Research map / Architectures and efficiency / Attention and transformers
Lite Vision Transformer With Enhanced Self-Attention
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Use LVT instead of PVTv2-B0 as the SegFormer encoder | SegFormer · ADE20K | mIoU: 37.4 -> 39.3 | Helped |
| Use LVT instead of PVTv2-B0 as the Panoptic FPN backbone | Panoptic FPN · COCO val | PQ: 41.3 -> 42.8 | Helped |
| Use LVT instead of PVTv2-B0 as the Panoptic FPN backbone | Panoptic FPN · COCO test-dev | PQ: 41.2 -> 43.0 | Helped |
| Add CSA to VOLO-D0 (ablation) | VOLO-D0 · ImageNet | Top-1 accuracy (%): 74.6 -> 75.2 | Helped |
| Add CSA to VOLO-D0 (ablation) | VOLO-D0 · ADE20K | mIoU: 39.3 -> 40.0 | Helped |
| Add RASA to VOLO-D0 + CSA (ablation) | VOLO-D0 + CSA · ImageNet | Top-1 accuracy (%): 75.2 -> 75.6 | Helped |
| Add RASA to VOLO-D0 + CSA (ablation) | VOLO-D0 + CSA · ADE20K | mIoU: 40.0 -> 41.0 | Helped |
| Increase RASA recursion from 1 to 2 (ablation) | LVT · ImageNet | Top-1 accuracy (%): 73.9 -> 74.8 | Helped |
| Increase RASA recursion from 2 to 3 (ablation) | LVT · ImageNet | Top-1 accuracy (%): 74.8 -> 74.6 | Hurt |
| Increase RASA recursion from 2 to 4 (ablation) | LVT · ImageNet | Top-1 accuracy (%): 74.8 -> 74.9 | Helped |
Closest papers by meaning
- You Only Need Less Attention at Each Stage in Vision Transformers (CVPR 2024 · similarity 0.89)
- ConViT: Improving Vision Transformers with Soft Convolutional Inductive Biases (ICML 2021 · similarity 0.89)
- Adder Attention for Vision Transformer (NeurIPS 2021 · similarity 0.88)
- Long-Short Transformer: Efficient Transformers for Language and Vision (NeurIPS 2021 · similarity 0.88)
- LSNet: See Large, Focus Small (CVPR 2025 · similarity 0.87)
- Skip-Attention: Improving Vision Transformers by Paying Less Attention (ICLR 2024 · similarity 0.86)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.