Kaggle cats vs dogs image recognition machine learning competition (using deep learning).
Cats vs Dogs - attempt I.pynb achieved a score of 0.09907 on the kaggle competition which would have placed it 426/1315 i.e. in the top third.
Note this code assumes the competition data is stored in a folder called data/. The data can be downloaded using the kaggle_cli as follows:
pip install kaggle-cli
kg config -g -u `username` -p `password` -c dogs-vs-cats-redux-kernels-edition
kg download
It's worth noting that if you have special characters such as $ in your password you should enclose your pass word in single quotes.
Then put the three downloaded files in a folder called data/ and unzip the data
unzip train.zip
unzip test.zip
Move the test data into a subfolder called data/test/test, i.e. create one extra subfolder inside data/test.
The code will then structure the data files appropriately from there.