二 . voc 数据集转yolov5

一. 数据集拆分 train&val

根据第一步得到如下数据集组成结构:

 由下代码变成:

import xml.etree.ElementTree as ET
import os


sets = ['train', 'val']  # 需要转换训练集. 验证集
Root = '../datasets/voc-end/' # 数据集目录
classes = ["person", "chef_uniform", "voc_clothes"]  # 修改为自己的label


def convert(size, box):
    dw = 1./(size[0]) 
    dh = 1./(size[1]) 
    x = (box[0] + box[1])/2.0 - 1
    y = (box[2] + box[3])/2.0 - 1
    w =