i try this swintransformer on deeplabv3 (https://github.com/VainF/DeepLabV3Plus-Pytorch), errors are found:

Exception has occurred: EinopsError
Error while processing rearrange-reduction pattern "b (nw_h w_h) (nw_w w_w) (h d) -> b h (nw_h nw_w) (w_h w_w) d".
Input tensor shape: torch.Size([1, 104, 104, 96]). Additional info: {'h': 3, 'w_h': 7, 'w_w': 7}.
Shape mismatch, can't divide axis of length 104 in chunks of 7

During handling of the above exception, another exception occurred:

File "D:\TangYong\Src\VS\Python\PyTorch\deeplabv3-vainf\network\backbone\swintransformer.py", line 111, in
lambda t: rearrange(t, 'b (nw_h w_h) (nw_w w_w) (h d) -> b h (nw_h nw_w) (w_h w_w) d',
File "D:\TangYong\Src\VS\Python\PyTorch\deeplabv3-vainf\network\backbone\swintransformer.py", line 110, in forward
q, k, v = map(
File "D:\TangYong\Src\VS\Python\PyTorch\deeplabv3-vainf\network\backbone\swintransformer.py", line 32, in forward
return self.fn(self.norm(x), **kwargs)
File "D:\TangYong\Src\VS\Python\PyTorch\deeplabv3-vainf\network\backbone\swintransformer.py", line 22, in forward
return self.fn(x, **kwargs) + x
File "D:\TangYong\Src\VS\Python\PyTorch\deeplabv3-vainf\network\backbone\swintransformer.py", line 149, in forward

thank you for your answer.