peschn opened a new issue #7382: GRU not working with layout 'NTC' URL: https://github.com/apache/incubator-mxnet/issues/7382 For bugs or installation issues, please provide the following information. The more information you provide, the more likely people will be able to help you. ## Environment info Operating System: Ubuntu 16.04.2 LTS Package used (Python/R/Scala/Julia): Python MXNet version: mxnet-cu80==0.10.1b20170803 If you are using python package, please provide Python version and distribution: python 2.7.13, used with anaconda 4.2.23 R `sessionInfo()`: ## Error Message: Please paste the full error message, including stack trace. ??? Traceback (most recent call last): File "simply/mx/test_rnn.py", line 14, in <module> output, hn = layer(input, h0) File "/home/huxley/dev/miniconda2/envs/mxnet/lib/python2.7/site-packages/mxnet/gluon/block.py", line 251, in __call__ return self.forward(*args) File "/home/huxley/dev/miniconda2/envs/mxnet/lib/python2.7/site-packages/mxnet/gluon/rnn/rnn_layer.py", line 162, in forward str(info['shape']), str(state.shape))) ValueError: Invalid recurrent state shape. Expecting (2, 2L, 100), got (2L, 8L, 100L). ??? ## Minimum reproducible example if you are using your own code, please provide a short script that reproduces the error. ## Steps to reproduce or if you are running standard examples, please provide the commands you have run that lead to the error. ???python import mxnet as mx batch_size = 8 timesteps = 7 nhidden = 100 layers = 2 nin = 10 layer = mx.gluon.rnn.GRU(nhidden, layers, layout='NTC') layer.initialize() input = mx.nd.random_uniform(shape=(batch_size, timesteps, nin)) h0 = mx.nd.random_uniform(shape=(layers, batch_size, nhidden)) output, hn = layer(input, h0) ??? ## What have you tried to solve it? The only thing that solves it is changing the layout (and the data accordingly) to 'TNC'. ---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org
With regards, Apache Git Services
