README.md

TopCoder NCS Challenge: Project template to support https://developer.movidius.com/competition

This directory contains all supporting files needed to generate submissions.zip file, which would then be uploaded to the TopCoder leaderboard for automatic scoring.

Prerequisites

This code example requires that the following components are available:

1. Movidius Neural Compute Stick
2. Movidius Neural Compute SDK

Running the Example

mkdir -p ~/workspace
cd ~/workspace
git clone https://github.com/movidius/ncappzoo
cd ~/workspace/ncappzoo/apps/topcoder_example
make run

If everything went well, you should see an output similar to this:

Downloading training dataset...
...
Downloading provisional dataset...
...
Downloading labels file...
...
mvNCCompile v02.00, Copyright @ Movidius Ltd 2016
...
IMAGE_MEAN (108, 118, 128)
IMAGE_SCALE 0.017241379310344827
IMAGE_DIM (256, 256)
n_images = 2000
progess 100/2000 ...
progess 200/2000 ...
...
progess 1900/2000 ...
progess 2000/2000 ...

Now that you have a working framework, follow these instructions to create submissions.zip

1. Compile your custom trained network.
   • For Caffe based networks, rename your weights file as weights.caffemodel, deploy.prototxt as network.prototxt, and then call make compile.
   • For Tensorflow based networks, refer https://movidius.github.io/ncsdk/tf_compile_guidance.html.
2. Run inference on provisional dataset and generate inferences.csv
   • make infer
3. Compress all files required to make a valid submission
   • make zip

Troubleshooting

If the dataset download didn't complete, you can either delete the data directory and rerun make run, or manually download the files into this directory structure.

data/
data/provisional/
data/provisional/provisional_{00001..02000}.jpg
data/training/
data/training/training_{00001..80000}.jpg
data/training_ground_truth.csv
data/provisional.csv