question

[zhangkai07]

Starting Training Loop...
Traceback (most recent call last):
File "D:\project_python\doctor_project\LYC\20241130\222.py", line 42, in
wt.fit(X=X, lr=1e-2, num_epochs=10)
File "C:\ProgramData\miniconda3\envs\py10\lib\site-packages\awave\transform.py", line 86, in fit
trainer(train_loader, epochs=num_epochs)
File "C:\ProgramData\miniconda3\envs\py10\lib\site-packages\awave\utils\train.py", line 84, in call
mean_epoch_loss = self._train_epoch(train_loader, epoch)
File "C:\ProgramData\miniconda3\envs\py10\lib\site-packages\awave\utils\train.py", line 110, in train_epoch
iter_loss = self.train_iteration(data)
File "C:\ProgramData\miniconda3\envs\py10\lib\site-packages\awave\utils\train.py", line 158, in train_iteration
loss.backward()
File "C:\ProgramData\miniconda3\envs\py10\lib\site-packages\torch_tensor.py", line 525, in backward
torch.autograd.backward(
File "C:\ProgramData\miniconda3\envs\py10\lib\site-packages\torch\autograd_init.py", line 260, in backward
grad_tensors = make_grads(tensors, grad_tensors, is_grads_batched=False)
File "C:\ProgramData\miniconda3\envs\py10\lib\site-packages\torch\autograd_init.py", line 141, in _make_grads
raise RuntimeError(msg)
RuntimeError: grad can be implicitly created only for real scalar outputs but got torch.complex64