Our code was tested on Ubuntu 18.04 with CUDA 10.2, quad core system with NVIDIA GeForce RTX 2070 8 GB RAM GPU, and 48 GB system RAM. It needs at least 20 GB of free disk space.
Run all code within the same directory containing BlackBox-production.yml, DATA_TRAINING.RData and TRUE_DATA_RANKING.RData.
Install conda: https://docs.conda.io/projects/conda/en/latest/user-guide/install/#regular-installation
Create conda environment with Python and R packages.
$ conda env create -f BlackBox-production.yml
$ conda activate BlackBox-productionConvert data and generate random training masks.
$ python convert_data.py
$ python generate_extra_masks.py # takes up to 12 hours to complete
$ python shift_bitmaps.py
$ python downsample_bitmaps.pyTrain models and evaluate their ensemble prediction (approximately 7 days to complete):
$ bash train_models.sh # BASH script training an ensemble of 155 models, each with specific hyperparameters
$ python make_models_database.py -i Model
$ Rscript preprocesing_TRUE_DATA_RANKING.R # creates file true.observations.rda with data used in score calculation
$ python make_prediction.py -i Model -o Model_score --number_of_predictions=20 --prediction_suffix=SampleA
$ python make_prediction.py -i Model -o Model_ensemble_score --only_score=20:SampleA --global_ensembleFinal prediction is saved in ./predictions/. Calculated score is saved in the last line of ./Model_ensemble_score_20_SampleA.csv.
In our code we use the following Python libraries which rely on different random number generators: random from stdlib, numpy, pytorch, and fastai. Since at least pytorch does not guarantee reproducibility between different platforms, CPU/GPU runs, and library versions even with a fixed random seed, we have opted not to seed any of the random number generators in our solution. This means a certain amount of variation in final result is expected between runs.