Authors: Philippe Lelièvre, Julien Horwood, Ishaan Kumar, Vicki Anand
├── dummy_data <- Fake data for testing purpose
├── mappings <- Baseline model mapping (used for the baseline submission)
├── models <- Models used for the experiments
├── notebooks <- Notebooks used mainly for data visualization and prototyping
├── scripts <- All training and experiments scripts
├── test_results <- Examples of test prediction results
├── trained_model <- Trained models folder
├── utils <- Utilities used accross the application
├── xgboost <- [IGNORE THIS FOLDER FOR THE CODE EVALUATION] Source code of the XGBoost project (https://github.com/dmlc/xgboost) to work around the problem that we cannot install a library on the environment. To be removed in the future if we can install the library.
├── ift6758 <- Evaluation script
└── README.md <- The top-level README for developers using this project.
- python scripts/train.py --input_path ../../new_data/Train/ --model final --output_results_path ../trained_models
- python scripts/train.py --input_path ../../new_data/Train/ --model final --eval_model True
- python scripts/train.py --input_path ../../new_data/Train/ --model final --k_fold True
- Limit all lines to a maximum of 119 characters. Only exception is when it impact negatively the readibility of a sentence.
Like mentioned in the project structure section, we included the source code of the XGBoost project (https://github.com/dmlc/xgboost) to work around the problem that we cannot install any library on the main python environment.
For the evaluation procedure, ignore the code inside the folder named xgboost
This folder should be removed in the future if we can install the library on the main python environment.