home > examples > image_classification

Image Classification Examples

This directory contains various classification examples on image datasets:

• MNIST:

  • mnist_dnn.py - simple MNIST classification example. Note: The purpose of the example on MNIST is to demonstrate the use of a deep neural network for classification. As such, the network does not achieve State of the Art (SOTA) classification accurary. A Convolutional Neural Network (CNN) should be used for that purpose.

  • mnist_cnn.py - a CNN-based MNIST classification example.

  • mnist_dp.py - MNIST example with differential privacy.

• CIFAR10

  • cifar10_simple.py - very simple CIFAR10 classification example which demonstrated how to write basic training loop with data augmentation

  • cifar10_advanced.py - more advanced CIFAR10 example which allows user to configure neural network architecture and other hyperparameters. It also supports training on multiple GPUs using objax.Parallel.

• Imagenet

  • imagenet_pretrained_vgg.py - example which shows how to load pre-trained weights for a VGG model and use it to classify input images. For more details see documentation.

  • imagenet_resnet50_train.py - example which shows how to train Resnet50 model on Imagenet. For more details see example documentation.

• Horses or Humans

  • horses_or_humans_logistic.py - simple example using logistic regression.