dmlc / mxnet-gtc-tutorial

MXNet Tutorial for NVidia GTC 2016.

Distributed Machine Learning Common Codebase

DMLC-Core is the backbone library to support all DMLC projects, offers the bricks to build efficient and scalable distributed machine learning libraries.

Developer Channel

What's New

• Note on Parameter Module for Machine Learning

Contents

• Documentation and Tutorials
• Contributing

Known Issues

• RecordIO format is not portable across different processor endians. So it is not possible to save RecordIO file on a x86 machine and then load it on a SPARC machine, because x86 is little endian while SPARC is big endian.

Contributing

Contributing to dmlc-core is welcomed! dmlc-core follows google's C style guide. If you are interested in contributing, take a look at feature wishlist and open a new issue if you like to add something.

• DMLC-Core uses C++11 standard. Ensure that your C++ compiler supports C++11.
• Try to introduce minimum dependency when possible

CheckList before submit code

• Type make lint and fix all the style problems.
• Type make doc and fix all the warnings.

NOTE

deps:

libcurl4-openssl-dev

mxnet-gtc-tutorial's Issues

Error running the deep3d tutorial

Hi,

I'm getting an error running the deep3d tutorial. I get a SoftmaxActivation error - cannot find argument.

Can someone help with this?
thanks

edit: its all good, working now.

Error on hand drawing test of the main tutorial

Hi！

I got good result from the training model on the MNIST dataset. But when I'm trying to test my own hand writing numbers, I find that the code given in the tutorial doesn't seem to work. There are no error reported, but the program quits without an Ipython window showing up as expected.

Did anyone know a fix to this problem?

Process finished with exit code 139 (interrupted by signal 11: SIGSEGV)

I am building a CNN+RNN model for speech recognition. But it seems a lot problems with MXNet's rnn API?
I followed the rnn example and trained my network finally after someone's help.
But when it comes to the predict step, some issues occur:
Here are my codes:

x = mx.io.NDArrayIter(data=Pxxs_arr, label=None, batch_size=10, shuffle=False, last_batch_handle='pad')
pred = self.model.predict(x)

hint: Pxx.shape = (10, 1, x, y)
(batch_size, chnnel, w, h)

Then:
Process finished with exit code 139 (interrupted by signal 11: SIGSEGV)

Since My codes is work pretty well in my single cnn network experienments:
pred = self.model.predict(Pxx)

hint: Pxx is a numpy array with shape(1, 1, x, y)

So, what's the true skills of MXNet's RNN API?

What version of mxnet is this code based on?

As the title.

Tutorial Notebook example out of date and not working with mxnet-0.9.1

1. mx.model.Feedforward is deprecated