Research map / Detection and segmentation / CNN features
You Only Look at One Sequence: Rethinking Transformer in Vision through Object Detection
Read the paperCheck 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 |
|---|---|---|---|
| Pre-train on ImageNet-1k instead of training from random initialization | YOLOS-Ti on COCO · COCO val | AP: 19.7 -> 28.7 | Helped |
| Increase supervised pre-training from 200 to 300 epochs (ablation) | YOLOS-Ti · COCO val | AP: 26.9 -> 28.7 | Helped |
| Increase supervised pre-training from 200 to 300 epochs (ablation) | YOLOS-S · COCO val | AP: 36.1 -> 36.1 | No clear effect |
| Add transformer-specific distillation during ImageNet pre-training (ablation) | YOLOS-Ti and YOLOS-S · COCO val | AP: 28.7 -> 29.7; 36.1 -> 37.2 | Helped |
| Use DINO rather than MoCo-v3 self-supervised pre-training (ablation) | YOLOS-S · COCO val | AP: 33.6 -> 36.2 | Helped |
| Apply uniform compound scaling rather than width scaling (ablation) | YOLOS-Ti scaled to about 4.5 GFLOPs · COCO val | AP: 36.1 -> 36.2 | No clear effect |
| Apply fast scaling rather than width scaling (ablation) | YOLOS-Ti scaled to about 4.5 GFLOPs · COCO val | AP: 36.1 -> 37.6 | Helped |
| Freeze the detection tokens instead of learning them (ablation) | YOLOS-Ti · COCO val | AP: 28.7 -> 28.3 | Hurt |
| Freeze the detection tokens instead of learning them (ablation) | YOLOS-S · COCO val | AP: 36.1 -> 36.4 | Helped |
| Use one large first-layer positional embedding and omit intermediate positional embeddings (ablation) | YOLOS-Ti · COCO val | AP: 28.3 -> 28.7 | Helped |
| Use one large first-layer positional embedding and omit intermediate positional embeddings (ablation) | YOLOS-S · COCO val | AP: 36.1 -> 36.6 | Helped |
| Apply dropout or stochastic depth during COCO fine-tuning (ablation) | YOLOS · COCO object detection | AP | Hurt |
Closest papers by meaning
- Roboflow100-VL: A Multi-Domain Object Detection Benchmark for Vision-Language Models (NeurIPS 2025 · similarity 0.87)
- TGBFormer: Transformer-GraphFormer Blender Network for Video Object Detection (AAAI 2025 · similarity 0.87)
- YOLOv10: Real-Time End-to-End Object Detection (NeurIPS 2024 · similarity 0.86)
- Mamba YOLO: A Simple Baseline for Object Detection with State Space Model (AAAI 2025 · similarity 0.86)
- YOLO-RD: Introducing Relevant and Compact Explicit Knowledge to YOLO by Retriever-Dictionary (ICLR 2025 · similarity 0.86)
- DETRs Beat YOLOs on Real-time Object Detection (CVPR 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.