search

carpedm20 / ENAS-pytorch

PyTorch implementation of "Efficient Neural Architecture Search via Parameters Sharing"
Apache License 2.0
2.69k stars 492 forks source link

RuntimeError: saved_forINTERNAL ASSERT FAILED at #57

Closed Shn9909 closed 1 year ago

Shn9909 commented 1 year ago

I encountered this strange error. Here is the output, thank you.

Traceback (most recent call last): File "D:/xiangmu/ENAS-pytorch-master/main.py", line 56, in main(args) File "D:/xiangmu/ENAS-pytorch-master/main.py", line 35, in main trnr.train() File "D:\xiangmu\ENAS-pytorch-master\trainer.py", line 223, in train self.train_shared(dag=dag) File "D:\xiangmu\ENAS-pytorch-master\trainer.py", line 314, in train_shared loss.backward() File "C:\Users\sunhaonan.conda\envs\enas\lib\site-packages\torch_tensor.py", line 307, in backward torch.autograd.backward(self, gradient, retain_graph, create_graph, inputs=inputs) File "C:\Users\sunhaonan.conda\envs\enas\lib\site-packages\torch\autograd__init__.py", line 156, in backward allow_unreachable=True, accumulate_grad=True) # allow_unreachable flag RuntimeError: saved_forINTERNAL ASSERT FAILED at "..\torch\csrc\autograd\saved_variable.cpp":133, please report a bug to PyTorch. No grad_fn for non-leaf saved tensor

Gangle-Sun commented 1 year ago

how do you solve this problem? I meet the same error