训练集，测试集数量的确定

import osimport random

g_root_path = "E:/DeepLearning/Faster-RCNN-TensorFlow-Python3.5-master/data/VOCdevkit2007/VOC2007/"xmlfilepath = "E:\DeepLearning\Faster-RCNN-TensorFlow-Python3.5-master\data\VOCdevkit2007\VOC2007\Annotations"  # 标注文件存放路径saveBasePath = "E:/DeepLearning/Faster-RCNN-TensorFlow-Python3.5-master/data/VOCdevkit2007/VOC2007/ImageSets/Main/"  # ImageSets信息生成路径trainval_percent = 0.66train_percent = 0.5

os.chdir(g_root_path)total_xml = os.listdir(xmlfilepath)num = len(total_xml)xml_list = range(num)tv = int(num * trainval_percent)tr = int(tv * train_percent)trainval = random.sample(xml_list, tv)train = random.sample(trainval, tr)

print("train and val size", tv)print("train  size", tr)ftrainval = open(saveBasePath + "trainval.txt", "w")ftest = open(saveBasePath + "test.txt", "w")ftrain = open(saveBasePath + "train.txt", "w")fval = open(saveBasePath + "val.txt", "w")

for i in xml_list:    name = total_xml[i][:-4] + "\n"    if i in trainval:        ftrainval.write(name)        if i in train:            ftrain.write(name)        else:            fval.write(name)    else:        ftest.write(name)

ftrainval.close()ftrain.close()fval.close()ftest.close()

原文地址：https://www.cnblogs.com/lwjkz/p/11416599.html