tensorflow 2.0:数据加载
1.对于一些小型常用的数据集，TensorFlow有相关的API可以调用——keras.datasets
其中有以下经典数据集
1.boston housing 波士顿房价
2.mnist/fasion mnist 手写数字集/时髦品集
3.cifar10/100 物象分类
4.imdb
2.使用 tf.data.Dataset 的好处：
1.既能让后面有迭代的方式，又能直接对数据（tensor类型）进行预处理，还能支持batch和多线程的方式处理
2.提供了 .shuffle（打散）, .map（预处理） 功能
现在我们假设 db 是个32*32的RGB三通道图片，即 [32,32,3]
.shuffle ：用于打乱数据集但不影响映射关系,将数据打乱，数值越大，混乱程度越大
db = tf.data.Dataset.from_tensor_slices( (x_test, y_test) )
db = db.shuffle(10000) # x_test,y_test映射关系不变
.map ： 用于使用预处理映射
def preprocess(x,y):
# 定义一个与处理函数 用于将numpy数据类型转化为Tensor的类型（dtype=float32）
x = tf.cast(x, dtype=tf.float32) / 255 # 将灰度级归一化
y = tf.cast(y, dtype=tf.float32)
y = tf.one_hot(y, depth=10) # 对数字编码 y 进行one_hot编码，10个0-1序列中只有一个1
return x, y
db2 = db.map(preprocess)
res = next(iter(db2)) # iter(db2):取得db2的迭代器，next(iter(db2))：迭代
.batch ：批处理
db3 = db2.batch(32) # (32张图片,32个label)为一个batch
res = next(iter(db3)) # 进行迭代
res[0].shape, res[1].shape # 分别是一个batch中图片格式与label格式的shape
(TensorShape([32,32,32,3]), TensorShape([32,1,10]))
# 图片格式是(32张，32*32大小，3个通道) # (32张图片对应的label,1个label——通常会squeeze掉，10个one_hot深度)
.repeat : 整个数据集的循环次数
db4 = db3.repeat() # 这样就是一直repeat迭代，死循环
db4 = db3.repeat(2) # 这个是迭代2次
For Eaxmple：
def prepare_mnist_features_and_labels(x,y):
x = tf.cast(x, tf.float32) / 255.0
y = tf.cast(y, tf.float64)
return x,y
def mnist_dataset():
(x, y),(x_val, y_val) = datasets.fashion_mnist.load_data() # 1.加载图像数据和通用数据（val指的是validation，测试数据集）
y = tf.one_hot(y, depth=10) # 2.数据 one_hot编码
y_val = tf.one_hot(y_val, depth=10) # label one_hot编码
ds = tf.data.Dataset.from_tensor_slices((x, y)) # 3.转换为Dataset类型
ds = ds.map(prepare_mnist_features_and_labels) # 4.预处理函数映射
ds = ds.shuffle(60000).batch(100) # 5.其他处理——如本处的前60000个打乱，100个为一个批次
ds_val = tf.data.Dataset.from_tensor_slices((x_val, y_val))
ds_val = ds_val.map(prepare_mnist_features_and_labels)
ds_val = ds_val.shuffle(10000).batch(100)
return ds, ds_val
3.现在我们来看看实战
import tensorflow as tf
from tensorflow.python import keras # 在pycharm中keras在tensorflow.python下
from tensorflow.keras import datasets
import os
os.environ['TF_CPP_MIN_LOG_LEVEL'] = '2'
# x: [60k, 28, 28], x_test: [10k, 28, 28]
# y: [60k], y_test: [10k]
(x, y), (x_test, y_test) = datasets.mnist.load_data()
# x: [0~255] => [0~1.]
x = tf.convert_to_tensor(x, dtype=tf.float32) / 255.
y = tf.convert_to_tensor(y, dtype=tf.int32)
x_test = tf.convert_to_tensor(x_test, dtype=tf.float32) / 255.
