are not even respond

so what is the solution ?

i face this issue when i try to test your model
torch.Size([1, 512, 122, 90]) torch.Size([1, 512, 121, 90])
Traceback (most recent call last):
File "c:\mostafa_farouk_work\DeOccNet\DeOccNet_codes\test25.py", line 35, in
main(int(1+i))
File "c:\mostafa_farouk_work\DeOccNet\DeOccNet_codes\test25.py", line 28, in main
result = test_(test_loader, SceneIdx)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "c:\mostafa_farouk_work\DeOccNet\DeOccNet_codes\test25.py", line 18, in test_
out = net(data)
^^^^^^^^^
File "C:\Users\Mostafa\AppData\Local\anaconda3\envs\MainPyTorchEnv\Lib\site-packages\torch\nn\modules\module.py", line 1518, in _wrapped_call_impl
return self._call_impl(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Users\Mostafa\AppData\Local\anaconda3\envs\MainPyTorchEnv\Lib\site-packages\torch\nn\modules\module.py", line 1527, in call_impl
return forward_call(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "c:\mostafa_farouk_work\DeOccNet\DeOccNet_codes\model.py", line 41, in forward
buffer_3L = torch.cat((buffer_3L, buffer_3), 1)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
RuntimeError: Sizes of tensors must match except in dimension 1. Expected size 122 but got size 121 for tensor number 1 in the list.

i printed this torch.Size([1, 512, 122, 90]) torch.Size([1, 512, 121, 90]) to ensure about the dimensions of the two tensors and actually the third parameter is different in the two parameters so whet is the solution?