The original neural network from Torch7, nn, contains stable and widely used modules. 'nnx' contains more experimental, unproven modules, and optimizations. Modules that become stable and which are proven useful make their way into 'nn' (some already have).
This section includes documentation for the following objects:
- SoftMaxTree : a hierarchical log-softmax Module;
- TreeNLLCriterion : a negative log-likelihood Criterion for the SoftMaxTree;
- PushTable (and PullTable) : extracts a table element and inserts it later in the network;
- MultiSoftMax : performs a softmax over the last dimension of a 2D or 3D input;
- SpatialReSampling : performs bilinear resampling of a 3D or 4D input image;
- Recurrent : a generalized recurrent neural network container;
The constructor takes 2 mandatory and 4 optional arguments :
inputSize: the number of units in the input embedding representation;hierarchy: a Tensor mapping oneparent_idto manychild_id(a tree);rootId: a number identifying the root node in the hierarchy. Defaults to-1;accUpdate: when the intent is to usebackwardUpdateoraccUpdateGradParameters, set this to true to save memory. Defaults to false;static: when true (the defualt), returns parameters with keys that don't change from batch to batch;verbose: prints some additional information concerning the hierarchy during construction.
The forward method returns an output Tensor of size 1D, while
backward returns a table {gradInput, gradTarget}. The second
variable is just a Tensor of zeros , such that the targets can be
propagated through Containers
like ParallelTable.
> input = torch.randn(5,10)
> target = torch.IntTensor{20,24,27,10,12}
> gradOutput = torch.randn(5)
> root_id = 29
> input_size = 10
> hierarchy = {
>> [29]=torch.IntTensor{30,1,2}, [1]=torch.IntTensor{3,4,5},
>> [2]=torch.IntTensor{6,7,8}, [3]=torch.IntTensor{9,10,11},
>> [4]=torch.IntTensor{12,13,14}, [5]=torch.IntTensor{15,16,17},
>> [6]=torch.IntTensor{18,19,20}, [7]=torch.IntTensor{21,22,23},
>> [8]=torch.IntTensor{24,25,26,27,28}
>> }
> smt = nn.SoftMaxTree(input_size, hierarchy, root_id)
> smt:forward{input, target}
-3.5186
-3.8950
-3.7433
-3.3071
-3.0522
[torch.DoubleTensor of dimension 5]
> smt:backward({input, target}, gradOutput)
{
1 : DoubleTensor - size: 5x10
2 : IntTensor - size: 5
}
An example utilizing the above SoftMaxTree Module
and a Linear Module demonstrates how the PushTable can be used to
forward the target Tensor without any other
Table Modules:
> mlp = nn.Sequential()
> linear = nn.Linear(50,100)
> push = nn.PushTable(2)
> pull = push:pull(2)
> mlp:add(push)
> mlp:add(nn.SelectTable(1))
> mlp:add(linear)
> mlp:add(pull)
> mlp:add(smt) --smt is a SoftMaxTree instance
> mlp:forward{input, target} -- input and target are defined above
-3.5186
-3.8950
-3.7433
-3.3071
-3.0522
[torch.DoubleTensor of dimension 5]
> mlp:backward({input, target}, gradOutput) -- so is gradOutput
{
1 : DoubleTensor - size: 5x10
2 : IntTensor - size: 5
}The above code is equivalent to the following:
> mlp2 = nn.Sequential()
> para = nn.ParallelTable()
> para:add(linear)
> para:add(nn.Identity())
> mlp2:add(para)
> mlp2:add(smt)
> mlp2:forward{input, target}
-3.5186
-3.8950
-3.7433
-3.3071
-3.0522
[torch.DoubleTensor of dimension 5]
> mlp2:backward({input, target}, gradOutput)
{
1 : DoubleTensor - size: 5x10
2 : IntTensor - size: 5
}In some cases, this can simplify the digraph of Modules. Note that a PushTable can be associated to many PullTables, but each PullTable is associated to only one PushTable.
### MultiSoftMax ### This Module takes 2D or 3D input and performs a softmax over the last dimension. It uses the existing [SoftMax](https://github.com/torch/nn/blob/master/doc/transfer.md#nn.SoftMax) CUDA/C code to do so such that the Module can be used on both GPU and CPU. This can be useful for [keypoint detection](https://github.com/nicholas-leonard/dp/blob/master/doc/facialkeypointstutorial.md#multisoftmax). ### SpatialReSampling ### Applies a 2D re-sampling over an input image composed of several input planes (or channels, colors). The input tensor in `forward(input)` is expected to be a 3D or 4D tensor of size : `[batchSize x] nInputPlane x width x height`. The number of output planes will be the same as the number of input planes.The re-sampling is done using bilinear interpolation.
For a simple nearest-neihbor upsampling, use nn.SpatialUpSampling(),
and for a simple average-based down-sampling, use
nn.SpatialDownSampling().
If the input image is a 3D tensor of size nInputPlane x height x width,
the output image size will be nInputPlane x oheight x owidth where
owidth and oheight are given to the constructor.
Instead of owidth and oheight, one can provide rwidth and rheight,
such that owidth = iwidth*rwidth and oheight = iheight*rheight.
As an example, we can run the following code on the famous Lenna image:
require 'image'
require 'nnx'
input = image.loadPNG('doc/image/Lenna.png')
l = nn.SpatialReSampling{owidth=150,oheight=150}
output = l:forward(input)
image.save('doc/image/Lenna-150x150-bilinear.png', output)The input:
The re-sampled output:
- Torch7 (www.torch.ch)
- Install Torch7 (refer to its own documentation).
- clone this project into dev directory of Torch7.
- Rebuild torch, it will include new projects too.
First run torch, and load nnx:
$ torch> require 'nnx'Once loaded, tab-completion will help you navigate through the library (note that most function are added directly to nn):
> nnx. + TAB
...
> nn. + TABIn particular, it's good to verify that all modules provided pass their tests:
> nnx.test_all()
> nnx.test_omp()DEPRECATED July 6th, 2015. Use rnn instead.