Skip to content

rushhan/WGAN-with-Feedback-Cross-Attention-LayerNorm

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

4 Commits
 
 
 
 
 
 
 
 

Repository files navigation

WGAN-with-Feedback-Cross-Attention-LayerNorm

WGAN with feedback from discriminator& LayerNorm instead of BatchNorm

Modifications to the original WGAN implementation:

  1. Added feedback from the discriminator in the form of attention 2.Modified BN to LayerNorm as BatchNorm creates correlation between the examples. Note that the results from LayerNorm were not promising

Execute

python main.py --dataset lsun --dataroot /workspace/lsun --cuda --clamp --fdback --save_dir sample

Notes

Need the lsun data. Follow the instruction in original implementation link

Sources

WGAN Model based on original paper link.

Cross Attention is my own implementation

To-Do

Rerun and recheck the results