Saved in:
| Main Authors: | , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2024
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2412.06295 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866917862845710336 |
|---|---|
| author | Liu, Yunpeng Liu, Boxiao Zhang, Yi Hou, Xingzhong Song, Guanglu Liu, Yu You, Haihang |
| author_facet | Liu, Yunpeng Liu, Boxiao Zhang, Yi Hou, Xingzhong Song, Guanglu Liu, Yu You, Haihang |
| contents | Significant advances have been made in the sampling efficiency of diffusion models and flow matching models, driven by Consistency Distillation (CD), which trains a student model to mimic the output of a teacher model at a later timestep. However, we found that the learning complexity of the student model varies significantly across different timesteps, leading to suboptimal performance in CD.To address this issue, we propose the Curriculum Consistency Model (CCM), which stabilizes and balances the learning complexity across timesteps. Specifically, we regard the distillation process at each timestep as a curriculum and introduce a metric based on Peak Signal-to-Noise Ratio (PSNR) to quantify the learning complexity of this curriculum, then ensure that the curriculum maintains consistent learning complexity across different timesteps by having the teacher model iterate more steps when the noise intensity is low. Our method achieves competitive single-step sampling Fréchet Inception Distance (FID) scores of 1.64 on CIFAR-10 and 2.18 on ImageNet 64x64.Moreover, we have extended our method to large-scale text-to-image models and confirmed that it generalizes well to both diffusion models (Stable Diffusion XL) and flow matching models (Stable Diffusion 3). The generated samples demonstrate improved image-text alignment and semantic structure, since CCM enlarges the distillation step at large timesteps and reduces the accumulated error. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2412_06295 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | See Further When Clear: Curriculum Consistency Model Liu, Yunpeng Liu, Boxiao Zhang, Yi Hou, Xingzhong Song, Guanglu Liu, Yu You, Haihang Computer Vision and Pattern Recognition Significant advances have been made in the sampling efficiency of diffusion models and flow matching models, driven by Consistency Distillation (CD), which trains a student model to mimic the output of a teacher model at a later timestep. However, we found that the learning complexity of the student model varies significantly across different timesteps, leading to suboptimal performance in CD.To address this issue, we propose the Curriculum Consistency Model (CCM), which stabilizes and balances the learning complexity across timesteps. Specifically, we regard the distillation process at each timestep as a curriculum and introduce a metric based on Peak Signal-to-Noise Ratio (PSNR) to quantify the learning complexity of this curriculum, then ensure that the curriculum maintains consistent learning complexity across different timesteps by having the teacher model iterate more steps when the noise intensity is low. Our method achieves competitive single-step sampling Fréchet Inception Distance (FID) scores of 1.64 on CIFAR-10 and 2.18 on ImageNet 64x64.Moreover, we have extended our method to large-scale text-to-image models and confirmed that it generalizes well to both diffusion models (Stable Diffusion XL) and flow matching models (Stable Diffusion 3). The generated samples demonstrate improved image-text alignment and semantic structure, since CCM enlarges the distillation step at large timesteps and reduces the accumulated error. |
| title | See Further When Clear: Curriculum Consistency Model |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2412.06295 |