tensorflow1.x学习之 5-构建非线性回归模型

原链接

回归模型

import numpy as np
import pandas as pd
import tensorflow as tf
import numpy as np
import matplotlib.pyplot as plt

使用numpy 生成均匀分布的200个数据点,包含首尾数据

生成从-0.5 到 0.5 均匀分布的200个数据点，包含首尾数据

x_data = np.linspace(-0.5,0.5,200)

x_data

array([-0.5       , -0.49497487, -0.48994975, -0.48492462, -0.4798995 ,
       -0.47487437, -0.46984925, -0.46482412, -0.45979899, -0.45477387,
       -0.44974874, -0.44472362, -0.43969849, -0.43467337, -0.42964824,
       -0.42462312, -0.41959799, -0.41457286, -0.40954774, -0.40452261,
       -0.39949749, -0.39447236, -0.38944724, -0.38442211, -0.37939698,
       -0.37437186, -0.36934673, -0.36432161, -0.35929648, -0.35427136,
       -0.34924623, -0.34422111, -0.33919598, -0.33417085, -0.32914573,
       -0.3241206 , -0.31909548, -0.31407035, -0.30904523, -0.3040201 ,
       -0.29899497, -0.29396985, -0.28894472, -0.2839196 , -0.27889447,
       -0.27386935, -0.26884422, -0.2638191 , -0.25879397, -0.25376884,
       -0.24874372, -0.24371859, -0.23869347, -0.23366834, -0.22864322,
       -0.22361809, -0.21859296, -0.21356784, -0.20854271, -0.20351759,
       -0.19849246, -0.19346734, -0.18844221, -0.18341709, -0.17839196,
       -0.17336683, -0.16834171, -0.16331658, -0.15829146, -0.15326633,
       -0.14824121, -0.14321608, -0.13819095, -0.13316583, -0.1281407 ,
       -0.12311558, -0.11809045, -0.11306533, -0.1080402 , -0.10301508,
       -0.09798995, -0.09296482, -0.0879397 , -0.08291457, -0.07788945,
       -0.07286432, -0.0678392 , -0.06281407, -0.05778894, -0.05276382,
       -0.04773869, -0.04271357, -0.03768844, -0.03266332, -0.02763819,
       -0.02261307, -0.01758794, -0.01256281, -0.00753769, -0.00251256,
        0.00251256,  0.00753769,  0.01256281,  0.01758794,  0.02261307,
        0.02763819,  0.03266332,  0.03768844,  0.04271357,  0.04773869,
        0.05276382,  0.05778894,  0.06281407,  0.0678392 ,  0.07286432,
        0.07788945,  0.08291457,  0.0879397 ,  0.09296482,  0.09798995,
        0.10301508,  0.1080402 ,  0.11306533,  0.11809045,  0.12311558,
        0.1281407 ,  0.13316583,  0.13819095,  0.14321608,  0.14824121,
        0.15326633,  0.15829146,  0.16331658,  0.16834171,  0.17336683,
        0.17839196,  0.18341709,  0.18844221,  0.19346734,  0.19849246,
        0.20351759,  0.20854271,  0.21356784,  0.21859296,  0.22361809,
        0.22864322,  0.23366834,  0.23869347,  0.24371859,  0.24874372,
        0.25376884,  0.25879397,  0.2638191 ,  0.26884422,  0.27386935,
        0.27889447,  0.2839196 ,  0.28894472,  0.29396985,  0.29899497,
        0.3040201 ,  0.30904523,  0.31407035,  0.31909548,  0.3241206 ,
        0.32914573,  0.33417085,  0.33919598,  0.34422111,  0.34924623,
        0.35427136,  0.35929648,  0.36432161,  0.36934673,  0.37437186,
        0.37939698,  0.38442211,  0.38944724,  0.39447236,  0.39949749,
        0.40452261,  0.40954774,  0.41457286,  0.41959799,  0.42462312,
        0.42964824,  0.43467337,  0.43969849,  0.44472362,  0.44974874,
        0.45477387,  0.45979899,  0.46482412,  0.46984925,  0.47487437,
        0.4798995 ,  0.48492462,  0.48994975,  0.49497487,  0.5       ])

将一维数据升成2维数据

np.newaxis

np.newaxis 的功能是增加新的维度，但是要注意 np.newaxis 放的位置不同，产生的矩阵形状也不同。

通常按照如下规则：

np.newaxis 放在哪个位置，就会给哪个位置增加维度

x[:, np.newaxis] ，放在后面，会给列上增加维度
x[np.newaxis, :] ，放在前面，会给行上增加维度

用途： 通常用它将一维的数据转换成一个矩阵，这样就可以与其他矩阵进行相乘。

x_data = x_data[:,np.newaxis]

