This repository contains code and data for classifying spectra from the Digitized First Byurakan Survey (DFBS).
This folder contains code and data for training and testing convolutional neural network models for classifying spectra into four classes: UV-excess galaxies, hot subdwarfs, carbon stars, and other objects. The code is the implementation of the paper: https://doi.org/10.1016/j.ascom.2020.100442.
This folder contains code and data for training and testing convolutional neural network models for the sub-object classification of the abovementioned groups.
This folder contains code and data for the cloud-based service for classifying astronomical images in a Google Colab environment. The service allows users to upload their spectra or use sample spectra from DFBS and get the classification results. The folder has three main notebooks: train_colab.ipynb, test_colab.ipynb, and infer_colab.ipynb.