tensorflow2.0学习笔记（二）

有注释，复制到Pycharm或者VScode里面看

import tensorflow as tf from tensorflow import keras from tensorflow.keras import datasets import os os.environ['TF_CPP_MIN_LOG_LEVEL'] = '2' # x: [60000, 28, 28], # y: [60000] (x, y), _ = datasets.mnist.load_data() # x: [0-255] => [0-1.0] x = tf.convert_to_tensor(x, dtype=tf.float32) / 255. y = tf.convert_to_tensor(y, dtype=tf.int32) print("x.shape:{} y.shape:{} x.dtype:{} y.dtype:{}".format(x.shape, y.shape, x.dtype, y.dtype)) print(tf.reduce_min(x), tf.reduce_max(x)) # tf.reduce_max(x)，计算一个张量的各个维度上元素的最大值 print(tf.reduce_min(y), tf.reduce_max(y)) x.shape:(60000, 28, 28) y.shape:(60000,) x.dtype:<dtype: 'float32'> y.dtype:<dtype: 'int32'> tf.Tensor(0.0, shape=(), dtype=float32) tf.Tensor(1.0, shape=(), dtype=float32) tf.Tensor(0, shape=(), dtype=int32) tf.Tensor(9, shape=(), dtype=int32) train_db = tf.data.Dataset.from_tensor_slices((x,y)).batch(128) train_iter = iter(train_db) # 用来生成迭代器。 sample = next(train_iter) print('batch:', sample[0].shape, sample[1].shape) batch: (128, 28, 28) (128,) # [b, 784] => [b, 256] => [b, 128] => [b, 10] # [dim_in, dim_out], [dim_out] # tf.truncated_normal函数，生成的值是在距离均值两个标准差范围之内的，（μ-2σ，μ+2σ），保证了生成的值都在均值附近。避免梯度爆炸产生NAN 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]) 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)) 0 0 loss: 0.30855444073677063 0 100 loss: 0.1943640410900116 0 200 loss: 0.20688548684120178 0 300 loss: 0.18792247772216797 0 400 loss: 0.1854521483182907 1 0 loss: 0.15025171637535095 1 100 loss: 0.14979198575019836 1 200 loss: 0.1626807451248169 1 300 loss: 0.15183447301387787 1 400 loss: 0.15254397690296173 2 0 loss: 0.1266193687915802 2 100 loss: 0.1301368772983551 2 200 loss: 0.1398109793663025 2 300 loss: 0.13149653375148773 2 400 loss: 0.13328148424625397 3 0 loss: 0.1121378168463707 3 100 loss: 0.11777482181787491 3 200 loss: 0.1255255490541458 3 300 loss: 0.11861814558506012 3 400 loss: 0.12068285793066025 4 0 loss: 0.10235831886529922 4 100 loss: 0.10916677862405777 4 200 loss: 0.11564607918262482 4 300 loss: 0.1097228154540062 4 400 loss: 0.11179308593273163 5 0 loss: 0.09519632160663605 5 100 loss: 0.10276556015014648 5 200 loss: 0.10830307006835938 5 300 loss: 0.10308554023504257 5 400 loss: 0.10502723604440689 6 0 loss: 0.0897449478507042 6 100 loss: 0.09781351685523987 6 200 loss: 0.10259167104959488 6 300 loss: 0.09793788939714432 6 400 loss: 0.09971939027309418 7 0 loss: 0.08542229235172272 7 100 loss: 0.09375528991222382 7 200 loss: 0.09798217564821243 7 300 loss: 0.0937323346734047 7 400 loss: 0.09538793563842773 8 0 loss: 0.08188784122467041 8 100 loss: 0.09030032157897949 8 200 loss: 0.09406483173370361 8 300 loss: 0.09022961556911469 8 400 loss: 0.09179739654064178 9 0 loss: 0.07893332093954086 9 100 loss: 0.08734345436096191 9 200 loss: 0.09066981077194214 9 300 loss: 0.08724066615104675 9 400 loss: 0.08874273300170898

参考：龙龙老师