Skip to content

Code for Second Annual Data Science Bowl. 16th place.

Notifications You must be signed in to change notification settings

mehrtash/kaggle-ndsb2

 
 

Repository files navigation

Second Annual Data Science Bowl

Code for Second Annual Data Science Bowl. 16th place.

Sumamry

A Hybrid Deep Neural Network using CNN and MLP.

It is used only 3 SAX slices to predict the actual volume, is not used segmentation techniques, is not needed hand-labelings.

neural network

Developer Environment

  • Ubuntu 14.04
  • 12GB RAM
  • GPU & CUDA (I used EC2 g2.2xlarge instance)
  • Torch7
  • Ruby
  • dicom (rubygems)
  • graphicsmagick (luarocks)

Installation

Install CUDA and Torch7 first. See NVIDIA CUDA Getting Started Guide for Linux and Getting started with Torch.

sudo apt-get install libgraphicsmagick-dev ruby rubygems
sudo gem install dicom
luarocks install graphicsmagick

Data

Place the data files into a subfolder ./data.

% ls ./data
test  train  train.csv  validate  validate.csv

For validation set

./run_all.sh

For test set

./run_all_test.sh

NOTICE: I used 8 g2.xlarge instances to execute this script. See comments in ./run_all_test.sh.

About

Code for Second Annual Data Science Bowl. 16th place.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Lua 92.1%
  • Shell 6.9%
  • Ruby 1.0%