tensorflow(2)——線性迴歸
學習《Tensorflow入門教程》記錄
import numpy as np import tensorflow as tf import matplotlib.pyplot as plt # 隨機生成1000個點,圍繞在y=0.1x+0.3的直線周圍 num_points = 1000 vectors_set = [] for i in range(num_points): x1 = np.random.normal(0.0, 0.55) y1 = x1 * 0.1 + 0.3 + np.random.normal(0.0, 0.03) vectors_set.append([x1, y1]) # 生成一些樣本 x_data = [v[0] for v in vectors_set] y_data = [v[1] for v in vectors_set] plt.scatter(x_data,y_data,c='r') plt.show()
結果如下所示:
對上面的資料集進行訓練:
# 生成1維的W矩陣,取值是[-1,1]之間的隨機數 W = tf.Variable(tf.random_uniform([1], -1.0, 1.0), name='W') # 生成1維的b矩陣,初始值是0 b = tf.Variable(tf.zeros([1]), name='b') # 經過計算得出預估值y y = W * x_data + b # 以預估值y和實際值y_data之間的均方誤差作為損失 loss = tf.reduce_mean(tf.square(y - y_data), name='loss') # 採用梯度下降法來優化引數 optimizer = tf.train.GradientDescentOptimizer(0.5) # 訓練的過程就是最小化這個誤差值 train = optimizer.minimize(loss, name='train') sess = tf.Session() init = tf.global_variables_initializer() sess.run(init) # 初始化的W和b是多少 print ("W =", sess.run(W), "b =", "seess.run(b),s =", sess.run(loss)) # 執行20次訓練 for step in range(20): sess.run(train) # 輸出訓練好的W和b print ("W =", sess.run(W), "b =", sess.run(b), "loss =", sess.run(loss))
訓練的結果為:
W = [-0.68253565] b = seess.run(b),s = 0.28643712
W = [-0.44002083] b = [0.32005146] loss = 0.08808484
W = [-0.2786493] b = [0.31371173] loss = 0.04372422
W = [-0.16553827] b = [0.30949324] loss = 0.021932669
W = [-0.08626061] b = [0.30653635] loss = 0.011227837
W = [-0.03069621] b = [0.3044639] loss = 0.005969218
W = [0.00824796] b = [0.30301136] loss = 0.003385987
W = [0.03554329] b = [0.3019933] loss = 0.0021170068
W = [0.05467413] b = [0.30127975] loss = 0.0014936357
W = [0.06808263] b = [0.30077967] loss = 0.0011874122
W = [0.07748042] b = [0.30042914] loss = 0.0010369839
W = [0.08406717] b = [0.30018348] loss = 0.0009630878
W = [0.08868372] b = [0.30001128] loss = 0.00092678727
W = [0.09191938] b = [0.2998906] loss = 0.000908955
W = [0.09418721] b = [0.299806] loss = 0.0009001951
W = [0.09577668] b = [0.29974672] loss = 0.00089589204
W = [0.09689073] b = [0.29970518] loss = 0.00089377817
W = [0.09767154] b = [0.29967606] loss = 0.0008927398
W = [0.0982188] b = [0.29965565] loss = 0.0008922296
W = [0.09860236] b = [0.29964134] loss = 0.000891979
W = [0.09887119] b = [0.2996313] loss = 0.00089185586
由結果可知,損失越來越小,W和b越來越逼近實際值。