Image classification using convolutional neural network with CIFAR-10 dataset https://www.cs.toronto.edu/~kriz/cifar.html
We experimented with different approaches in order to improve test accuracy. Please note the results for the same test might vary.
-
ImageClassificationCNN1 produced test accuracy at 0.7049999833106995 and test loss: of 0.8769986033439636
-
ImageClassificationCNN2 produced test accuracy at 0.7807000279426575 and test loss: of 0.6588695645332336
-
ImageClassificationCNN3 produced test accuracy at 0.8248999714851379 and test loss of 0.5760216116905212
-
ImageClassificationCNN4 produced test accuracy at 0.8363000154495239 and test loss of 0.8086720108985901
-
ImageClassificationCNN5 is the same as ImageClassificationCNN4 but has added visualization 🐶 for sample prediction