tensorflow(十七)：數據的加載：map()、shuffle()、tf.data.Dataset.from_tensor_slices()

一、數據集簡介

 

 

 

 

二、MNIST數據集介紹

 

 三、CIFAR 10/100數據集介紹

 

 

 

 四、tf.data.Dataset.from_tensor_slices()

 

 五、shuffle()隨機打散

 

 六、map()數據預處理

 

 

 

 

 

 

 七、實戰

import tensorflow as tf
import tensorflow.keras as keras
import os

os.environ['TF_CPP_MIN_LOG_LEVEL'] = '2'

def prepare_mnist_features_and_labels(x,y):
    x = tf.cast(x, tf.float32) / 255.0
    y = tf.cast(y, tf.int64)
    return x,y

def mnist_dataset():
    (x,y), (x_test,y_test) = keras.datasets.fashion_mnist.load_data() #numpy中的格式

    y = tf.one_hot(y, depth=10)                     #[10k] ==> [10k,10]的tensor
    y_test = tf.one_hot(y_test, depth=10)

    ds = tf.data.Dataset.from_tensor_slices((x,y))
    ds = ds.map(prepare_mnist_features_and_labels)  #數據預處理,注意：tf.map中傳進的參數
    ds = ds.shuffle(60000).batch(100)               #隨機打散，讀取一個batch的樣本

    ds_val = tf.data.Dataset.from_tensor_slices((x_test,y_test))
    ds_val = ds_val.map(prepare_mnist_features_and_labels)
    ds_val = ds_val.shuffle(10000).batch(100)
    return ds, ds_val


def main():
    ds, ds_val = mnist_dataset()

    print("訓練集信息如下：")
    iteration_ds = iter(ds)
    iter_ds = next(iteration_ds)
    print(iter_ds[0].shape, iter_ds[1].shape)

    print("測試集信息如下：")
    iteration_ds_val = iter(ds_val)
    iter_ds_val = next(iteration_ds_val)
    print(iter_ds_val[0].shape, iter_ds_val[1].shape)

if __name__ == '__main__':
    main()