cifar10 classification based on alexnet and vgg16 using TensorFlow
- model.py Alexnet model, change the size and number of kernels and add dropout and batch normalization.
- train.py train Alexnet, and use Tensorboard to help visualize.
- vgg16.py vgg model, add another fc layer with 10 output and add batch normalization. Reference: Davi Frossard, 2016, VGG16 implementation in TensorFlow
- trainVGG.py train vgg model, nearly the same as train.py
- cifar10_input.py load data, including data augmentation.
Best result:85%