x_data

array([[-0.5       ],
       [-0.49497487],
       [-0.48994975],
       [-0.48492462],
       [-0.4798995 ],
       [-0.47487437],
       [-0.46984925],
       [-0.46482412],
       [-0.45979899],
       [-0.45477387],
       [-0.44974874],
       [-0.44472362],
       [-0.43969849],
       [-0.43467337],
       [-0.42964824],
       [-0.42462312],
       [-0.41959799],
       [-0.41457286],
       [-0.40954774],
       [-0.40452261],
       [-0.39949749],
       [-0.39447236],
       [-0.38944724],
       [-0.38442211],
       [-0.37939698],
       [-0.37437186],
       [-0.36934673],
       [-0.36432161],
       [-0.35929648],
       [-0.35427136],
       [-0.34924623],
       [-0.34422111],
       [-0.33919598],
       [-0.33417085],
       [-0.32914573],
       [-0.3241206 ],
       [-0.31909548],
       [-0.31407035],
       [-0.30904523],
       [-0.3040201 ],
       [-0.29899497],
       [-0.29396985],
       [-0.28894472],
       [-0.2839196 ],
       [-0.27889447],
       [-0.27386935],
       [-0.26884422],
       [-0.2638191 ],
       [-0.25879397],
       [-0.25376884],
       [-0.24874372],
       [-0.24371859],
       [-0.23869347],
       [-0.23366834],
       [-0.22864322],
       [-0.22361809],
       [-0.21859296],
       [-0.21356784],
       [-0.20854271],
       [-0.20351759],
       [-0.19849246],
       [-0.19346734],
       [-0.18844221],
       [-0.18341709],
       [-0.17839196],
       [-0.17336683],
       [-0.16834171],
       [-0.16331658],
       [-0.15829146],
       [-0.15326633],
       [-0.14824121],
       [-0.14321608],
       [-0.13819095],
       [-0.13316583],
       [-0.1281407 ],
       [-0.12311558],
       [-0.11809045],
       [-0.11306533],
       [-0.1080402 ],
       [-0.10301508],
       [-0.09798995],
       [-0.09296482],
       [-0.0879397 ],
       [-0.08291457],
       [-0.07788945],
       [-0.07286432],
       [-0.0678392 ],
       [-0.06281407],
       [-0.05778894],
       [-0.05276382],
       [-0.04773869],
       [-0.04271357],
       [-0.03768844],
       [-0.03266332],
       [-0.02763819],
       [-0.02261307],
       [-0.01758794],
       [-0.01256281],
       [-0.00753769],
       [-0.00251256],
       [ 0.00251256],
       [ 0.00753769],
       [ 0.01256281],
       [ 0.01758794],
       [ 0.02261307],
       [ 0.02763819],
       [ 0.03266332],
       [ 0.03768844],
       [ 0.04271357],
       [ 0.04773869],
       [ 0.05276382],
       [ 0.05778894],
       [ 0.06281407],
       [ 0.0678392 ],
       [ 0.07286432],
       [ 0.07788945],
       [ 0.08291457],
       [ 0.0879397 ],
       [ 0.09296482],
       [ 0.09798995],
       [ 0.10301508],
       [ 0.1080402 ],
       [ 0.11306533],
       [ 0.11809045],
       [ 0.12311558],
       [ 0.1281407 ],
       [ 0.13316583],
       [ 0.13819095],
       [ 0.14321608],
       [ 0.14824121],
       [ 0.15326633],
       [ 0.15829146],
       [ 0.16331658],
       [ 0.16834171],
       [ 0.17336683],
       [ 0.17839196],
       [ 0.18341709],
       [ 0.18844221],
       [ 0.19346734],
       [ 0.19849246],
       [ 0.20351759],
       [ 0.20854271],
       [ 0.21356784],
       [ 0.21859296],
       [ 0.22361809],
       [ 0.22864322],
       [ 0.23366834],
       [ 0.23869347],
       [ 0.24371859],
       [ 0.24874372],
       [ 0.25376884],
       [ 0.25879397],
       [ 0.2638191 ],
       [ 0.26884422],
       [ 0.27386935],
       [ 0.27889447],
       [ 0.2839196 ],
       [ 0.28894472],
       [ 0.29396985],
       [ 0.29899497],
       [ 0.3040201 ],
       [ 0.30904523],
       [ 0.31407035],
       [ 0.31909548],
       [ 0.3241206 ],
       [ 0.32914573],
       [ 0.33417085],
       [ 0.33919598],
       [ 0.34422111],
       [ 0.34924623],
       [ 0.35427136],
       [ 0.35929648],
       [ 0.36432161],
       [ 0.36934673],
       [ 0.37437186],
       [ 0.37939698],
       [ 0.38442211],
       [ 0.38944724],
       [ 0.39447236],
       [ 0.39949749],
       [ 0.40452261],
       [ 0.40954774],
       [ 0.41457286],
       [ 0.41959799],
       [ 0.42462312],
       [ 0.42964824],
       [ 0.43467337],
       [ 0.43969849],
       [ 0.44472362],
       [ 0.44974874],
       [ 0.45477387],
       [ 0.45979899],
       [ 0.46482412],
       [ 0.46984925],
       [ 0.47487437],
       [ 0.4798995 ],
       [ 0.48492462],
       [ 0.48994975],
       [ 0.49497487],
       [ 0.5       ]])

生成噪声数据与x_data的shape一致

noise 为均值0，方差为0.02的高斯分布

知识点：

np.random.normal(mean, stddev, shape)用于生成一个随机正态分布（高斯分布）的数据，正态分布的均值为mean，方差为stddev，数据的维度为shape。

tf.random_normal([1,10])在tensorflow中形成[1,10 ]大小的张量，数据服从高斯分布。默认的均值为0，方差为0.1。

noise = np.random.normal(0,0.02,x_data.shape)

