zcq15/spdet

spdet's Issues

Code Error:

Traceback (most recent call last):
File "demo.py", line 37, in
_,output = model(img)
File "/home/mayiwen/.conda/envs/py38/lib/python3.8/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl
return self._call_impl(*args, **kwargs)
File "/home/mayiwen/.conda/envs/py38/lib/python3.8/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl
return forward_call(*args, **kwargs)
File "/home/mayiwen/python_code/SPDET/models/spdet/spdet.py", line 54, in forward
_tokens, token = self.embed(equi)
File "/home/mayiwen/.conda/envs/py38/lib/python3.8/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl
return self._call_impl(*args, **kwargs)
File "/home/mayiwen/.conda/envs/py38/lib/python3.8/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl
return forward_call(*args, **kwargs)
File "/home/mayiwen/python_code/SPDET/models/spdet/modules.py", line 403, in forward
x3 = self.embed3(torch.cat([x3,torch.unsqueeze(feats[-1],0)],dim=1))
RuntimeError: Sizes of tensors must match except in dimension 1. Expected size 16 but got size 5 for tensor number 1 in the list.

I'd like to know how to tackle this problem?

About the loss function

Hello. I'd like to know why you use 'S(p') = cos(u')*cos(v')' in your loss function. Could you elaborate it? Thanks!