metadata
license: apache-2.0
library_name: libreyolo
tags:
- object-detection
- rt-detr
- transformer
- real-time
- pytorch
datasets:
- coco
LibreYOLO RT-DETRv2-M*
RT-DETRv2-M* (r34vd) - 49.9 AP on COCO
Model Details
- Architecture: RT-DETRv2 (Real-Time Detection Transformer v2)
- Backbone: ResNet-34 (r34vd)
- Framework: PyTorch
- License: Apache 2.0
Performance
| Model | Dataset | Input Size | AP | AP50 | Params | FPS |
|---|---|---|---|---|---|---|
| RT-DETRv2-M* | COCO | 640 | 49.9 | 67.5 | 31M | 161 |
Usage
from libreyolo import LIBREYOLO
# Load model
model = LIBREYOLO("librertdetrms.pth", size="ms")
# Run inference
result = model("image.jpg", conf_thres=0.5)
# Access results
print(f"Detected {result['num_detections']} objects")
for box, score, cls in zip(result['boxes'], result['scores'], result['classes']):
print(f" Class {cls}: {score:.2f} @ {box}")
Installation
pip install libreyolo
Citation
@misc{lv2024rtdetrv2,
title={RT-DETRv2: Improved Baseline with Bag-of-Freebies for Real-Time Detection Transformer},
author={Wenyu Lv and Yian Zhao and Qinyao Chang and Kui Huang and Guanzhong Wang and Yi Liu},
year={2024},
eprint={2407.17140},
archivePrefix={arXiv},
primaryClass={cs.CV}
}