Error when i tried to train rnn

carpedm20 / ENAS-pytorch

PyTorch implementation of "Efficient Neural Architecture Search via Parameters Sharing"

Apache License 2.0

2.69k stars 492 forks source link

Error when i tried to train rnn #22

Open Olivia-xu opened 6 years ago

Olivia-xu commented 6 years ago

when i tried to train rnn with the command python main.py --network_type rnn --dataset ptb --controller_optim adam --controller_lr 0.00035 \ --shared_optim sgd --shared_lr 20.0 --entropy_coeff 0.0001 one error occures at util.detach(h), it seems that when h is a tensor, this function will be infinitely recursive, causing an error. Can you tell me why this is happening?

d-zhou12 commented 6 years ago

the version of pytorch is not 0.4.0, you need 0.3.1 or other

yuezha01 commented 6 years ago

Thanks @d-zhou12 , I had the same error as @Olivia-xu. Pytorch 0.3.1 works for me.

NdaAzr commented 5 years ago

It's giving me a different error which is File "main.py", line 30, in main trnr = trainer.Trainer(args, dataset) File "D:\Neda\Pytorch\NAS\trainer.py", line 160, in __init__ self.build_model() File "D:\Neda\Pytorch\NAS\trainer.py", line 182, in build_model self.shared = models.RNN(self.args, self.dataset) AttributeError: module 'models' has no attribute 'RNN'

My pytorch version is 1.1.0. Could you please point me to the right direction?