Building efficient models with MXNet

The MXNet deep learning framework allows you to build efficient deep learning models in Python. Next to Python, it also let you build models in popular languages as R, Scala, and Julia. Apache MXNet is supported by Amazon and Baidu, amongst others. MXNet has proven to be fast in benchmarks and it supports GPU and multi-GPU usages. By using lazy evaluation, MXNet is able to automatically execute operations in parallel. Furthermore, the MXNet frameworks uses a symbolic interface, called Symbol. This simplifies building neural network architectures.

How to do it...

1. To install MXNet on Ubuntu with GPU support, we can use the following command in the terminal:

pip install mxnet-cu80==0.11.0

For other platforms and non-GPU support, have a look at https://mxnet.incubator.apache.org/get_started/install.html.

2. Next, we are ready to import mxnet in our Python environment:

import mxnet as mx

3. We create some simple dummy data that we assign to the GPU and CPU:

import numpy as np
x_input = mx.nd.empty((1, 5), mx.gpu())
x_input[:] = np.array([[1,2,3,4,5]], np.float32)

y_input = mx.nd.empty((1, 5), mx.cpu())
y_input[:] = np.array([[10, 15, 20, 22.5, 25]], np.float32)

4. We can easily copy and adjust the data. Where possible MXNet will automatically execute operations in parallel:

x_input
w_input = x_input
z_input = x_input.copyto(mx.cpu())
x_input += 1
w_input /= 2
z_input *= 2

5. We can print the output as follows:

print(x_input.asnumpy())
print(w_input.asnumpy())
print(z_input.asnumpy())

6. If we want to feed our data to a model, we should create an iterator first:

batch_size = 1
train_iter = mx.io.NDArrayIter(x_input, y_input, batch_size, shuffle=True, data_name='input', label_name='target')

7. Next, we can create the symbols for our model:

X = mx.sym.Variable('input')
Y = mx.symbol.Variable('target')
fc1 = mx.sym.FullyConnected(data=X, name='fc1', num_hidden = 5)
lin_reg = mx.sym.LinearRegressionOutput(data=fc1, label=Y, name="lin_reg")

8. Before we can start training, we need to define our model:

model = mx.mod.Module(
    symbol = lin_reg,
    data_names=['input'], 
    label_names = ['target']
)

9. Let's start training:

model.fit(train_iter,
            optimizer_params={'learning_rate':0.01, 'momentum': 0.9},
            num_epoch=100,
            batch_end_callback = mx.callback.Speedometer(batch_size, 2))

10. To use the trained model for prediction we:

model.predict(train_iter).asnumpy()

We've shortly introduced the MXNet framework. In this introduction, we've demonstrated how easily one can assign variables and computations to a CPU or GPU and how to use the Symbol interface. However, there is much more to explore and the MXNet is a powerful framework for building flexible and efficient deep learning models.