When I run the colab notebook i got below error.
RuntimeError: The expanded size of the tensor (165) must match the existing size (231) at non-singleton dimension 1. Target sizes: [32, 165, 80]. Tensor sizes: [32, 231, 1]
I think this raise because of the target tensor and input tensors are not in same dimension.
Because that decoder output stop dynamically.Anyone can help ?
torch.Size([32, 165, 80])
torch.Size([32, 231, 80])
