Kaggle Amazon deforestation challenge
This repository contains the source code of the Kaggle Amazon deforestation challenge
How to use and view the jupyter notebook
The jupyter notebook is stripped to a .py file for version control in the notebook/ folder. To recreate the original .ipynb file and use it with jupyter execute tools/extract_py_notebook.sh (after cloning the repo for example).
When you'll work on the notebook, git will ignore the changes you made to it as specified in the .gitignore file. You need to strip it to a .py file first before pushing it to the repo.
To do so and let jupyter automatically create the .py file each time you save your notebook you need to add a script to its configuration simply by executing the following command:
tools/add_jupyter_vc.sh
This command should be only executed once on a system. If you ever run into issues with this new configuration you can restore your original jupyter configuration with:
tools/restore_jupyter_config.sh
/!\ Be careful to not run tools/add_jupyter_vc.sh twice in a row as it will erase your original configuration file forever.
Dependencies
- Kaggle Data downloader
- Tensorflow 1.1
- Scipy
- Seaborn