tensorFlow2基本操作（三）

5 数据加载

keras.datasets:

boston housing

mnist/fasion mnist

cifar10/100

imdb

1)载入cifar100

import tensorflow as tf 
from tensorflow import keras

(x, y), (x_test, y_test) = keras.datasets.mnist.load_data()

x.shape

(60000, 28, 28)

y.shape

(60000,)

x.min(),x.max(),x.mean()

(0, 255, 33.318421449829934)

x_test.shape,y_test.shape

((10000, 28, 28), (10000,))

y[:4]

array([5, 0, 4, 1], dtype=uint8)

y_onehot = tf.one_hot(y,depth = 10)

y_onehot[:2]

<tf.Tensor: id=8, shape=(2, 10), dtype=float32, numpy=
array([[0., 0., 0., 0., 0., 1., 0., 0., 0., 0.],
       [1., 0., 0., 0., 0., 0., 0., 0., 0., 0.]], dtype=float32)>

2)载入cifar10/100

(x,y) = (x_test,y_test) = keras.datasets.cifar10.load_data()
#此处本人暂停了，因为下载速度实在是太慢。其中遇到个小问题，网址不对，后来进入python文件更新了网址。

Downloading data from https://www.cs.toronto.edu/~kriz/cifar-10-python.tar.gz
  4890624/170498071 [..............................] - ETA: 9:40:41

import tensorflow as tf 
from tensorflow import keras

(x,y),(x_test,y_test) = keras.datasets.cifar10.load_data()       
#因为本人直接在网上下载好了数据集，并且直接拷贝在了目标文件夹‘C:\Users\wanfuchun\.keras\datasets’，因此此处加载特别快

x.shape,y.shape,x_test.shape,y_test.shape

((50000, 32, 32, 3), (50000, 1), (10000, 32, 32, 3), (10000, 1))

x.min(),x.max()

(0, 255)

y[:4]

array([[6],
       [9],
       [9],
       [4]], dtype=uint8)

db = tf.data.Dataset.from_tensor_slices(x_test)

next(iter(db)).shape

TensorShape([32, 32, 3])

db = tf.data.Dataset.from_tensor_slices((x_test,y_test))

next(iter(db))[0].shape

TensorShape([32, 32, 3])

next(iter(db))[1].shape

TensorShape([1])

3）Dataset api的shuffle打散功能

db = db.shuffle(10000)

4).map 数据预处理功能

def preprocess(x,y):
    x = tf.cast(x,dtype=tf.float32)/255.
    y = tf.cast(y,dtype=tf.int32)
    y = tf.one_hot(y,depth = 10)
    return x,y
db2 = db.map(preprocess)
res = next(iter(db2))

res[0].shape,res[1].shape

(TensorShape([32, 32, 3]), TensorShape([1, 10]))

res[1][:2] 

<tf.Tensor: id=221, shape=(1, 10), dtype=float32, numpy=array([[0., 0., 0., 0., 1., 0., 0., 0., 0., 0.]], dtype=float32)>

db3 = db2.batch(32)

res = next(iter(db3))

res[0].shape,res[1].shape

(TensorShape([32, 32, 32, 3]), TensorShape([32, 1, 10]))

6 误差计算

1）熵的概念

import tensorflow as tf

a = tf.fill([4],0.25)

a * tf.math.log(a)/tf.math.log(