VLASS is expected to catalog approximately 10 million radio sources, generating corresponding quicklook images during its lifetime operation. An efficient and automated way to classify these images needs to be developed. Training an Image Classifier using PyTorch to distinguish between common radio sources classifications e.g. Fanaroff-Riley or a simplified classification scheme.
A Neural Network model with PyTorch was developed, and began initial model training. The model is able to classify the quicklook images into 3 categories: single, multiple, and diffuse sources. A classification tool like this is relevant to the astronomy community in general, to be able to access specific classifications regarding radio sources.
Keywords: Radio Astronomy, Machine Learning, Image Classification, Radio Sources, VLASS, PyTorch.