suchaoxiao / keras-cifar10

practice on CIFAR10 with Keras

Keras for CIFAR10

This project demonstrates some personal examples with Keras on CIFAR10.

---

Introduction

---

The CIFAR10 dataset is 32x32 size, 50000 train images and 10000 test images. The dataset is divided into 40000 train images, 10000 validation images, and 10000 images.

Result

All result is tested on 10000 test images.You can view and run in the jupyter environment.

Model	Notebook	Accuracy
SVM	svm	33.36%
Softmax	softmax	33.11%
simple_cnn	simple_cnn	66.75%
vgg	vgg	92.32%
inceptionV1	GoogLeNet	93.08%
ResNet18	resnet18	93.47%
small-ResNet20	small_resnet20	91.25%
small-ResNet32	small_resnet32	92.34%
small-ResNet56	small_resnet56	92.37%