Tensorflow K折交叉验证使用方法

K折交叉验证

 

 

imageList,labelList,totalData=getData.genImageList(csvname)
    index = [i for i in range(len(imageList))]
    random.shuffle(index)
    imageList = imageList[index]
    labelList = labelList[index]
    kf = KFold(k)
    print(k,'folds finished splited,start trainning!')
    i=0
    for train_index, test_index in kf.split(imageList):
        i=i+1
        modelDir='./Model/model_'+str(i)+'/'
        getData.mkdir(modelDir)
        print('model 路径创建成功——',modelDir)
        train_X, train_y = imageList[train_index], labelList[train_index]
        test_X, test_y = imageList[test_index], labelList[test_index]
        #使用上边的新的train test数据进行训练

先需要根据获取数据的imagelist，labellist ,然后定义K

  for train_index, test_index in kf.split(imageList):

遍历，然后获取每一折的数据用作训练，random.shuffle用于打乱数据

random.shuffle(index)
    imageList = imageList[index]
    labelList = labelList[index]