y_test = tf.convert_to_tensor(y_test, dtype=tf.int32)
print(x.shape, y.shape, x.dtype, y.dtype)
print(tf.reduce_min(x), tf.reduce_max(x))
print(tf.reduce_min(y), tf.reduce_max(y))
train_db = tf.data.Dataset.from_tensor_slices((x,y)).batch(128)
test_db = tf.data.Dataset.from_tensor_slices((x_test,y_test)).batch(128)
train_iter = iter(train_db)
sample = next(train_iter)
print('batch:', sample[0].shape, sample[1].shape)
# [b, 784] => [b, 256] => [b, 128] => [b, 10]
# [dim_in, dim_out], [dim_out]
w1 = tf.Variable(tf.random.truncated_normal([784, 256], stddev=0.1))
b1 = tf.Variable(tf.zeros([256]))
w2 = tf.Variable(tf.random.truncated_normal([256, 128], stddev=0.1))
b2 = tf.Variable(tf.zeros([128]))
w3 = tf.Variable(tf.random.truncated_normal([128, 10], stddev=0.1))
b3 = tf.Variable(tf.zeros([10]))
lr = 1e-3
for epoch in range(10): # iterate db for 10
for step, (x, y) in enumerate(train_db): # for every batch
# x:[128, 28, 28]
# y: [128]
# [b, 28, 28] => [b, 28*28]
x = tf.reshape(x, [-1, 28*28])
with tf.GradientTape() as tape: # tf.Variable
# x: [b, 28*28]
# h1 = x@w1 + b1
# [b, 784]@[784, 256] + [256] => [b, 256] + [256] => [b, 256] + [b, 256]
h1 = x@w1 + tf.broadcast_to(b1, [x.shape[0], 256])
h1 = tf.nn.relu(h1)
# [b, 256] => [b, 128]
h2 = h1@w2 + b2
h2 = tf.nn.relu(h2)
# [b, 128] => [b, 10]
out = h2@w3 + b3
# compute loss
# out: [b, 10]
# y: [b] => [b, 10]
y_onehot = tf.one_hot(y, depth=10)
# mse = mean(sum(y-out)^2)
# [b, 10]
loss = tf.square(y_onehot - out)
# mean: scalar
loss = tf.reduce_mean(loss)
# compute gradients
grads = tape.gradient(loss, [w1, b1, w2, b2, w3, b3])
# print(grads)
# w1 = w1 - lr * w1_grad
w1.assign_sub(lr * grads[0]) # assign_sub 相当于 -= 原地更新
b1.assign_sub(lr * grads[1])
w2.assign_sub(lr * grads[2])
b2.assign_sub(lr * grads[3])
w3.assign_sub(lr * grads[4])
b3.assign_sub(lr * grads[5])
if step % 100 == 0:
print(epoch, step, 'loss:', float(loss))
# test/evluation
# [w1, b1, w2, b2, w3, b3]
total_correct, total_num = 0, 0
for step, (x,y) in enumerate(test_db):
# [b, 28, 28] => [b, 28*28]
x = tf.reshape(x, [-1, 28*28])
# [b, 784] => [b, 256] => [b, 128] => [b, 10]
h1 = tf.nn.relu(x@w1 + b1)
h2 = tf.nn.relu(h1@w2 + b2)
out = h2@w3 + b3
# out : [b, 10] ~ R
# prob: [b, 10] ~ [0, 1]
prob = tf.nn.softmax(out, axis=1) # 在第一维度归一化
# [b, 10] => [b]
pred = tf.argmax(prob, axis=1) # 在第一维度选择索引号
pred = tf.cast(pred, dtype=tf.int32)
# y: [b]
# b: int32
# print(pred.dtype, y.dtype)
correct = tf.cast(tf.equal(pred, y), dtype=tf.int32)
coorect = tf.reduce_sum(correct)
# 原句是 totoal_correct += int(correct)，但会报错，好像说不能转化为标量，所以我只能取第一维数据了(correct[0])
# 即（TypeError: only size-1 arrays can be converted to Python scalars）
total_correct += int(correct[0]) # 正确了的个数 total_correct
total_num += x.shape[0] # 总的标签数
acc = total_correct / total_num
print("test acc:", acc)
Reference:
https://blog.csdn.net/weixin_43469047/article/details/90744677
