Research map / Architectures and efficiency / Attention and transformers
Rethinking Spatial Dimensions of Vision Transformers
Byeongho Heo, Sangdoo Yun, Dongyoon Han, Sanghyuk Chun, Junsuk Choe, Seong Joon Oh
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (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 |
|---|---|---|---|
| add ResNet-style spatial reduction to ResNet50 (ablation) | ResNet50 · ImageNet-1k | training loss / validation accuracy | Improved |
| add pooling layers (PiT) to ViT, vanilla training | ViT-Ti · ImageNet-1k val | top-1 accuracy: 68.7% -> 71.3% | Improved |
| add pooling layers (PiT) to ViT, vanilla training | ViT-S · ImageNet-1k val | top-1 accuracy: 68.7% -> 73.3% | Improved |
| add pooling layers (PiT) to ViT, vanilla training | ViT-B · ImageNet-1k val | top-1 accuracy: 69.3% -> 76.1% | Improved |
| add pooling layers (PiT) to ViT, CutMix training | ViT-B · ImageNet-1k val | top-1 accuracy: 75.3% -> 79.9% | Improved |
| add pooling layers (PiT) to ViT, DeiT training | ViT-Ti · ImageNet-1k val | top-1 accuracy: 72.2% -> 73.0% | Improved |
| add pooling layers (PiT) to ViT, DeiT training | ViT-B · ImageNet-1k val | top-1 accuracy: 81.8% -> 82.0% | Improved |
| add pooling layers (PiT) to ViT with distillation | ViT-Ti · ImageNet-1k val | top-1 accuracy: 74.5% -> 74.6% | No clear effect |
| add pooling layers (PiT) to ViT, 1000 epoch training | ViT-Ti · ImageNet-1k val | top-1 accuracy: 76.6% -> 76.4% | Hurt |
| add pooling layers (PiT) to ViT, 384x384 fine-tune | ViT-B · ImageNet-1k val | top-1 accuracy: 83.1% -> 83.0% | No clear effect |
| use PiT-S backbone instead of ViT-S | Deformable DETR · COCO val2017 | AP: 36.9 -> 39.4 | Improved |
| use PiT-S backbone instead of ViT-S | ImageNet-A · ImageNet-A | accuracy: 19.1 -> 21.7 | Improved |
Closest papers by meaning
- Do Vision Transformers See Like Convolutional Neural Networks? (NeurIPS 2021 · similarity 0.91)
- A Survey on Efficient Vision Transformers: Algorithms, Techniques, and Performance Benchmarking (TPAMI 2024 · similarity 0.89)
- Vision Transformers provably learn spatial structure (NeurIPS 2022 · similarity 0.89)
- CvT: Introducing Convolutions to Vision Transformers (ICCV 2021 · similarity 0.89)
- Scalable Vision Transformers With Hierarchical Pooling (ICCV 2021 · similarity 0.89)
- Your ViT is Secretly an Image Segmentation Model (CVPR 2025 · similarity 0.88)
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.