noise

array([[-2.02515881e-02],
       [-3.98700785e-04],
       [ 1.61803122e-02],
       [ 6.51774294e-03],
       [-1.52915571e-02],
       [ 5.28524003e-04],
       [-1.35583316e-02],
       [-1.16639357e-02],
       [ 1.12795088e-02],
       [ 2.17958202e-02],
       [-5.10629033e-03],
       [-3.42689421e-02],
       [-2.73630598e-02],
       [-3.73493961e-03],
       [ 1.05498899e-04],
       [-1.84174916e-02],
       [-2.34951662e-02],
       [-1.97783659e-02],
       [-1.90812955e-02],
       [ 9.58075145e-03],
       [-3.07560733e-02],
       [-3.16563044e-02],
       [ 2.01417254e-02],
       [ 1.37672276e-02],
       [ 3.99399233e-02],
       [ 5.08037999e-02],
       [ 3.89761871e-02],
       [-1.42584732e-02],
       [-9.00674559e-03],
       [ 8.72212103e-03],
       [ 1.05172674e-03],
       [ 3.28943871e-03],
       [ 3.17770175e-02],
       [ 1.07188843e-02],
       [-2.18477136e-02],
       [ 2.30739671e-02],
       [ 1.42992443e-03],
       [ 5.64206720e-03],
       [ 1.12339933e-02],
       [ 1.17920375e-02],
       [-1.76954406e-02],
       [ 7.75792215e-04],
       [-5.00778192e-03],
       [ 1.14254882e-02],
       [ 2.58287480e-02],
       [ 2.07157841e-02],
       [-2.34783762e-02],
       [ 2.14233757e-02],
       [ 3.07079511e-02],
       [ 9.31029018e-03],
       [-3.16327294e-02],
       [ 8.21574331e-03],
       [-2.40266443e-02],
       [-7.39963396e-04],
       [-1.54603538e-02],
       [ 6.03391882e-03],
       [ 2.00442330e-02],
       [ 1.00569368e-02],
       [ 9.82289659e-03],
       [-5.85642119e-04],
       [ 2.40038008e-03],
       [ 2.69370432e-02],
       [-1.37818764e-02],
       [ 2.47578426e-02],
       [-4.45396331e-03],
       [ 2.17608302e-03],
       [-2.11030685e-02],
       [-1.95243842e-03],
       [-4.52981725e-03],
       [ 9.29389670e-03],
       [-6.48160747e-03],
       [-2.90058751e-02],
       [ 2.45138067e-02],
       [-6.68889030e-03],
       [-1.07238813e-02],
       [ 1.54892926e-02],
       [-3.46983956e-02],
       [-7.58514693e-03],
       [ 4.10520294e-03],
       [ 2.38037479e-02],
       [ 2.42883804e-02],
       [-7.03287474e-04],
       [-1.27985822e-02],
       [-2.03246987e-02],
       [ 6.95005345e-03],
       [-1.49340876e-02],
       [ 3.02910049e-02],
       [ 1.75900519e-02],
       [-8.09429871e-03],
       [ 3.72621254e-02],
       [-7.74786634e-03],
       [-7.57327275e-03],
       [ 2.16679673e-03],
       [ 2.78236634e-02],
       [ 1.91890482e-02],
       [-1.67659468e-02],
       [-1.58632175e-02],
       [-1.28388472e-03],
       [-3.53630722e-03],
       [-2.01801723e-03],
       [ 1.53658041e-02],
       [ 1.24663789e-02],
       [-1.52842040e-02],
       [ 6.77654868e-05],
       [-2.90278049e-02],
       [ 1.78198188e-02],
       [ 6.84986231e-03],
       [ 1.26168449e-02],
       [ 2.75176592e-02],
       [ 4.65492235e-03],
       [ 7.11446594e-03],
       [ 2.99692534e-02],
       [ 1.46054853e-02],
       [-6.30385578e-03],
       [-2.08958882e-02],
       [ 1.47020914e-02],
       [-2.02600867e-02],
       [ 3.96328661e-03],
       [-3.15813684e-02],
       [ 3.09509993e-03],
       [ 2.00029087e-03],
       [-2.88060987e-02],
       [ 2.97860739e-02],
       [-1.64964120e-02],
       [-3.56814269e-02],
       [ 2.97813064e-02],
       [ 1.31154156e-02],
       [-6.69908095e-03],
       [ 3.09366668e-02],
       [ 2.39595221e-02],
       [ 2.47953496e-02],
       [ 2.20327257e-03],
       [-6.00919408e-04],
       [ 6.85003266e-03],
       [-2.41446522e-02],
       [-1.43629595e-02],
       [ 3.44114035e-02],
       [ 1.76974462e-02],
       [-1.26406158e-03],
       [-2.01851128e-02],
       [ 4.66711699e-03],
       [ 3.37930376e-03],
       [-1.56299057e-02],
       [-7.31714719e-03],
       [-2.90333968e-02],
       [ 1.32564815e-02],
       [ 8.62453846e-03],
       [ 5.87748232e-03],
       [-1.30342070e-02],
       [ 1.66746834e-02],
       [ 8.47028139e-03],
       [-2.49514994e-03],
       [-5.91114136e-02],
       [ 7.52942835e-03],
       [-3.49852240e-02],
       [-1.20737562e-02],
       [ 2.39032205e-02],
       [ 1.82399971e-02],
       [-3.42108155e-02],
       [ 1.96295786e-02],
       [-1.00512367e-02],
       [ 2.56140687e-02],
       [ 5.69259106e-02],
       [-2.19836979e-02],
       [-1.95835444e-03],
       [ 1.61181804e-02],
       [ 2.03444155e-02],
       [-1.96351898e-02],
       [ 2.24982506e-02],
       [ 4.83362322e-03],
       [-2.75322699e-02],
       [-8.67047701e-03],
       [ 2.02351784e-02],
       [ 1.72292532e-02],
       [-3.45015698e-03],
       [-2.92746604e-03],
       [-3.93184989e-02],
       [ 8.00309461e-03],
       [-2.68324409e-03],
       [ 1.74606570e-02],
       [-2.78318598e-03],
       [ 1.14454183e-03],
       [ 5.01713826e-02],
       [-2.15924094e-02],
       [ 7.83265419e-03],
       [-9.66560920e-03],
       [-2.54746325e-02],
       [-2.28332672e-02],
       [-1.16610213e-03],
       [ 4.14485952e-02],
       [ 3.74661756e-03],
       [-8.70117611e-03],
       [ 7.60622823e-03],
       [ 1.24471458e-03],
       [-2.40383989e-02],
       [ 2.60671683e-03],
       [ 1.42489497e-03],
       [-2.69978331e-02],
       [ 2.65559720e-03],
       [ 3.04472983e-02]])

