Research map / Detection and segmentation / CNN features
YOLOv7: Trainable Bag-of-Freebies Sets New State-of-the-Art for Real-Time Object Detectors
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 |
|---|---|---|---|
| compound scaling (depth and transition width) vs width-only scaling (ablation) | YOLOv7-X light · MS COCO val 2017 | AP val: +0.5 | Helped |
| compound scaling vs depth-only scaling (ablation) | YOLOv7-X light · MS COCO val 2017 | AP val: +0.2 | Helped |
| lead head guided coarse-to-fine label assigner for auxiliary head (ablation) | YOLOv7-E6 · MS COCO val 2017 | AP val: +0.3 | Helped |
| independent label assigner for auxiliary head (vs no aux head) (ablation) | YOLOv7-E6 · MS COCO val 2017 | AP val: +0.2 | Helped |
| upper bound constraint on auxiliary decoder objectness (ablation) | YOLOv7-E6 auxiliary head · MS COCO val 2017 | AP val: 0.0 | No clear effect |
| partial auxiliary head connected before merge cardinality in E-ELAN (ablation) | YOLOv7-E6E · MS COCO val 2017 | AP val: +0.5 | Helped |
| planned re-parameterized RepRCSP (no identity connection) in reversed CSPDarknet (ablation) | YOLOR-CSP · MS COCO val 2017 | AP val: +0.1 | Helped |
| planned re-parameterized RepRCSP in YOLOR-W6 backbone (ablation) | YOLOR-W6 · MS COCO val 2017 | AP val: +0.03 | Helped |
| RepCSP (RepConv with identity) in YOLOR-W6 backbone (ablation) | YOLOR-W6 · MS COCO val 2017 | AP val: -0.15 | Hurt |
| planned RepConv placement in 3-stacked ELAN (Figure (a) and (e) placements) (ablation) | 3-stacked ELAN · MS COCO val 2017 | AP val: -0.08 to +0.07 | Mixed |
| YOLOv7 vs YOLOv4 (architecture plus trainable bag-of-freebies) | YOLOv7 · MS COCO val 2017 | AP val: +1.5 | Helped |
| YOLOv7-tiny vs YOLOv4-tiny at 416 input | YOLOv7-tiny · MS COCO val 2017 | AP val: +10.3 | Helped |
Closest papers by meaning
- YOLOv10: Real-Time End-to-End Object Detection (NeurIPS 2024 · similarity 0.90)
- YOLOv12: Attention-Centric Real-Time Object Detectors (NeurIPS 2025 · similarity 0.87)
- Mamba YOLO: A Simple Baseline for Object Detection with State Space Model (AAAI 2025 · similarity 0.86)
- DETRs Beat YOLOs on Real-time Object Detection (CVPR 2024 · similarity 0.85)
- Roboflow100-VL: A Multi-Domain Object Detection Benchmark for Vision-Language Models (NeurIPS 2025 · similarity 0.84)
- RemDet: Rethinking Efficient Model Design for UAV Object Detection (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.