Notebook

In [ ]:

%matplotlib inline

nn package¶

We’ve redesigned the nn package, so that it’s fully integrated with autograd. Let's review the changes.

Replace containers with autograd:

You no longer have to use Containers like ``ConcatTable``, or modules like ``CAddTable``, or use and debug with nngraph. We will seamlessly use autograd to define our neural networks. For example, * ``output = nn.CAddTable():forward({input1, input2})`` simply becomes ``output = input1 + input2`` * ``output = nn.MulConstant(0.5):forward(input)`` simply becomes ``output = input * 0.5``

State is no longer held in the module, but in the network graph:

Using recurrent networks should be simpler because of this reason. If you want to create a recurrent network, simply use the same Linear layer multiple times, without having to think about sharing weights. .. figure:: /_static/img/torch-nn-vs-pytorch-nn.png :alt: torch-nn-vs-pytorch-nn torch-nn-vs-pytorch-nn

Simplified debugging:

Debugging is intuitive using Python’s pdb debugger, and **the debugger and stack traces stop at exactly where an error occurred.** What you see is what you get.

Example 1: ConvNet¶

Let’s see how to create a small ConvNet.

All of your networks are derived from the base class nn.Module:

In the constructor, you declare all the layers you want to use.
In the forward function, you define how your model is going to be run, from input to output

In [ ]:

importtorchfromtorch.autogradimportVariableimporttorch.nnasnnimporttorch.nn.functionalasFclassMNISTConvNet(nn.Module):def__init__(self):# this is the place where you instantiate all your modules# you can later access them using the same names you've given them in# heresuper(MNISTConvNet,self).__init__()self.conv1=nn.Conv2d(1,10,5)self.pool1=nn.MaxPool2d(2,2)self.conv2=nn.Conv2d(10,20,5)self.pool2=nn.MaxPool2d(2,2)self.fc1=nn.Linear(320,50)self.fc2=nn.Linear(50,10)# it's the forward function that defines the network structure# we're accepting only a single input in here, but if you want,# feel free to use moredefforward(self,input):x=self.pool1(F.relu(self.conv1(input)))x=self.pool2(F.relu(self.conv2(x)))# in your model definition you can go full crazy and use arbitrary# python code to define your model structure# all these are perfectly legal, and will be handled correctly# by autograd:# if x.gt(0) > x.numel() / 2:# ...## you can even do a loop and reuse the same module inside it# modules no longer hold ephemeral state, so you can use them# multiple times during your forward pass# while x.norm(2) < 10:# x = self.conv1(x)x=x.view(x.size(0),-1)x=F.relu(self.fc1(x))x=F.relu(self.fc2(x))returnx

Let's use the defined ConvNet now. You create an instance of the class first.

In [ ]:

net=MNISTConvNet()print(net)

Note

``torch.nn`` only supports mini-batches The entire ``torch.nn`` package only supports inputs that are a mini-batch of samples, and not a single sample.

For example, ``nn.Conv2d`` will take in a 4D Tensor of ``nSamples x nChannels x Height x Width``. If you have a single sample, just use ``input.unsqueeze(0)`` to add a fake batch dimension.</p></div>

Create a mini-batch containing a single sample of random data and send the sample through the ConvNet.

In [ ]:

input=Variable(torch.randn(1,1,28,28))out=net(input)print(out.size())

Define a dummy target label and compute error using a loss function.

In [ ]:

target=Variable(torch.LongTensor([3]))loss_fn=nn.CrossEntropyLoss()# LogSoftmax + ClassNLL Losserr=loss_fn(out,target)err.backward()print(err)

The output of the ConvNet out is a Variable. We compute the loss using that, and that results in err which is also a Variable. Calling .backward on err hence will propagate gradients all the way through the ConvNet to it’s weights

Let's access individual layer weights and gradients:

In [ ]:

print(net.conv1.weight.grad.size())

In [ ]:

print(net.conv1.weight.data.norm())# norm of the weightprint(net.conv1.weight.grad.data.norm())# norm of the gradients

Forward and Backward Function Hooks¶

We’ve inspected the weights and the gradients. But how about inspecting / modifying the output and grad_output of a layer?

We introduce hooks for this purpose.

You can register a function on a Module or a Variable. The hook can be a forward hook or a backward hook. The forward hook will be executed when a forward call is executed. The backward hook will be executed in the backward phase. Let’s look at an example.

We register a forward hook on conv2 and print some information

In [ ]:

defprintnorm(self,input,output):# input is a tuple of packed inputs# output is a Variable. output.data is the Tensor we are interestedprint('Inside '+self.__class__.__name__+' forward')print('')print('input: ',type(input))print('input[0]: ',type(input[0]))print('output: ',type(output))print('')print('input size:',input[0].size())print('output size:',output.data.size())print('output norm:',output.data.norm())net.conv2.register_forward_hook(printnorm)out=net(input)

We register a backward hook on conv2 and print some information

In [ ]:

defprintgradnorm(self,grad_input,grad_output):print('Inside '+self.__class__.__name__+' backward')print('Inside class:'+self.__class__.__name__)print('')print('grad_input: ',type(grad_input))print('grad_input[0]: ',type(grad_input[0]))print('grad_output: ',type(grad_output))print('grad_output[0]: ',type(grad_output[0]))print('')print('grad_input size:',grad_input[0].size())print('grad_output size:',grad_output[0].size())print('grad_input norm:',grad_input[0].data.norm())net.conv2.register_backward_hook(printgradnorm)out=net(input)err=loss_fn(out,target)err.backward()

A full and working MNIST example is located here https://github.com/pytorch/examples/tree/master/mnist

Example 2: Recurrent Net¶

Next, let’s look at building recurrent nets with PyTorch.

Since the state of the network is held in the graph and not in the layers, you can simply create an nn.Linear and reuse it over and over again for the recurrence.

In [ ]:

classRNN(nn.Module):# you can also accept arguments in your model constructordef__init__(self,data_size,hidden_size,output_size):super(RNN,self).__init__()self.hidden_size=hidden_sizeinput_size=data_size+hidden_sizeself.i2h=nn.Linear(input_size,hidden_size)self.h2o=nn.Linear(hidden_size,output_size)defforward(self,data,last_hidden):input=torch.cat((data,last_hidden),1)hidden=self.i2h(input)output=self.h2o(hidden)returnhidden,outputrnn=RNN(50,20,10)

A more complete Language Modeling example using LSTMs and Penn Tree-bank is located here <https://github.com/pytorch/examples/tree/master/word\_language\_model>_

PyTorch by default has seamless CuDNN integration for ConvNets and Recurrent Nets

In [ ]:

loss_fn=nn.MSELoss()batch_size=10TIMESTEPS=5# Create some fake databatch=Variable(torch.randn(batch_size,50))hidden=Variable(torch.zeros(batch_size,20))target=Variable(torch.zeros(batch_size,10))loss=0fortinrange(TIMESTEPS):# yes! you can reuse the same network several times,# sum up the losses, and call backward!hidden,output=rnn(batch,hidden)loss+=loss_fn(output,target)loss.backward()