Transfer learning with Xception architecture in Keras
This repository contains Jupyter notebooks with code described in this post: https://habrahabr.ru/post/347564/
To reproduce experiment:
- Download dataset from Kaggle
- Place it under data/train_raw
- Create directories
- test/cats
- test/dogs
- train/cats
- train/dogs
- validation/cats
- validation/dogs
- Run data-preparation.ipynb to split dataset
- Run transfer-learning.ipynb