kg-dog-breed

Code for kaggle's Dog Breed Identification problem.

Setup

1. Clone the repository and navigate to its root.

git clone https://github.com/hjwk/kg-dog-breed
cd kg-dog-project

2. Create a virtual environment from the requirements files (I used the same environment as for Udacity's dog project.

conda env create -f requirements/dog-windows.yml
activate kg-dog-breed

3. Download the training and testing data and put them in data/train and data/test.

4. Execute the dataCleaning.py script in order to re-organize the training data into a training and a validation datasets. After running this script you should have a new folder named data_gen and withing it a test and train folder in which the photos are organized into folders named after their classes.

Generating the bottleneck features and training the CNN

You can now use kg-dog-breed.py to generate the bottleneck features and train your model. If you do not want to rebuild the bottleneck features each tme you run the script simply comment the appropriate lines.