This repository is an implementation of Feature Generating Networks for Zero Shot Learning (https://arxiv.org/pdf/1712.00981.pdf) in Tensorflow.

1. Compatible with Python-3, Tensorflow 1.5
2. Download datasets for zero shot learning from http://datasets.d2.mpi-inf.mpg.de/xian/xlsa17.zip

  sh zsl_awa.sh

Make sure you provide the correct dataset path in the shell script. You can change the arguments passed in the shell script to generate results for any dataset.

I have also provided with the weights of the generator discriminator model on AWA2 at https://drive.google.com/file/d/1gSsokbQNrn0bpJtKQYRIqPK3-0tf0mRl/view?usp=sharing.