f-CLSWGAN

Introduction

This work improves the performance of the model proposed in the paper "Feature Generating Networks for Zero-Shot Learning." CVPR (2018) by Yongqin Xian, Tobias Lorenz, Bernt Schiele, Zeynep Akata. To improve the performance of the generator and the discriminator I have used axial attention transformer. It is a simple but powerful technique to attend to multi-dimensional data efficiently.

Environment

• Python: 3.7,

• PyTorch: 1.2,

• scipy.

Dataset

The datasets can be downloaded from here. The datasets are 2048-d extracted feature maps from resnet-101.

Acknowledgement

The Axial Attention code is taken from this amazing repository.