Research map / Architectures and efficiency / Neural architecture search
Searching for BurgerFormer with Micro-Meso-Macro Space Design
Longxing Yang, Yu Hu, Shun Lu, Zihao Sun, Jilin Mei, Yinhe Han, Xiaowei Li
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 |
|---|---|---|---|
| hybrid sampling (SPOS + sandwich-style min/max macro sampling) instead of random search (ablation) | supernet training · ImageNet-1k, 1G FLOPs | Top-1 acc (%): +2.4 vs random search | Improved |
| hybrid sampling instead of SPOS (ablation) | supernet training · ImageNet-1k, 1G FLOPs | Top-1 acc (%): +1.0 vs SPOS | Improved |
| hybrid sampling instead of sandwich sampling (ablation) | supernet training · ImageNet-1k, 1G FLOPs | Top-1 acc (%): +0.5 vs Sandwich | Improved |
| include first Norm-Op-Norm-Act (bread) in hamburger block (ablation) | BurgerFormer-Tiny · ImageNet-1k | Top-1 acc (%): +0.3 (78.0 vs 77.7) | Improved |
| include second Norm-Op-Norm-Act (bread) in hamburger block (ablation) | BurgerFormer-Tiny · ImageNet-1k | Top-1 acc (%): +0.8 (78.0 vs 77.2) | Improved |
| include meat Norm-Op-Norm-Act in hamburger block (ablation) | BurgerFormer-Tiny · ImageNet-1k | Top-1 acc (%): +1.1 (78.0 vs 76.9) | Improved |
| searched BurgerFormer vs manually designed and auto-searched Transformers at similar FLOPs | BurgerFormer-Tiny/Base/Large · ImageNet-1k | Top-1 acc (%): BurgerFormer-Base 82.7 vs Swin-T 81.3 at fewer FLOPs | Improved |
| BurgerFormer-Large matches Swin-S accuracy at fewer FLOPs | BurgerFormer-Large · ImageNet-1k | Top-1 acc (%): 83.0 vs 83.0 with 25% fewer FLOPs | No clear effect |
| BurgerFormer-Base as backbone instead of PVT-Small/Swin-T in detectors | RetinaNet and Mask R-CNN on COCO · COCO | AP: RetinaNet 41.2 vs PVT-Small 40.4 (better); Mask R-CNN box AP 44.0 vs Swin-T 43.7 (better), but Swin-T is 48.0M params vs | Mixed |
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.
- AutoFormer-tiny — AutoFormer: Searching Transformers for Visual Recognition (ICCV 2021) · better on 1 of 1 shared result
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
- AutoFormer: Searching Transformers for Visual Recognition (ICCV 2021 · similarity 0.87)
- MetaFormer Is Actually What You Need for Vision (CVPR 2022 · similarity 0.86)
- Searching the Search Space of Vision Transformer (NeurIPS 2021 · similarity 0.85)
- Masked Distillation Advances Self-Supervised Transformer Architecture Search (ICLR 2024 · similarity 0.84)
- Auto-Prox: Training-Free Vision Transformer Architecture Search via Automatic Proxy Discovery (AAAI 2024 · similarity 0.83)
- Mesoscopic Insights: Orchestrating Multi-Scale & Hybrid Architecture for Image Manipulation Localization (AAAI 2025 · similarity 0.83)
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.
