Deep Networks with Stochastic Depth
This repository hosts the Torch 7 code for the paper Deep Networks with Stochastic Depth
available at http://arxiv.org/abs/1603.09382. For now, the code reproduces the results in Figure 3 for CIFAR-10 and CIFAR-100, and Figure 4 left for SVHN. The code for the 1202-layer network is easily modified from the repo
fb.resnet.torch using our provided module for stochastic depth.
Table of Contents
- Torch 7 and CUDA with the basic packages (nn, optim, image, cutorch, cunn).
- cudnn (https://developer.nvidia.com/cudnn) and torch bindings (https://github.com/soumith/cudnn.torch).
- nninit torch package (https://github.com/Kaixhin/nninit);
luarocks install nninitshould do the trick.
- CIFAR-10 and CIFAR-100 datasets in Torch format; this script https://github.com/soumith/cifar.torch should very conveniently handle it for you.
- SVHN dataset in Torch format, available at this website http://torch7.s3-website-us-east-1.amazonaws.com/data/svhn.t7.tgz. Please note that running on SVHN requires roughly 28GB of RAM for dataset loading.
Getting Started on CIFAR-10
git clone https://github.com/yueatsprograms/Stochastic_Depth cd Stochastic_Depth git clone https://github.com/soumith/cifar.torch cd cifar.torch th Cifar10BinToTensor.lua cd .. mkdir results th main.lua -dataRoot cifar.torch/ -resultFolder results/ -deathRate 0.5
th main.lua -dataRoot path_to_data -resultFolder path_to_save -deathRate 0.5
This command runs the 110-layer ResNet on CIFAR-10 with stochastic depth, using linear decay survival probabilities ending in 0.5. The
-device flag allows you to specify which GPU to run on. On our machine with a TITAN X, each epoch takes about 60 seconds, and the program ends with a test error (selected by best validation error) of 5.25%.
The default deathRate is set to 0. This is equivalent to a constant depth network, so to run our baseline, enter:
th main.lua -dataRoot path_to_data -resultFolder path_to_save
On our machine with a TITAN X, each epoch takes about 75 seconds, and this baseline program ends with a test error (selected by best validation error) of 6.41% (see Figure 3 in the paper).
You can run on CIFAR-100 by adding the flag
-dataset cifar100. Our program provides other options, for example, your network depth (
-N), data augmentation (
-augmentation), batch size (
-batchSize) etc. You can change the optimization hyperparameters in the sgdState variable, and learning rate schedule in the the main function. The program saves a file every epoch to
deathRate, which has a table of tuples containing your test and validation errors until that epoch.
The architecture and number of epochs for SVHN used in our paper are slightly different from the code's default, please use the following command if you would like to replicate our result of 1.75% on SVHN:
th main.lua -dataRoot path_to_data -resultFolder path_to_save -dataset svhn -N 25 -maxEpochs 50 -deathRate 0.5
- It is normal to get a +/- 0.2% difference from our reported results on CIFAR-10, and analogously for the other datasets. Networks are initialized differently, and most importantly, the validation set is chosen at random (determined by your seed).
- If you train on SVHN and the model doesn't converge for the first 1600 or so iterations, that's ok, just wait for a little longer.
- Xavier reported that the model is able to converge for him on CIFAR-10 only after he uses the following initalization for Batch Normalization
model:add(cudnn.SpatialBatchNormalization(_dim_):init('weight', nninit.normal, 1.0, 0.002):init('bias', nninit.constant, 0)). We could not replicate the non-convergence and thus won't put this initialization into our code, but recognize that machines (or the versions of Torch installed) might be different.
My email is ys646 at cornell.edu. I'm happy to answer any of your questions, and I'd very much appreciate your suggestions. My academic website is at http://yueatsprograms.github.io.