How to use from the
Use from the
ultralytics library
# Couldn't find a valid YOLO version tag.
# Replace XX with the correct version.
from ultralytics import YOLOvXX

model = YOLOvXX.from_pretrained("liberty666/yolo11n-chinese-signature")
source = 'http://images.cocodataset.org/val2017/000000039769.jpg'
model.predict(source=source, save=True)

yolo11n-chinese-signature

中文文档 手写签名区域 单类目标检测模型(YOLOv11-nano),类别:signature

下载与使用

命令行(推荐)

pip install huggingface_hub ultralytics
hf download liberty666/yolo11n-chinese-signature --local-dir ./my_model

Python

from huggingface_hub import hf_hub_download
from ultralytics import YOLO

path = hf_hub_download(repo_id="liberty666/yolo11n-chinese-signature", filename="best.pt")
model = YOLO(path)
model.predict("document.jpg", conf=0.25)

本项目脚本

python scripts/08_download_from_hub.py
python scripts/06_detect_and_crop.py --model hf://liberty666/yolo11n-chinese-signature

仓库文件

文件 说明
best.pt Ultralytics 最佳权重
yolo11n_signature.onnx ONNX 部署(若已导出)
data.yaml 数据集类别配置
train.yaml 训练超参备份
metrics/ 训练指标摘要

标签

  • ultralytics
  • yolo11
  • yolo11n
  • object-detection
  • signature-detection
  • chinese-document
  • chisig

最佳权重(best.pt,epoch 47)

指标
Recall 0.9516
mAP@50 0.9913

上传文件为 best.pt(非 last.pt)。 模型卡片勿将训练末轮 epoch 指标误认为部署权重性能。

引用

基于 ChiSig 与 YOLOv11。详见项目 Yolo_fit_guidance.md 与论文 Yan et al., CVPRW 2022.

Downloads last month
133
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support