构建数据y_data

$ y_data = x_data^{2} + noise $

y_data = np.square(x_data) + noise

y_data

array([[ 0.22974841],
       [ 0.24460143],
       [ 0.25623107],
       [ 0.24166963],
       [ 0.21501197],
       [ 0.22603419],
       [ 0.20719998],
       [ 0.20439753],
       [ 0.22269462],
       [ 0.22861509],
       [ 0.19716764],
       [ 0.16351015],
       [ 0.1659717 ],
       [ 0.185206  ],
       [ 0.18470311],
       [ 0.1618873 ],
       [ 0.15256731],
       [ 0.15209229],
       [ 0.14864805],
       [ 0.1732193 ],
       [ 0.12884217],
       [ 0.12395214],
       [ 0.17181088],
       [ 0.16154759],
       [ 0.183882  ],
       [ 0.19095809],
       [ 0.1753932 ],
       [ 0.11847176],
       [ 0.12008722],
       [ 0.13423032],
       [ 0.12302466],
       [ 0.12177761],
       [ 0.14683093],
       [ 0.12238904],
       [ 0.0864892 ],
       [ 0.12812813],
       [ 0.10325185],
       [ 0.10428225],
       [ 0.10674295],
       [ 0.10422026],
       [ 0.07170255],
       [ 0.08719406],
       [ 0.07848127],
       [ 0.09203583],
       [ 0.10361087],
       [ 0.0957202 ],
       [ 0.04879884],
       [ 0.09102389],
       [ 0.09768227],
       [ 0.07370892],
       [ 0.03024071],
       [ 0.0676145 ],
       [ 0.03294793],
       [ 0.05386093],
       [ 0.03681737],
       [ 0.05603897],
       [ 0.06782712],
       [ 0.05566816],
       [ 0.05331296],
       [ 0.04083377],
       [ 0.04179964],
       [ 0.06436665],
       [ 0.02172859],
       [ 0.05839967],
       [ 0.02736973],
       [ 0.03223214],
       [ 0.00723586],
       [ 0.02471987],
       [ 0.02052637],
       [ 0.03278447],
       [ 0.01549385],
       [-0.00849503],
       [ 0.04361055],
       [ 0.01104425],
       [ 0.00569616],
       [ 0.03064674],
       [-0.02075304],
       [ 0.00519862],
       [ 0.01577789],
       [ 0.03441585],
       [ 0.03389041],
       [ 0.00793917],
       [-0.00506519],
       [-0.01344987],
       [ 0.01301682],
       [-0.00962488],
       [ 0.03489316],
       [ 0.02153566],
       [-0.00475474],
       [ 0.04004615],
       [-0.00546888],
       [-0.00574882],
       [ 0.00358722],
       [ 0.02889056],
       [ 0.01995292],
       [-0.0162546 ],
       [-0.01555388],
       [-0.00112606],
       [-0.00347949],
       [-0.0020117 ],
       [ 0.01537212],
       [ 0.0125232 ],
       [-0.01512638],
       [ 0.0003771 ],
       [-0.02851645],
       [ 0.01858369],
       [ 0.00791675],
       [ 0.01403726],
       [ 0.02934211],
       [ 0.00693391],
       [ 0.00989849],
       [ 0.03330882],
       [ 0.01855109],
       [-0.0017017 ],
       [-0.01558668],
       [ 0.02076886],
       [-0.01338526],
       [ 0.01169668],
       [-0.02293891],
       [ 0.01269713],
       [ 0.0126124 ],
       [-0.01713341],
       [ 0.04256984],
       [-0.00255106],
       [-0.02052398],
       [ 0.04620135],
       [ 0.03084855],
       [ 0.01239766],
       [ 0.05144751],
       [ 0.04593498],
       [ 0.04828592],
       [ 0.02725946],
       [ 0.02607139],
       [ 0.03518896],
       [ 0.00591141],
       [ 0.01746073],
       [ 0.06805323],
       [ 0.05320791],
       [ 0.03616555],
       [ 0.01921414],
       [ 0.04608653],
       [ 0.04686937],
       [ 0.02998132],
       [ 0.04046574],
       [ 0.02097165],
       [ 0.0655342 ],
       [ 0.06322543],
       [ 0.06285205],
       [ 0.04636455],
       [ 0.07854812],
       [ 0.07286891],
       [ 0.06447917],
       [ 0.0104891 ],
       [ 0.07980664],
       [ 0.0400192 ],
       [ 0.06570837],
       [ 0.10451356],
       [ 0.10172905],
       [ 0.05220746],
       [ 0.10902757],
       [ 0.08237698],
       [ 0.12112302],
       [ 0.1555661 ],
       [ 0.07983823],
       [ 0.10309581],
       [ 0.12445509],
       [ 0.13201458],
       [ 0.09541872],
       [ 0.14098642],
       [ 0.12680655],
       [ 0.09797592],
       [ 0.12042349],
       [ 0.15296541],
       [ 0.15364626],
       [ 0.13670413],
       [ 0.14101461],
       [ 0.10846186],
       [ 0.15967224],
       [ 0.1529252 ],
       [ 0.1770589 ],
       [ 0.16085536],
       [ 0.16887389],
       [ 0.22204204],
       [ 0.15447006],
       [ 0.18813744],
       [ 0.174932  ],
       [ 0.1634663 ],
       [ 0.1705015 ],
       [ 0.19661299],
       [ 0.24372253],
       [ 0.21056589],
       [ 0.20271394],
       [ 0.22366769],
       [ 0.22200303],
       [ 0.20146727],
       [ 0.23291024],
       [ 0.23657679],
       [ 0.21305292],
       [ 0.24765572],
       [ 0.2804473 ]])

