Using classical and neural image embeddings and finetuned end-to-end networks to achieve top-tier performance on a vehicle type classification task. Containerized and deployed model as a web app
Good to have this up and running as soon as possible. just feed whatever feature vector we finish first to multinomial logistic regression (potentially with l1 penalty or other regularization).
Good to have this up and running as soon as possible. just feed whatever feature vector we finish first to multinomial logistic regression (potentially with l1 penalty or other regularization).