Returns car brand with about 69% accuracy given an image.
See https://www.kaggle.com/code/dima806/car-brands-image-detection-vit for details.
Classification report:
precision recall f1-score support
Acura 0.3799 0.5658 0.4546 2066
Alfa Romeo 0.7487 0.9424 0.8344 2067
Aston Martin 0.9377 0.8162 0.8727 2067
Audi 0.3810 0.6623 0.4837 2067
BMW 0.4379 0.1824 0.2575 2067
Bentley 0.7206 0.8360 0.7740 2067
Bugatti 0.9862 1.0000 0.9930 2067
Buick 0.5081 0.4981 0.5031 2066
Cadillac 0.7252 0.4315 0.5411 2067
Chevrolet 0.3715 0.1553 0.2190 2067
Chrysler 0.6298 0.7551 0.6868 2066
Citroen 0.9597 0.9903 0.9748 2067
Daewoo 0.9745 1.0000 0.9871 2067
Dodge 0.5020 0.6618 0.5710 2067
Ferrari 0.9238 0.9908 0.9561 2067
Fiat 0.8116 0.8670 0.8384 2067
Ford 0.4484 0.0798 0.1355 2067
GMC 0.5630 0.7842 0.6555 2067
Genesis 0.6549 0.8916 0.7552 2067
Honda 0.3684 0.3880 0.3779 2067
Hudson 0.9584 0.8132 0.8798 2066
Hyundai 0.3593 0.3527 0.3560 2067
Infiniti 0.4569 0.6546 0.5382 2067
Jaguar 0.4496 0.2975 0.3581 2067
Jeep 0.8256 0.8563 0.8407 2067
Kia 0.3308 0.1035 0.1577 2067
Lamborghini 0.9252 0.9811 0.9523 2067
Land Rover 0.5205 0.8365 0.6417 2067
Lexus 0.4655 0.2221 0.3007 2067
Lincoln 0.5455 0.5244 0.5348 2067
MG 0.7773 0.9879 0.8700 2067
Maserati 0.7179 0.8162 0.7639 2067
Mazda 0.4517 0.4664 0.4589 2067
McLaren 0.9782 1.0000 0.9890 2066
Mercedes-Benz 0.3383 0.0329 0.0600 2067
Mini 0.8048 0.9337 0.8645 2067
Mitsubishi 0.4671 0.7928 0.5878 2066
Nissan 0.5305 0.0672 0.1194 2067
Oldsmobile 0.8832 0.9918 0.9344 2067
Peugeot 0.9070 1.0000 0.9512 2067
Pontiac 0.9641 0.9884 0.9761 2067
Porsche 0.5380 0.6376 0.5836 2067
Ram 0.8475 0.9652 0.9025 2067
Ram Trucks 0.9626 0.9831 0.9727 2067
Renault 0.9686 1.0000 0.9840 2066
Rolls-Royce 0.8737 0.9671 0.9180 2067
Saab 0.9311 1.0000 0.9643 2067
Smart 0.9247 0.9627 0.9433 2066
Studebaker 0.9645 1.0000 0.9819 2067
Subaru 0.4404 0.3112 0.3647 2066
Suzuki 0.9425 1.0000 0.9704 2067
Tesla 0.7482 0.9390 0.8328 2066
Toyota 0.2884 0.0755 0.1196 2067
Volkswagen 0.4282 0.4964 0.4598 2067
Volvo 0.4807 0.5300 0.5041 2066
accuracy 0.6925 113674
macro avg 0.6733 0.6925 0.6638 113674
weighted avg 0.6733 0.6925 0.6638 113674