A classification model trained with
Supervised Contrastive Learning
(Prannay Khosla et al.).
The training procedure was done as seen in the example on
keras.io
by Khalid Salama.
The model was
trained on cifar10
, which includes ten classes: airplane, automobile, bird, cat, deer, dog, frog, horse, ship, truck.
The
test accuracy after 50 epochs
of the model with contrastive learning was
81.06%
(opposed to 79.88% without contrastive learning).