yuhsianglin / SAAE

Semantic Adversarial Autoencoder

Semantic Adversarial Autoencoder

Python code (using TensorFlow) of Semantic Adversarial Autoencoder (SAAE), including using explicit matching and implicit matching. Also include the implementation of Semantic Autoencoder (SAE, arXiv:1704.08345), and its variation with GAN.

Getting Started

In the code directory, run_*.py are the run script for training and evaluating the four models. To run one of the model, after setting up the environment, type, for example,

python run_aaeexp_AwA_sZSL.py

Built With

• Python
• TensorFlow

Author

• Yu-Hsiang Lin - LinkedIn
• Kangyan Zhou
• Shihui Li

Documentation

• Shihui Li, Yu-Hsiang Lin, Kangyan Zhou, Semantic Adversarial Autoencoder for Zero-Shot Learning.