Research map / Architectures and efficiency / Attention and transformers
Patch Slimming for Efficient Vision Transformers
Yehui Tang, Kai Han, Yunhe Wang, Chang Xu, Jianyuan Guo, Chao Xu, Dacheng Tao
Check 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 |
|---|---|---|---|
| Apply static top-down patch slimming | DeiT-Ti · ImageNet | Top-1 accuracy: 72.2 -> 72.0 | Hurt |
| Apply dynamic top-down patch slimming | DeiT-Ti · ImageNet | FLOPs reduction: 53.8 | Improved |
| Apply static top-down patch slimming | DeiT-S · ImageNet | Top-1 accuracy: 79.8 -> 79.4 | Hurt |
| Apply dynamic top-down patch slimming | DeiT-B · ImageNet | Top-1 accuracy: 81.8 -> 81.6 | Hurt |
| Apply patch slimming | T2T-ViT-14 · ImageNet | FLOPs: 5.2 -> 3.1 | Improved |
| Use learned, layer-specific patch selection instead of uniform pruning (ablation) | DeiT-S · ImageNet | Top-1 accuracy: 77.2 -> 79.4 | Improved |
| Increase the global tolerant error to allow more patches to be pruned (ablation) | DeiT-S · ImageNet | FLOPs reduction | Improved |
| Prune patches in deeper layers rather than the same fraction in shallower layers (ablation) | DeiT-S · ImageNet | Top-1 accuracy | Improved |
| Select patches using propagated impact scores instead of random or current-layer attention scores (ablation) | DeiT-S · ImageNet | Top-1 accuracy | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- SCOP — SCOP: Scientific Control for Reliable Neural Network Pruning (NeurIPS 2020) · better on 2 of 2 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Synergistic Patch Pruning for Vision Transformer: Unifying Intra- & Inter-Layer Patch Importance (ICLR 2024 · similarity 0.89)
- Scaling Laws in Patchification: An Image Is Worth 50,176 Tokens And More (ICML 2025 · similarity 0.88)
- An Image is Worth More Than 16x16 Patches: Exploring Transformers on Individual Pixels (ICLR 2025 · similarity 0.87)
- Skip-Attention: Improving Vision Transformers by Paying Less Attention (ICLR 2024 · similarity 0.87)
- Transformer Fusion with Optimal Transport (ICLR 2024 · similarity 0.87)
- Slicing Vision Transformer for Flexible Inference (NeurIPS 2024 · similarity 0.87)
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.
