Awesome
Kaggle CIFAR-10
Code for CIFAR-10 competition. http://www.kaggle.com/c/cifar-10
Summary
| Description | |
|---|---|
| Data Augmentation | cropping, scaling and horizontal reflection. see lib/data_augmentation.lua |
| Preprocessing | Global Contrast Normalization (GCN) and ZCA whitening. see lib/preprocessing.lua |
| Model | Network In Network (NIN). see nin_model.lua |
| Training Time | 30 hours on GTX760. |
| Prediction Time | 5 hours on GTX760. |
| Result | 0.92210 in public leaderboard. |
Developer Environment
- Ubuntu 14.04
- LuaJit/Torch7 latest
- 32GB RAM
- CUDA environment (GTX760 or more higher GPU)
Installation
Install CUDA (on Ubuntu 14.04):
apt-get install nvidia-331
apt-get install nvidia-cuda-toolkit
Install Torch7:
curl -s https://raw.githubusercontent.com/torch/ezinstall/master/install-all | bash
Install(or upgrade) dependency packages:
luarocks install torch
luarocks install nn
luarocks install cutorch
luarocks install cunn
Checking CUDA environment
th cuda_test.lua
Please check your Torch7/CUDA environment when this code fails.
Convert dataset
Please place the data files into a subfolder ./data.
ls ./data
test train trainLabels.csv
- th convert_data.lua
Local testing
th validate.lua
dataset:
| train | test |
|---|---|
| 1-40000 | 40001-50000 |
Generating the submission.txt
th train.lua
th predict.lua
Figure
data augmentation + preprocessing
References
- Alex Krizhevsky, Ilya Sutskever, Geoffrey E. Hinton, "ImageNet Classification with Deep Convolutional Neural Networks"
- Min Lin, Qiang Chen, Shuicheng Yan, "Network In Network"