Saved in:
| Main Authors: | , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2023
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2312.17742 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866909057333329920 |
|---|---|
| author | Tian, Yonglong Fan, Lijie Chen, Kaifeng Katabi, Dina Krishnan, Dilip Isola, Phillip |
| author_facet | Tian, Yonglong Fan, Lijie Chen, Kaifeng Katabi, Dina Krishnan, Dilip Isola, Phillip |
| contents | We introduce SynCLR, a novel approach for learning visual representations exclusively from synthetic images and synthetic captions, without any real data. We synthesize a large dataset of image captions using LLMs, then use an off-the-shelf text-to-image model to generate multiple images corresponding to each synthetic caption. We perform visual representation learning on these synthetic images via contrastive learning, treating images sharing the same caption as positive pairs. The resulting representations transfer well to many downstream tasks, competing favorably with other general-purpose visual representation learners such as CLIP and DINO v2 in image classification tasks. Furthermore, in dense prediction tasks such as semantic segmentation, SynCLR outperforms previous self-supervised methods by a significant margin, e.g., improving over MAE and iBOT by 6.2 and 4.3 mIoU on ADE20k for ViT-B/16. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2312_17742 |
| institution | arXiv |
| publishDate | 2023 |
| record_format | arxiv |
| spellingShingle | Learning Vision from Models Rivals Learning Vision from Data Tian, Yonglong Fan, Lijie Chen, Kaifeng Katabi, Dina Krishnan, Dilip Isola, Phillip Computer Vision and Pattern Recognition We introduce SynCLR, a novel approach for learning visual representations exclusively from synthetic images and synthetic captions, without any real data. We synthesize a large dataset of image captions using LLMs, then use an off-the-shelf text-to-image model to generate multiple images corresponding to each synthetic caption. We perform visual representation learning on these synthetic images via contrastive learning, treating images sharing the same caption as positive pairs. The resulting representations transfer well to many downstream tasks, competing favorably with other general-purpose visual representation learners such as CLIP and DINO v2 in image classification tasks. Furthermore, in dense prediction tasks such as semantic segmentation, SynCLR outperforms previous self-supervised methods by a significant margin, e.g., improving over MAE and iBOT by 6.2 and 4.3 mIoU on ADE20k for ViT-B/16. |
| title | Learning Vision from Models Rivals Learning Vision from Data |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2312.17742 |