Skip to content

YOLO3 动漫人脸检测 (Based on keras and tensorflow) 2019-1-19

License

Notifications You must be signed in to change notification settings

wmylxmj/YOLO-V3-IOU

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

69 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

YOLO-V3-IOU

YOLO3 动漫人脸检测 2019-1-19

环境

  • tensorflow-gpu 1.12.0
  • Keras 2.2.4

注意

数据集的准备

训练模型

  • 加载权重,将权重h5文件放入models文件夹
  • 最新权重下载地址:https://pan.baidu.com/s/1QKODAhRU_Rw5lpf8NcIR4g
  • 若重新开始训练,将load_pretrained置为False
  • 设置学习率,若重新训练,建议设0.001,若加载权重,建议设较小的学习率
  • 运行train.py

预测

  • 加载权重,将训练好的权重h5文件放入models文件夹
  • 运行run.py,对数据集进行预测输出,输出在outputs文件夹中
  • 预测新的图片
  1. from predict import YOLO
  2. from PIL import Image
  3. yolo = YOLO()
  4. img = Image.open(r'./test.jpg')
  5. output = yolo.detect_image(img)
  6. output.show()

Demo

模型

Releases

No releases published

Packages

No packages published

Languages