定义placeholder用于接收训练的数据

可以认为是输入层

x = tf.placeholder(tf.float32, [None, 1], name="x_input")
y = tf.placeholder(tf.float32, [None, 1], name="y_input")

x

<tf.Tensor 'x_input:0' shape=(?, 1) dtype=float32>

y

<tf.Tensor 'y_input:0' shape=(?, 1) dtype=float32>

定义隐藏层

即：
$$a_1=w_1\ast x+b_1$$
$$ou{t}_1=activation\left(a_1\right)$$

tf.random_normal()函数用于从“服从指定正态分布的序列”中随机取出指定个数的值。

tf.random_normal(shape, mean=0.0, stddev=1.0, dtype=tf.float32, seed=None, name=None)

shape: 输出张量的形状，必选
mean: 正态分布的均值，默认为0
stddev: 正态分布的标准差，默认为1.0
dtype: 输出的类型，默认为tf.float32
seed: 随机数种子，是一个整数，当设置之后，每次生成的随机数都一样
name: 操作的名称

tf.zeros(
shape,
dtype=tf.float32,
name=None
)

其他重要结论：
变量w1声明之后并没有被赋值，需要在Session中调用run(tf.global_variables_initializer())方法初始化之后才会被具体赋值。

tf中张量与常规向量不同的是执行"print w1"输出的是w1的形状和数据类型等属性信息，获取w1的值需要调用sess.run(w1)方法。

tf.nn.tanh()是激活函数，图形类似于sigmod，但是tanh比sigmod更陡。
————————————————
原文链接

W_1 = tf.Variable(tf.random_normal([1,10]))
b_1 = tf.Variable(tf.zeros([1,10]))
a_1 = tf.matmul(x, W_1) + b_1
out_1 = tf.nn.tanh(a_1)

W_1

<tf.Variable 'Variable:0' shape=(1, 10) dtype=float32_ref>

b_1

<tf.Variable 'Variable_1:0' shape=(1, 10) dtype=float32_ref>

a_1

<tf.Tensor 'add:0' shape=(?, 10) dtype=float32>

out_1

<tf.Tensor 'Tanh:0' shape=(?, 10) dtype=float32>

定义输出层

即输出变为维度为1的数据
$$a_2=ou{t}_1\ast w_2+b_2$$
$$ou{t}_2=activation\left(a_2\right)$$

W_2 = tf.Variable(tf.random_normal([10,1]))
b_2 = tf.Variable(tf.zeros([1,1]))
a_2 = tf.matmul(out_1,W_2) + b_2
out_2 = tf.nn.tanh(a_2)

W_2

<tf.Variable 'Variable_2:0' shape=(10, 1) dtype=float32_ref>

b_2

<tf.Variable 'Variable_3:0' shape=(1, 1) dtype=float32_ref>

a_2

<tf.Tensor 'add_1:0' shape=(?, 1) dtype=float32>

out_2

<tf.Tensor 'Tanh_1:0' shape=(?, 1) dtype=float32>

定义损失函数

loss = tf.reduce_mean(tf.square(out_2 - y_data))

loss

<tf.Tensor 'Mean:0' shape=() dtype=float32>

train_step = tf.train.GradientDescentOptimizer(0.1).minimize(loss)

train_step

<tf.Operation 'GradientDescent' type=NoOp>

初始化变量

init = tf.global_variables_initializer()

init

<tf.Operation 'init' type=NoOp>

训练

with tf.Session() as sess:
    sess.run(init)
    for epc in range(10000):
        sess.run([loss, train_step],{x:x_data,y:y_data})
        if epc % 10 == 0:
            # loss_value = sess.run([loss])
            loss_value = sess.run([loss],{x:x_data,y:y_data})
            
            print(epc, loss_value)
    prediction_value = sess.run(out_2, feed_dict={x:x_data})

