Image Classification using CNN
Using Convolutional Neural Network (CNN) based deep learning method implemented two models for performing classification on CIFAR-10 dataset.
CIFAR-10 dataset consists of 60000 images of size 32x32x3.
I have implemented two models one without data augmentation and other with data augmentation. Model trained without data augmentation reached its limit in just 20 epochs from there it is overfitting the data. But for the model trained using data augmentation learns for more than 120 epochs and also shows good accuracy and loss compared to previous one.
| Model | val_Loss | val_Accuracy |
|---|---|---|
| Without Data Augmentation | 0.976 | 66.52 % |
| With Data Augmentation | 0.7310 | 75.26 % |
For using the code Download it and upload in google colab then you can run it on GPU with tensorflow framework.