Research map / Architectures and efficiency / Attention and transformers
Vision Permutator: A Permutable MLP-Like Architecture for Visual Recognition
Qibin Hou, Zihang Jiang, Li Yuan, Ming-Ming Cheng, Shuicheng Yan, Jiashi Feng
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 |
|---|---|---|---|
| smaller initial patch size (16 to 14) (ablation) | ViP-Small · ImageNet-1k val | top-1 accuracy (%): 79.8 -> 80.5 | Improved |
| add fine-level tokens (7x7 patches plus 4 Permutators) (ablation) | ViP-Small · ImageNet-1k val | top-1 accuracy (%): 80.5 -> 81.5 | Improved |
| encode fine-level tokens (ablation) | ViP-Small · ImageNet-1k val | throughput (img/s): 789 img/s -> 719 img/s | Hurt |
| scale model from Small/7 to Medium/7 (ablation) | ViP · ImageNet-1k val | top-1 accuracy (%): 81.5 -> 82.7 | Improved |
| scale model from Medium/7 to Large/7 (ablation) | ViP · ImageNet-1k val | top-1 accuracy (%): 82.7 -> 83.2 | Improved |
| add Random Augmentation (ablation) | ViP-Small/14 · ImageNet-1k val | top-1 accuracy (%): 75.3 -> 77.7 (+2.4) | Improved |
| add CutOut (ablation) | ViP-Small/14 · ImageNet-1k val | top-1 accuracy (%): 77.7 -> 78.0 (+2.7) | Improved |
| add MixUp (ablation) | ViP-Small/14 · ImageNet-1k val | top-1 accuracy (%): 78.0 -> 80.2 (+4.9) | Improved |
| add CutMix (ablation) | ViP-Small/14 · ImageNet-1k val | top-1 accuracy (%): 80.2 -> 80.6 (+5.3) | Improved |
| add height information encoding (ablation) | ViP-Small/14 · ImageNet-1k val | top-1 accuracy (%): 72.8 -> 80.2 | Improved |
| add width information encoding (ablation) | ViP-Small/14 · ImageNet-1k val | top-1 accuracy (%): 72.7 -> 80.2 | Improved |
| use Weighted Permute-MLP instead of vanilla Permute-MLP (ablation) | ViP-Small/14 · ImageNet-1k val | top-1 accuracy (%): 80.2 -> 80.6 | Improved |
Closest papers by meaning
- DeepPermNet: Visual Permutation Learning (CVPR 2017 · similarity 0.90)
- SparseFormer: Sparse Visual Recognition via Limited Latent Tokens (ICLR 2024 · similarity 0.86)
- Linear Differential Vision Transformer: Learning Visual Contrasts via Pairwise Differentials (NeurIPS 2025 · similarity 0.86)
- LookWhere? Efficient Visual Recognition by Learning Where to Look and What to See from Self-Supervision (NeurIPS 2025 · similarity 0.86)
- Perception Encoder: The best visual embeddings are not at the output of the network (NeurIPS 2025 · similarity 0.86)
- AsCAN: Asymmetric Convolution-Attention Networks for Efficient Recognition and Generation (NeurIPS 2024 · similarity 0.85)
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.