0 [0.088102065]
10 [0.008092111]
20 [0.0068610567]
30 [0.0059195864]
40 [0.005191134]
50 [0.004624322]
60 [0.004180253]
70 [0.0038295502]
80 [0.0035500634]
90 [0.0033250647]
100 [0.0031419084]
110 [0.0029910123]
120 [0.0028650912]
130 [0.0027586007]
140 [0.0026672974]
150 [0.00258793]
160 [0.002517999]
170 [0.0024555754]
180 [0.0023991656]
190 [0.0023476128]
200 [0.00230001]
210 [0.0022556535]
220 [0.002213992]
230 [0.0021745914]
240 [0.0021371064]
250 [0.0021012647]
260 [0.0020668525]
270 [0.002033692]
280 [0.0020016485]
290 [0.0019706015]
300 [0.0019404693]
310 [0.0019111665]
320 [0.0018826395]
330 [0.0018548326]
340 [0.0018277002]
350 [0.0018012094]
360 [0.0017753274]
370 [0.0017500247]
380 [0.0017252808]
390 [0.0017010701]
400 [0.0016773775]
410 [0.0016541854]
420 [0.0016314754]
430 [0.001609237]
440 [0.0015874541]
450 [0.0015661144]
460 [0.0015452087]
470 [0.0015247224]
480 [0.0015046478]
490 [0.0014849738]
500 [0.001465691]
510 [0.0014467903]
520 [0.0014282635]
530 [0.0014101005]
540 [0.0013922951]
550 [0.0013748398]
560 [0.0013577223]
570 [0.0013409385]
580 [0.001324485]
590 [0.0013083491]
600 [0.0012925273]
610 [0.0012770089]
620 [0.0012617889]
630 [0.0012468616]
640 [0.0012322213]
650 [0.0012178608]
660 [0.001203776]
670 [0.0011899606]
680 [0.001176408]
690 [0.0011631139]
700 [0.0011500707]
710 [0.0011372764]
720 [0.0011247236]
730 [0.0011124085]
740 [0.0011003258]
750 [0.0010884696]
760 [0.0010768368]
770 [0.0010654222]
780 [0.0010542229]
790 [0.0010432326]
800 [0.0010324479]
810 [0.001021865]
820 [0.0010114787]
830 [0.0010012871]
840 [0.0009912843]
850 [0.0009814667]
860 [0.0009718312]
870 [0.0009623742]
880 [0.0009530929]
890 [0.00094398216]
900 [0.00093503983]
910 [0.0009262621]
920 [0.0009176459]
930 [0.0009091884]
940 [0.0009008854]
950 [0.00089273485]
960 [0.00088473404]
970 [0.00087687845]
980 [0.00086916686]
990 [0.00086159585]
1000 [0.00085416285]
1010 [0.00084686594]
1020 [0.0008397007]
1030 [0.00083266536]
1040 [0.00082575827]
1050 [0.000818976]
1060 [0.000812317]
1070 [0.0008057776]
1080 [0.00079935655]
1090 [0.0007930518]
1100 [0.0007868596]
1110 [0.0007807794]
1120 [0.000774809]
1130 [0.000768946]
1140 [0.0007631874]
1150 [0.00075753254]
1160 [0.0007519795]
1170 [0.0007465253]
1180 [0.00074116816]
1190 [0.00073590665]
1200 [0.0007307394]
1210 [0.0007256637]
1220 [0.0007206787]
1230 [0.00071578263]
1240 [0.0007109733]
1250 [0.0007062496]
1260 [0.0007016087]
1270 [0.00069705123]
1280 [0.0006925741]
1290 [0.000688176]
1300 [0.00068385573]
1310 [0.00067961146]
1320 [0.0006754424]
1330 [0.0006713466]
1340 [0.0006673233]
1350 [0.0006633699]
1360 [0.0006594866]
1370 [0.00065567123]
1380 [0.00065192394]
1390 [0.00064824015]
1400 [0.00064462255]
1410 [0.00064106775]
1420 [0.00063757517]
1430 [0.0006341433]
1440 [0.0006307714]
1450 [0.0006274588]
1460 [0.0006242032]
1470 [0.0006210049]
1480 [0.00061786204]
1490 [0.00061477325]
1500 [0.00061173865]
1510 [0.00060875644]
1520 [0.0006058264]
1530 [0.0006029463]
1540 [0.0006001171]
1550 [0.00059733563]
1560 [0.0005946035]
1570 [0.000591917]
1580 [0.0005892783]
1590 [0.0005866838]
1600 [0.0005841345]
1610 [0.00058162963]
1620 [0.0005791669]
1630 [0.0005767466]
1640 [0.0005743682]
1650 [0.0005720304]
1660 [0.0005697327]
1670 [0.00056747434]
1680 [0.0005652542]
1690 [0.0005630722]
1700 [0.0005609275]
1710 [0.0005588197]
1720 [0.00055674737]
1730 [0.00055471034]
1740 [0.0005527083]
1750 [0.00055074017]
1760 [0.00054880534]
1770 [0.0005469036]
1780 [0.000545034]
1790 [0.0005431957]
1800 [0.00054138916]
1810 [0.0005396124]
1820 [0.00053786614]
1830 [0.000536149]
1840 [0.00053446123]
1850 [0.0005328015]
1860 [0.00053117017]
1870 [0.0005295665]
1880 [0.00052798924]
1890 [0.00052643823]
1900 [0.00052491395]
1910 [0.000523415]
1920 [0.0005219409]
1930 [0.0005204917]
1940 [0.00051906635]
1950 [0.00051766547]
1960 [0.00051628763]
1970 [0.00051493256]
1980 [0.0005136007]
1990 [0.00051229045]
2000 [0.0005110018]
2010 [0.00050973526]
2020 [0.000508489]
2030 [0.00050726376]
2040 [0.00050605944]
2050 [0.0005048739]
2060 [0.0005037088]
2070 [0.0005025626]
2080 [0.00050143566]
2090 [0.00050032703]
2100 [0.00049923634]
2110 [0.000498164]
2120 [0.00049710914]
2130 [0.00049607205]
2140 [0.00049505115]
2150 [0.0004940473]
2160 [0.0004930606]
2170 [0.00049208937]
2180 [0.0004911343]
2190 [0.0004901946]
2200 [0.00048927026]
2210 [0.0004883614]
2220 [0.00048746713]
2230 [0.00048658706]
2240 [0.00048572134]
2250 [0.00048486955]
2260 [0.00048403215]
2270 [0.00048320784]
2280 [0.00048239712]
2290 [0.00048159956]
