用numpy.random python替换random.sample

我有40个文件夹，每个文件夹包含10张图像。此代码将每个图像文件夹分成两半，从这10张图像中随机选择5张用于训练，5张用于测试

for Class in allClasses:
path = folderPath + '/' + Class
allImages = os.listdir(path)
imageCount = len(allImages)
# seperating training images
randomlist = random.sample(range(imageCount), (int)(imageCount / 2))
trainingImagesForClass = [allImages[i] for i in randomlist]
trainingImages.append([readImage(path + '/' + i) for i in trainingImagesForClass])
# seperating test images
testList = [x for x in range(imageCount) if x not in randomlist]
testImagesForClass = [allImages[i] for i in testList]
testImages.append([readImage(path + '/' + i) for i in testImagesForClass])

这段代码工作得很好，但我想使用numpy random实现