Deep learning-based classification of galaxy morphologies using images from the Galaxy10 DECals dataset. This repository compares state-of-the-art models, including ConvNeXt, EfficientNet, and Swin Transformer, to automate and improve the accuracy of galaxy type identification for astrophysical research.
computer-vision deep-learning astronomy python3 pytorch transformer image-classification galaxy-morphology efficientnet astronn swin-transformer convnext
-
Updated
Sep 26, 2025 - Jupyter Notebook