Your Local GAN: Designing Two Dimensional Local Attention Mechanisms for Generative Models
Preprint
Abstract
We introduce a new local sparse attention layer that preserves two-dimensional geometry and locality. We show that by just replacing the dense attention layer of SAGAN with our construction, we obtain very significant FID, Inception score and pure visual improvements. FID score is improved from $18.65$ to $15.94$ on ImageNet, keeping all other parameters the same. The sparse attention patterns that we propose for our new layer are designed using a novel information theoretic criterion that uses information flow graphs. We also present a novel way to invert Generative Adversarial Networks with attention. Our method extracts from the attention layer of the discriminator a saliency map, which we use to construct a new loss function for the inversion. This allows us to visualize the newly introduced attention heads and show that they indeed capture interesting aspects of two-dimensional geometry of real images.
Keywords
All Related Versions
This publication has 3 references indexed in Scilit:
- A Style-Based Generator Architecture for Generative Adversarial NetworksIEEE Transactions on Pattern Analysis and Machine Intelligence, 2020
- Global Guarantees for Enforcing Deep Generative Priors by Empirical RiskIEEE Transactions on Information Theory, 2019
- Adaptive Attention Span in Transformers2019