Switti: Designing Scale-Wise Transformers for Text-to-Image Synthesis
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2024
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866910885049532416 |
|---|---|
| author | Voronov, Anton Kuznedelev, Denis Khoroshikh, Mikhail Khrulkov, Valentin Baranchuk, Dmitry |
| author_facet | Voronov, Anton Kuznedelev, Denis Khoroshikh, Mikhail Khrulkov, Valentin Baranchuk, Dmitry |
| contents | This work presents Switti, a scale-wise transformer for text-to-image generation. We start by adapting an existing next-scale prediction autoregressive (AR) architecture to T2I generation, investigating and mitigating training stability issues in the process. Next, we argue that scale-wise transformers do not require causality and propose a non-causal counterpart facilitating ~21% faster sampling and lower memory usage while also achieving slightly better generation quality. Furthermore, we reveal that classifier-free guidance at high-resolution scales is often unnecessary and can even degrade performance. By disabling guidance at these scales, we achieve an additional sampling acceleration of ~32% and improve the generation of fine-grained details. Extensive human preference studies and automated evaluations show that Switti outperforms existing T2I AR models and competes with state-of-the-art T2I diffusion models while being up to 7x faster. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2412_01819 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Switti: Designing Scale-Wise Transformers for Text-to-Image Synthesis Voronov, Anton Kuznedelev, Denis Khoroshikh, Mikhail Khrulkov, Valentin Baranchuk, Dmitry Computer Vision and Pattern Recognition This work presents Switti, a scale-wise transformer for text-to-image generation. We start by adapting an existing next-scale prediction autoregressive (AR) architecture to T2I generation, investigating and mitigating training stability issues in the process. Next, we argue that scale-wise transformers do not require causality and propose a non-causal counterpart facilitating ~21% faster sampling and lower memory usage while also achieving slightly better generation quality. Furthermore, we reveal that classifier-free guidance at high-resolution scales is often unnecessary and can even degrade performance. By disabling guidance at these scales, we achieve an additional sampling acceleration of ~32% and improve the generation of fine-grained details. Extensive human preference studies and automated evaluations show that Switti outperforms existing T2I AR models and competes with state-of-the-art T2I diffusion models while being up to 7x faster. |
| title | Switti: Designing Scale-Wise Transformers for Text-to-Image Synthesis |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2412.01819 |