This repository has been archived by the owner on Nov 17, 2023. It is now read-only.
-
Notifications
You must be signed in to change notification settings - Fork 6.8k
How to get the symbol of gluon's RNN model by hybridizing #7551
Labels
Comments
RNN is not a HybridBlock so you cannot use it as a children of a HybridBlock. This is currently a limitation due to sym.RNN doesn't support CPU. RNN will be changed to a hybridblock in the future. |
@piiswrong Any ETA for this feature? |
@piiswrong Is this feature available now? |
looking forward to this new feature |
same here. thanks! |
same here. |
is it still a limitation yet? Can someone post a workaround? |
This request is tracked in #10873. The overall project is tracked in https://github.com/apache/incubator-mxnet/projects/11 |
Resolving it since it is now supported - #11482 |
Sign up for free
to subscribe to this conversation on GitHub.
Already have an account?
Sign in.
I follow the instruction of http://gluon.mxnet.io/P14-C05-hybridize.html#, but fail. the error message is:
File "/usr/lib/python2.7/site-packages/mxnet-0.11.0-py2.7.egg/mxnet/gluon/block.py", line 312, in setattr
super(HybridBlock, self).setattr(name, value)
File "/usr/lib/python2.7/site-packages/mxnet-0.11.0-py2.7.egg/mxnet/gluon/block.py", line 177, in setattr
self.register_child(value)
File "/usr/lib/python2.7/site-packages/mxnet-0.11.0-py2.7.egg/mxnet/gluon/block.py", line 327, in register_child
str(block), str(type(block))))
ValueError: Children of HybridBlock must also be HybridBlock, but RNN(100 -> 100, TNC, num_layers=2, dropout=0.2) has type <class 'mxnet.gluon.rnn.rnn_layer.RNN'>. If you are using Sequential, please try HybridSequential instead.
I can not find the HybridSequential RNN in source code of gluon
The text was updated successfully, but these errors were encountered: