测试（张量）- 实战 - 好文

目录
import tensorflow as tf from tensorflow import keras from tensorflow.keras
import datasets import os # do not print irrelevant information #
os.environ['TF_CPP_MIN_LOG_LEVEL'] = '2' # x: [60k,28,28], [10,28,28] # y:
[60k], [10k] (x, y), (x_test, y_test) = datasets.mnist.load_data() # transform
Tensor # 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) f'x.shape: {x.shape}, y.shape:
{y.shape}, x.dtype: {x.dtype}, y.dtype: {y.dtype}' "x.shape: (60000, 28, 28),
y.shape: (60000,), x.dtype: <dtype: 'float32'>, y.dtype: <dtype: 'int32'>"
f'min_x: {tf.reduce_min(x)}, max_x: {tf.reduce_max(x)}' 'min_x: 0.0, max_x: 1.0'
f'min_y: {tf.reduce_min(y)}, max_y: {tf.reduce_max(y)}' 'min_y: 0, max_y: 9' #
batch of 128 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) f'batch:
{sample[0].shape,sample[1].shape}' 'batch: (TensorShape([128, 28, 28]),
TensorShape([128]))' # [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])) # learning rate lr = 1e-3
for epoch in range(10): # iterate db for 10 # tranin every train_db for step,
(x, y) in enumerate(train_db): # x: [128,28,28] # y: [128] # [b,28,28] ==>
[b,28*28] x = tf.reshape(x, [-1, 28 * 28]) with tf.GradientTape( ) as tape: #
only data types of tf.variable are logged # 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 = x@w2 + b2 # b2 can broadcast automatic 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]) # w1 = w1 - lr * w1_grad # w1 = w1 - lr * grads[0] # not in
situ update # in situ update 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(f'epoch:{epoch}, step: {step}, loss:{float(loss)}') # [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) # [b,10] ==> [b] pred = tf.argmax(prob, axis=1) pred
= tf.cast(pred, dtype=tf.int32) # y: [b] # [b], int32 correct =
tf.cast(tf.equal(pred, y), dtype=tf.int32) correct = tf.reduce_sum(correct)
total_correct += int(correct) total_num += x.shape[0] acc = total_correct /
total_num print(f'test acc: {acc}') epoch:0, step: 0, loss:0.4985736012458801
epoch:0, step: 100, loss:0.22939381003379822 epoch:0, step: 200,
loss:0.2018660604953766 epoch:0, step: 300, loss:0.18181894719600677 epoch:0,
step: 400, loss:0.1831897795200348 test acc: 0.1153 epoch:1, step: 0,
loss:0.1674182116985321 epoch:1, step: 100, loss:0.17186065018177032 epoch:1,
step: 200, loss:0.16210347414016724 epoch:1, step: 300, loss:0.1499405801296234
epoch:1, step: 400, loss:0.15070970356464386 test acc: 0.1769 epoch:2, step: 0,
loss:0.14020009338855743 epoch:2, step: 100, loss:0.14754906296730042 epoch:2,
step: 200, loss:0.13924123346805573 epoch:2, step: 300, loss:0.1308508813381195
epoch:2, step: 400, loss:0.1306917369365692 test acc: 0.235 epoch:3, step: 0,
loss:0.12297296524047852 epoch:3, step: 100, loss:0.13165466487407684 epoch:3,
step: 200, loss:0.12420644611120224 epoch:3, step: 300, loss:0.1179303377866745
epoch:3, step: 400, loss:0.11716334521770477 test acc: 0.2927 epoch:4, step: 0,
loss:0.11098697036504745 epoch:4, step: 100, loss:0.12046296894550323 epoch:4,
step: 200, loss:0.11333265155553818 epoch:4, step: 300,
loss:0.10868857055902481 epoch:4, step: 400, loss:0.10756760835647583 test acc:
0.3386 epoch:5, step: 0, loss:0.1022152453660965 epoch:5, step: 100,
loss:0.1120707243680954 epoch:5, step: 200, loss:0.10497119277715683 epoch:5,
step: 300, loss:0.10168357938528061 epoch:5, step: 400,
loss:0.10033649206161499 test acc: 0.379 epoch:6, step: 0,
loss:0.09566861391067505 epoch:6, step: 100, loss:0.10548736900091171 epoch:6,
step: 200, loss:0.09834134578704834 epoch:6, step: 300, loss:0.0961376205086708
epoch:6, step: 400, loss:0.09474694728851318 test acc: 0.4168 epoch:7, step: 0,
loss:0.09054075181484222 epoch:7, step: 100, loss:0.1001550704240799 epoch:7,
step: 200, loss:0.09303966909646988 epoch:7, step: 300,
loss:0.09163998067378998 epoch:7, step: 400, loss:0.09031815826892853 test acc:
0.453 epoch:8, step: 0, loss:0.08635123074054718 epoch:8, step: 100,
loss:0.0957597866654396 epoch:8, step: 200, loss:0.08867798745632172 epoch:8,
step: 300, loss:0.08790989965200424 epoch:8, step: 400,
loss:0.08668653666973114 test acc: 0.4831 epoch:9, step: 0,
loss:0.08282895386219025 epoch:9, step: 100, loss:0.09203790128231049 epoch:9,
step: 200, loss:0.0850382000207901 epoch:9, step: 300, loss:0.08473993837833405
epoch:9, step: 400, loss:0.0835738554596901 test acc: 0.5065
15 测试（张量）- 实战

热门工具换一换