2300 [0.00048081472]
2310 [0.00048004295]
2320 [0.00047928267]
2330 [0.0004785353]
2340 [0.0004777997]
2350 [0.0004770761]
2360 [0.00047636408]
2370 [0.0004756636]
2380 [0.00047497393]
2390 [0.00047429555]
2400 [0.00047362794]
2410 [0.00047297098]
2420 [0.0004723244]
2430 [0.00047168828]
2440 [0.00047106214]
2450 [0.00047044607]
2460 [0.0004698395]
2470 [0.00046924278]
2480 [0.00046865566]
2490 [0.0004680775]
2500 [0.0004675086]
2510 [0.00046694867]
2520 [0.0004663973]
2530 [0.00046585477]
2540 [0.0004653211]
2550 [0.00046479513]
2560 [0.0004642776]
2570 [0.00046376832]
2580 [0.0004632672]
2590 [0.0004627734]
2600 [0.00046228728]
2610 [0.00046180864]
2620 [0.00046133844]
2630 [0.00046087438]
2640 [0.00046041797]
2650 [0.00045996852]
2660 [0.00045952582]
2670 [0.0004590903]
2680 [0.00045866147]
2690 [0.00045823882]
2700 [0.00045782304]
2710 [0.0004574138]
2720 [0.00045701078]
2730 [0.00045661323]
2740 [0.00045622193]
2750 [0.00045583711]
2760 [0.00045545795]
2770 [0.00045508437]
2780 [0.00045471615]
2790 [0.0004543542]
2800 [0.00045399703]
2810 [0.00045364545]
2820 [0.00045329944]
2830 [0.00045295843]
2840 [0.00045262274]
2850 [0.00045229198]
2860 [0.000451966]
2870 [0.00045164465]
2880 [0.00045132867]
2890 [0.00045101703]
2900 [0.0004507101]
2910 [0.00045040727]
2920 [0.00045010904]
2930 [0.0004498162]
2940 [0.0004495264]
2950 [0.0004492414]
2960 [0.00044896072]
2970 [0.00044868438]
2980 [0.00044841124]
2990 [0.0004481426]
3000 [0.00044787786]
3010 [0.00044761656]
3020 [0.00044735923]
3030 [0.00044710565]
3040 [0.00044685593]
3050 [0.00044660963]
3060 [0.0004463672]
3070 [0.0004461277]
3080 [0.00044589108]
3090 [0.00044565886]
3100 [0.00044543014]
3110 [0.0004452036]
3120 [0.00044498112]
3130 [0.0004447614]
3140 [0.00044454483]
3150 [0.00044433097]
3160 [0.000444121]
3170 [0.000443913]
3180 [0.0004437085]
3190 [0.00044350655]
3200 [0.00044330745]
3210 [0.0004431111]
3220 [0.00044291755]
3230 [0.0004427265]
3240 [0.00044253803]
3250 [0.00044235185]
3260 [0.00044216882]
3270 [0.0004419884]
3280 [0.00044180977]
3290 [0.00044163392]
3300 [0.00044146052]
3310 [0.000441289]
3320 [0.00044112]
3330 [0.00044095368]
3340 [0.0004407888]
3350 [0.00044062675]
3360 [0.00044046674]
3370 [0.0004403086]
3380 [0.00044015245]
3390 [0.00043999852]
3400 [0.0004398466]
3410 [0.00043969668]
3420 [0.00043954843]
3430 [0.00043940215]
3440 [0.00043925812]
3450 [0.00043911568]
3460 [0.00043897502]
3470 [0.00043883678]
3480 [0.00043869932]
3490 [0.00043856387]
3500 [0.0004384306]
3510 [0.00043829836]
3520 [0.00043816835]
3530 [0.00043803937]
3540 [0.00043791244]
3550 [0.00043778692]
3560 [0.00043766276]
3570 [0.00043754003]
3580 [0.00043741948]
3590 [0.00043730022]
3600 [0.00043718188]
3610 [0.00043706544]
3620 [0.00043695018]
3630 [0.00043683636]
3640 [0.00043672405]
3650 [0.00043661267]
3660 [0.00043650277]
3670 [0.0004363945]
3680 [0.0004362873]
3690 [0.00043618082]
3700 [0.00043607614]
3710 [0.00043597282]
3720 [0.00043587005]
3730 [0.0004357688]
3740 [0.00043566845]
3750 [0.00043556967]
3760 [0.0004354723]
3770 [0.0004353749]
3780 [0.00043527916]
3790 [0.00043518477]
3800 [0.00043509132]
3810 [0.0004349982]
3820 [0.0004349073]
3830 [0.00043481632]
3840 [0.000434727]
3850 [0.00043463818]
3860 [0.00043455034]
3870 [0.000434464]
3880 [0.0004343779]
3890 [0.00043429286]
3900 [0.0004342092]
3910 [0.00043412566]
3920 [0.00043404356]
3930 [0.00043396198]
3940 [0.00043388127]
3950 [0.00043380202]
3960 [0.0004337225]
3970 [0.00043364457]
3980 [0.00043356713]
3990 [0.00043349044]
4000 [0.00043341465]
4010 [0.00043333956]
4020 [0.0004332652]
4030 [0.0004331911]
4040 [0.00043311817]
4050 [0.0004330458]
4060 [0.00043297416]
4070 [0.00043290333]
4080 [0.00043283254]
4090 [0.00043276348]
4100 [0.00043269433]
4110 [0.00043262585]
4120 [0.00043255818]
4130 [0.00043249095]
4140 [0.00043242468]
4150 [0.0004323585]
4160 [0.00043229398]
4170 [0.0004322289]
4180 [0.00043216452]
4190 [0.00043210105]
4200 [0.00043203763]
4210 [0.00043197558]
4220 [0.00043191345]
4230 [0.00043185183]
4240 [0.00043179086]
4250 [0.00043173053]
4260 [0.0004316707]
4270 [0.00043161103]
4280 [0.00043155163]
4290 [0.00043149322]
4300 [0.00043143533]
4310 [0.00043137762]
4320 [0.00043132054]
4330 [0.0004312641]
4340 [0.0004312077]
4350 [0.00043115148]
4360 [0.00043109656]
4370 [0.0004310414]
4380 [0.0004309873]
4390 [0.0004309325]
4400 [0.0004308792]
4410 [0.0004308261]
4420 [0.00043077313]
4430 [0.0004307202]
4440 [0.00043066777]
4450 [0.00043061597]
4460 [0.00043056524]
4470 [0.0004305138]
4480 [0.0004304629]
4490 [0.00043041285]
