Return a vehicle type probability based on an image with about 93% accuracy.
See https://www.kaggle.com/code/dima806/vehicle-10-types-detection-vit for more details.
Classification report:
precision recall f1-score support
SUV 0.8780 0.9000 0.8889 40
bus 1.0000 1.0000 1.0000 40
family sedan 0.8571 0.9000 0.8780 40
fire engine 0.8444 0.9500 0.8941 40
heavy truck 0.9459 0.8750 0.9091 40
jeep 0.9512 0.9750 0.9630 40
minibus 0.9500 0.9500 0.9500 40
racing car 1.0000 0.9500 0.9744 40
taxi 0.9750 0.9750 0.9750 40
truck 0.9722 0.8750 0.9211 40
accuracy 0.9350 400
macro avg 0.9374 0.9350 0.9354 400
weighted avg 0.9374 0.9350 0.9354 400