Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Preprint |
| Published: |
2024
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2410.06055 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866911304843788288 |
|---|---|
| author | Cao, Boyuan Ye, Jiaxin Wei, Yujie Shan, Hongming |
| author_facet | Cao, Boyuan Ye, Jiaxin Wei, Yujie Shan, Hongming |
| contents | While latent diffusion models (LDMs), such as Stable Diffusion, are designed for high-resolution (HR) image generation, they often struggle with significant structural distortions when generating images at resolutions higher than their training one. Instead of relying on extensive retraining, a more resource-efficient approach is to reprogram the pretrained model for HR image generation; however, existing methods often result in poor image quality and long inference time. We introduce RepLDM, a novel reprogramming framework for pretrained LDMs that enables high-quality, high-efficiency, high-resolution image generation; see Fig. 1. RepLDM consists of two stages: (i) an attention guidance stage, which generates a latent representation of a higher-quality training-resolution image using a novel training-free self-attention mechanism to enhance the structural consistency; and (ii) a progressive upsampling stage, which progressively performs upsampling in pixel space to mitigate the severe artifacts caused by latent space upsampling. The effective initialization from the first stage allows for denoising at higher resolutions with significantly fewer steps, improving the efficiency. Extensive experimental results demonstrate that RepLDM significantly outperforms state-of-the-art methods in both quality and efficiency for HR image generation, underscoring its advantages for real-world applications. Codes: https://github.com/kmittle/RepLDM. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2410_06055 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | RepLDM: Reprogramming Pretrained Latent Diffusion Models for High-Quality, High-Efficiency, High-Resolution Image Generation Cao, Boyuan Ye, Jiaxin Wei, Yujie Shan, Hongming Computer Vision and Pattern Recognition While latent diffusion models (LDMs), such as Stable Diffusion, are designed for high-resolution (HR) image generation, they often struggle with significant structural distortions when generating images at resolutions higher than their training one. Instead of relying on extensive retraining, a more resource-efficient approach is to reprogram the pretrained model for HR image generation; however, existing methods often result in poor image quality and long inference time. We introduce RepLDM, a novel reprogramming framework for pretrained LDMs that enables high-quality, high-efficiency, high-resolution image generation; see Fig. 1. RepLDM consists of two stages: (i) an attention guidance stage, which generates a latent representation of a higher-quality training-resolution image using a novel training-free self-attention mechanism to enhance the structural consistency; and (ii) a progressive upsampling stage, which progressively performs upsampling in pixel space to mitigate the severe artifacts caused by latent space upsampling. The effective initialization from the first stage allows for denoising at higher resolutions with significantly fewer steps, improving the efficiency. Extensive experimental results demonstrate that RepLDM significantly outperforms state-of-the-art methods in both quality and efficiency for HR image generation, underscoring its advantages for real-world applications. Codes: https://github.com/kmittle/RepLDM. |
| title | RepLDM: Reprogramming Pretrained Latent Diffusion Models for High-Quality, High-Efficiency, High-Resolution Image Generation |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2410.06055 |