I decorate a nngraph model with nn.Recursor to make it a nn.AbstractSequencer class. But when I try to save a train model, the following error raises:
Unwritable object at <?>.<?>.modules.1.<?>.rnn.test.rnntest._tests.Sequencer.._index.rejectWhile..rejectWhile.ipairs
It's very weird because when I test the code on my laptop, it works. But when I run it on a server, it raise the above error when executing torch.save(). Note that I have another torch.save() in front of it, and that works just fine. So Is there any problem when decorating nngraph model with nn.Recursor ?
Hi
I decorate a nngraph model with nn.Recursor to make it a nn.AbstractSequencer class. But when I try to save a train model, the following error raises: Unwritable object at <?>.<?>.modules.1.<?>.rnn.test.rnntest._tests.Sequencer.._index.rejectWhile..rejectWhile.ipairs
It's very weird because when I test the code on my laptop, it works. But when I run it on a server, it raise the above error when executing torch.save(). Note that I have another torch.save() in front of it, and that works just fine. So Is there any problem when decorating nngraph model with nn.Recursor ?
Thanks for your reply