Beronx86

I thin I fixed the bug. https://github.com/mila-udem/blocks/pull/1113 Because in graph/**init**.py :ComputationGraph.scan_variables supposed that there is no nested recurrent classes.

I think I fixed this bug [https://github.com/mila-udem/fuel/pull/327]

@abergeron Hi, I did as what you said and modified [line 73 in sever.py](https://github.com/mila-udem/fuel/blob/master/fuel/server.py#L73) to socket.recv(copy=False). It works.

@dwf HI， thanks for your response. The brick names are not unique. But enlighted by you, I modified the `Selector` code to select auxiliary variables, and rename the variable to...

I found `GradientDescent.total_gradient_norm` and `GradientDescent.total_step_norm` can be used to monitor the gradient.

The test error is 'ImportError: No module named matplotlib'. I think it is OK.

Ok, I'll write the test case.

In the original code, the `recurrent` method would require `initial_states` function whether there is `recurrent states` or not. If the `recurrent states` is left empty, an error would occur at...

You may produce the error with the following code. The error occurs when the class does not contain a recurrent method named `apply` ``` Python import numpy import theano from...