深度学习训练集问题

小弟是深度学习方面新手

最近想试试图像辨识但遇到了难题

我想辨认的三种东西分别是lad, lcx, rca

这三个类型的资料夹下分别又有100个资料夹里面是5张jpg档

train ->floder
-----lad ->floder
----------1 ->floder
---------------v1 v2 v3 v4 v5 ->jpg
----------2 ->floder
---------------v1 v2 v3 v4 v5 ->jpg
----------3 ->floder
---------------v1 v2 v3 v4 v5 ->jpg
-----lcx ->floder
----------1 ->floder
---------------v1 v2 v3 v4 v5 ->jpg
----------2->floder
---------------v1 v2 v3 v4 v5 ->jpg
----------3->floder
---------------v1 v2 v3 v4 v5 ->jpg

test 和 vaildation 资料夹也是同样结构

一般的文章或书籍所教之方法的资料结构都是
train ->floder
-----lad ->floder
----------1->jpg
----------2->jpg
----------3->jpg
----------4->jpg
lcx ->floder
----------1->jpg
----------2->jpg
----------3->jpg
----------4->jpg

5张图片放置在"rcx", "lad", "lcx"资料夹底下的"1", "2", "3", "4", "5" 资料夹

请问各位前备我该用哪种语法让模型知道训练的图片在资料夹下

我用的是python的tensorflow

程序码如下:

labels = ['lad', 'rca', 'lcx']
def get_data(data_dir):
    data = [] 
    for label in labels: 
        path = os.path.join(data_dir, label)
        class_num = labels.index(label)
        dirs = os.listdir(path)
        for file in dirs:
            pic_dir=os.path.join(path,file)
            for i in os.listdir(pic_dir):
                image_dir=os.path.join(pic_dir,i)
                img1 = cv2.imread(os.path.join(path, image_dir))
                img1 = cv2.cvtColor(img1, cv2.COLOR_RGB2BGR)
                data.append([img1, class_num])
    return np.array(data)
    train = get_data('C:/')
    test = get_data('C:/')
x_train = []
y_train = []
x_val = []
y_val = []

for feature, label in train:
    x_train.append(feature)
    y_train.append(label)
    
for feature, label in test:
    x_val.append(feature)
    y_val.append(label)
        
x_train = np.array(x_train) / 255
x_val = np.array(x_val) / 255
        
x_train.reshape(-1, img_size, img_size, 1)
y_train = np.array(y_train)
        
x_val.reshape(-1, img_size, img_size, 1)
y_val = np.array(y_val)

感谢大家的指教!

每个留言我都会认真看!!