Classify programming language

It uses ULMFIT model https://arxiv.org/abs/1801.06146 Model with default parameters achieves ~96% accuracy on validation set.

It requires:

• python 3.7
• gpu for training (recommended)

Basing training classifier:

python train.py --input data.csv --model-path ulmfit.pkl 

Running service that predicts programming language:

python app.py application.conf

Example of curl:

curl -X POST -H "Content-Type:application/json" 'http://0.0.0.0:8010/predict' 
--data '{"text": "#include <iostream>"}'

Running service from docker:

docker build -t detectorService -f DockerfileService .

docker run --rm -it detectorService python app.py application.conf

Further improvements:

• Compare with baseline models such as CNN, N-grams features
• Try different tokenizations instead of just on letters
• Try different hyperparameters, bigger model.