An exploratory definition of convolutional network for automatic classification of galaxies, based on visual images. Relevant files in the repository:
galaxy-convnet-v2.ipynb: Python notebook for training the model;galaxy-convnet-v2.h5: model weights stored in HDF5 format;galaxy-convnet-dataset.ipynb: Python notebook for creating the dataset;data.tgz: possible dataset for training and validation, created with the above notebook.
More information about this project available in this blog post.
Thanks to Mariana Júlio for her interest in this project that sourced updates and improvements.