99% accuracy on testing set

Used Pretrained VGG-16 Model with weights of imagenet and pooling layer as avg