I have trained YOLOv10 on the DocLayNet dataset for this project. Below is the results table. Feel free to use our fine-tuned models, and please remember to cite YOLOv10, DocLayNet, and our repository. If you find this repository useful, don't forget to give it a 🌟!
- 03/06/2024: 🚀 Uploaded Fine-tuned (check the table below).
- 02/06/2024: 🤗 HuggingFace demo is live with YOLOv10-x fine-tuned weights.
The models were fine-tuned using 4xA100 GPUs on the Doclaynet-base dataset, which consists of 6910 training images, 648 validation images, and 499 test images.
| Model | mAP50 | mAP50-95 | Model Weights |
|---|---|---|---|
| YOLOv10-x | 0.924 | 0.740 | Download |
| YOLOv10-b | 0.922 | 0.732 | Download |
| YOLOv10-l | 0.921 | 0.732 | Download |
| YOLOv10-m | 0.917 | 0.737 | Download |
| YOLOv10-s | 0.905 | 0.713 | Download |
| YOLOv10-n | 0.892 | 0.685 | Download |
conda create -n yolov10 python=3.9
conda activate yolov10
git clone https://github.com/THU-MIG/yolov10.git
cd yolov10
pip install -r requirements.txt
pip install -e .
- YOLOv10
BibTeX
@article{wang2024yolov10,
title={YOLOv10: Real-Time End-to-End Object Detection},
author={Wang, Ao and Chen, Hui and Liu, Lihao and Chen, Kai and Lin, Zijia and Han, Jungong and Ding, Guiguang},
journal={arXiv preprint arXiv:2405.14458},
year={2024}
}
- DocLayNet
@article{doclaynet2022,
title = {DocLayNet: A Large Human-Annotated Dataset for Document-Layout Analysis},
doi = {10.1145/3534678.353904},
url = {https://arxiv.org/abs/2206.01062},
author = {Pfitzmann, Birgit and Auer, Christoph and Dolfi, Michele and Nassar, Ahmed S and Staar, Peter W J},
year = {2022}
}
LinkedIn: https://www.linkedin.com/in/omar-moured/