zindi-spot-the-mask-challenge-standalone-notebook

Standalone code to show how to participated in this zindi challenge using TF.Keras on Google Colab. Read more here.

Path update

• You have to update PATH ( path place where the challenge repository will be created ) in notebook 1-BUILD_REPOSITORY_STRUCTURE.ipynb.

• You have to update REPO_PATH ( path to go to the created repository of mask challenge repository ) in notebooks 2-dataset_structuring.ipynb and 4-keras_starter-spot_the_mask.ipynb.

Instructions to RUN

1- Instruction about folder structure

• Run the notebook 1-BUILD_REPOSITORY_STRUCTURE.ipynb to create the challenge repository and its sub-folders.

• Move all notebooks in the created challenge repository to its sub-folder NOTEBOOKS.

2- Instruction about the dataset

• Download challenge dataset on ZINDI web site and put each file in corresponding sub-folder of the DATASET folder in regard of file type.

3- Instruction about dataset structuring

• Run the notebook 2-dataset_structuring.ipynb to unzip and process the dataset.

4- Instruction about starter notebook

• Run the notebook 4-keras_starter-spot_the_mask.ipynb to create your first submission file ( which will be saved in the challenge repository to its sub-folder SUBMISSIONS ) and make your furst submit on ZINDI. After that, play around model architecture, augmentation parameters of ImageDataGenerator and any other tips you want to improve in ranking on the leaderboard.