我正在从kaggle导入数据集，它有15个类，我只需要其中的10个类，我如何将这些类过滤到我的数据集中？
我正在尝试这个代码

image_size= 256
batch_size=8
channels=3
epochs=50

dataset = tf.keras.preprocessing.image_dataset_from_directory('/kaggle/input/plant-village/PlantVillage',
                                                              seed=123,
                                                              shuffle=True,
                                                              image_size=(image_size,image_size),
                                                              batch_size=batch_size)
dataset.class_names

字符串

结果是

找到20638个文件，属于15个类别。'Pepper__bell_Bacterial_spot'，'Pepper__bell_healthy'，'Potato_Early_blight'，'Potato_Late_blight'，'Potato_healthy'，'Tomato_Bacterial_spot'，'Tomato_Early_blight'，'Tomato_Late_blight'，'Tomato_Leaf_Mold'，'Tomato_Septoria_leaf_spot'，'Tomato_Spider_mites_Two_spotted_spider_mite'，“Tomato__Target_Spot”、“Tomato__Tomato_YellowLeaf__Curl_Virus”、“Tomato__Tomato_mosaic_virus”、“Tomato_healthy”]

我只希望这些条款

desired_classes = [
    'Tomato_Bacterial_spot',
    'Tomato_Early_blight',
    'Tomato_Late_blight',
    'Tomato_Leaf_Mold',
    'Tomato_Septoria_leaf_spot',
    'Tomato_Spider_mites_Two_spotted_spider_mite',
    'Tomato__Target_Spot',
    'Tomato__Tomato_YellowLeaf__Curl_Virus',
    'Tomato__Tomato_mosaic_virus',
    'Tomato_healthy'
]

型