Tensorflow models with simple ways
Some models with easy understanding code,it will help you understand what the model does.These models can only work with not very good results.
Models list
1.Adversarial Networks (GAN)
- AAE Adversarial Autoencoders
- ACGAN Conditional Image Synthesis With Auxiliary Classifier GANs
- Auto encoder Recent Advances in Autoencoder-Based Representation Learning
- BGAN Boundary-Seeking Generative Adversarial Networks
- BiGAN Bidirectional Generative Adversarial Network
- CCGAN Semi-Supervised Learning with Context-Conditional Generative Adversarial Networks
- CGAN Conditional Generative Adversarial Nets
- CoGAN Coupled generative adversarial networks
- CycleGAN Unpaired Image-to-Image Translation using Cycle-Consistent Adversarial Networks
- DCGAN Deep Convolutional Generative Adversarial Network
- GAN Generative Adversarial Network with a MLP generator and discriminator
- VAE Auto-Encoding Variational Bayes
2.Special structure Convolutional network (Not recurring paper)
- DenseNet Densely Connected Convolutional Networks
- ResNet Deep Residual Learning for Image Recognition
- HighwayNet Highway Networks
- MobileNet_v1&v2 MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applications
MobileNetV2: Inverted Residuals and Linear Bottlenecks
3.Other Models
- None
Results of some models
The results are based on MNIST or Fashion-MNIST
AAE
ACGAN
Auto_encoder
Input image → Hidden layer → Output image
BGAN
BiGAN
CCGAN
Real image → Random cropping image → Repaired image
CGAN
CoGAN
A: MNIST B: Rotate 90 degrees MNIST
The model try to convert between A and B.
A → B
B → A
CycleGAN
A: MNIST B: Rotate 90 degrees MNIST
The model try to convert between A and B.
A → B
B → A