Efficient Image Generation with Variadic Attention Heads
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2022
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866909660330590208 |
|---|---|
| author | Walton, Steven Hassani, Ali Xu, Xingqian Wang, Zhangyang Shi, Humphrey |
| author_facet | Walton, Steven Hassani, Ali Xu, Xingqian Wang, Zhangyang Shi, Humphrey |
| contents | While the integration of transformers in vision models have yielded significant improvements on vision tasks they still require significant amounts of computation for both training and inference. Restricted attention mechanisms significantly reduce these computational burdens but come at the cost of losing either global or local coherence. We propose a simple, yet powerful method to reduce these trade-offs: allow the attention heads of a single transformer to attend to multiple receptive fields.
We demonstrate our method utilizing Neighborhood Attention (NA) and integrate it into a StyleGAN based architecture for image generation. With this work, dubbed StyleNAT, we are able to achieve a FID of 2.05 on FFHQ, a 6% improvement over StyleGAN-XL, while utilizing 28% fewer parameters and with 4$\times$ the throughput capacity. StyleNAT achieves the Pareto Frontier on FFHQ-256 and demonstrates powerful and efficient image generation on other datasets. Our code and model checkpoints are publicly available at: https://github.com/SHI-Labs/StyleNAT |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2211_05770 |
| institution | arXiv |
| publishDate | 2022 |
| record_format | arxiv |
| spellingShingle | Efficient Image Generation with Variadic Attention Heads Walton, Steven Hassani, Ali Xu, Xingqian Wang, Zhangyang Shi, Humphrey Computer Vision and Pattern Recognition Artificial Intelligence Machine Learning While the integration of transformers in vision models have yielded significant improvements on vision tasks they still require significant amounts of computation for both training and inference. Restricted attention mechanisms significantly reduce these computational burdens but come at the cost of losing either global or local coherence. We propose a simple, yet powerful method to reduce these trade-offs: allow the attention heads of a single transformer to attend to multiple receptive fields. We demonstrate our method utilizing Neighborhood Attention (NA) and integrate it into a StyleGAN based architecture for image generation. With this work, dubbed StyleNAT, we are able to achieve a FID of 2.05 on FFHQ, a 6% improvement over StyleGAN-XL, while utilizing 28% fewer parameters and with 4$\times$ the throughput capacity. StyleNAT achieves the Pareto Frontier on FFHQ-256 and demonstrates powerful and efficient image generation on other datasets. Our code and model checkpoints are publicly available at: https://github.com/SHI-Labs/StyleNAT |
| title | Efficient Image Generation with Variadic Attention Heads |
| topic | Computer Vision and Pattern Recognition Artificial Intelligence Machine Learning |
| url | https://arxiv.org/abs/2211.05770 |