4500 [0.0004303623]
4510 [0.00043031285]
4520 [0.00043026317]
4530 [0.0004302138]
4540 [0.00043016527]
4550 [0.00043011724]
4560 [0.00043006917]
4570 [0.00043002106]
4580 [0.0004299736]
4590 [0.00042992647]
4600 [0.0004298798]
4610 [0.00042983308]
4620 [0.00042978636]
4630 [0.00042974055]
4640 [0.0004296945]
4650 [0.0004296494]
4660 [0.00042960397]
4670 [0.00042955892]
4680 [0.00042951456]
4690 [0.0004294701]
4700 [0.0004294261]
4710 [0.00042938173]
4720 [0.0004293382]
4730 [0.00042929448]
4740 [0.00042925126]
4750 [0.0004292084]
4760 [0.00042916567]
4770 [0.00042912338]
4780 [0.00042908083]
4790 [0.00042903863]
4800 [0.00042899646]
4810 [0.00042895557]
4820 [0.00042891383]
4830 [0.00042887262]
4840 [0.0004288312]
4850 [0.00042879037]
4860 [0.00042874983]
4870 [0.00042870964]
4880 [0.000428669]
4890 [0.00042862876]
4900 [0.00042858906]
4910 [0.0004285492]
4920 [0.00042850987]
4930 [0.00042847026]
4940 [0.00042843155]
4950 [0.00042839212]
4960 [0.00042835347]
4970 [0.00042831487]
4980 [0.00042827625]
4990 [0.0004282383]
5000 [0.00042819968]
5010 [0.0004281617]
5020 [0.00042812416]
5030 [0.00042808673]
5040 [0.00042804904]
5050 [0.0004280118]
5060 [0.00042797413]
5070 [0.00042793722]
5080 [0.00042790014]
5090 [0.00042786333]
5100 [0.00042782657]
5110 [0.00042779007]
5120 [0.00042775378]
5130 [0.00042771734]
5140 [0.00042768143]
5150 [0.000427645]
5160 [0.00042760954]
5170 [0.00042757348]
5180 [0.000427538]
5190 [0.00042750238]
5200 [0.00042746737]
5210 [0.00042743166]
5220 [0.00042739644]
5230 [0.00042736158]
5240 [0.00042732668]
5250 [0.00042729155]
5260 [0.00042725675]
5270 [0.00042722252]
5280 [0.0004271881]
5290 [0.0004271534]
5300 [0.000427119]
5310 [0.0004270849]
5320 [0.00042705107]
5330 [0.00042701702]
5340 [0.00042698305]
5350 [0.00042694926]
5360 [0.00042691565]
5370 [0.00042688212]
5380 [0.00042684854]
5390 [0.00042681492]
5400 [0.000426782]
5410 [0.00042674853]
5420 [0.00042671568]
5430 [0.00042668264]
5440 [0.00042664915]
5450 [0.00042661652]
5460 [0.00042658375]
5470 [0.0004265509]
5480 [0.00042651824]
5490 [0.00042648605]
5500 [0.0004264537]
5510 [0.00042642123]
5520 [0.0004263887]
5530 [0.0004263568]
5540 [0.0004263245]
5550 [0.00042629227]
5560 [0.00042626046]
5570 [0.00042622798]
5580 [0.00042619652]
5590 [0.00042616486]
5600 [0.00042613328]
5610 [0.0004261015]
5620 [0.00042606995]
5630 [0.00042603814]
5640 [0.0004260068]
5650 [0.0004259754]
5660 [0.00042594402]
5670 [0.00042591334]
5680 [0.00042588194]
5690 [0.00042585062]
5700 [0.00042581954]
5710 [0.00042578854]
5720 [0.00042575732]
5730 [0.00042572618]
5740 [0.00042569608]
5750 [0.00042566494]
5760 [0.00042563424]
5770 [0.00042560394]
5780 [0.0004255732]
5790 [0.00042554233]
5800 [0.00042551185]
5810 [0.00042548173]
5820 [0.0004254513]
5830 [0.0004254207]
5840 [0.0004253909]
5850 [0.00042536051]
5860 [0.00042533022]
5870 [0.00042529986]
5880 [0.00042526991]
5890 [0.00042524026]
5900 [0.0004252097]
5910 [0.00042518004]
5920 [0.0004251501]
5930 [0.00042512015]
5940 [0.00042509052]
5950 [0.00042506063]
5960 [0.00042503088]
5970 [0.00042500123]
5980 [0.00042497233]
5990 [0.00042494238]
6000 [0.00042491264]
6010 [0.00042488292]
6020 [0.00042485364]
6030 [0.00042482436]
6040 [0.00042479485]
6050 [0.00042476557]
6060 [0.0004247366]
6070 [0.00042470705]
6080 [0.00042467815]
6090 [0.00042464904]
6100 [0.00042461953]
6110 [0.0004245908]
6120 [0.00042456147]
6130 [0.0004245328]
6140 [0.0004245039]
6150 [0.0004244747]
6160 [0.00042444616]
6170 [0.00042441752]
6180 [0.00042438842]
6190 [0.00042435995]
6200 [0.00042433082]
6210 [0.0004243022]
6220 [0.0004242738]
6230 [0.00042424514]
6240 [0.0004242161]
6250 [0.00042418725]
6260 [0.00042415917]
6270 [0.00042413027]
6280 [0.00042410157]
6290 [0.00042407325]
6300 [0.00042404505]
6310 [0.0004240165]
6320 [0.0004239881]
6330 [0.00042395986]
6340 [0.00042393166]
6350 [0.00042390355]
6360 [0.00042387526]
6370 [0.00042384674]
6380 [0.0004238188]
6390 [0.0004237906]
6400 [0.00042376228]
6410 [0.00042373405]
6420 [0.0004237061]
6430 [0.00042367817]
6440 [0.00042365023]
6450 [0.00042362235]
6460 [0.0004235944]
6470 [0.0004235666]
6480 [0.00042353864]
6490 [0.00042351068]
6500 [0.00042348265]
6510 [0.00042345494]
6520 [0.0004234271]
6530 [0.00042339944]
6540 [0.00042337168]
6550 [0.00042334403]
6560 [0.00042331667]
6570 [0.00042328908]
6580 [0.00042326155]
6590 [0.00042323372]
6600 [0.00042320671]
6610 [0.0004231786]
6620 [0.00042315142]
6630 [0.00042312383]
6640 [0.0004230965]
6650 [0.00042306862]
6660 [0.00042304135]
6670 [0.00042301408]
6680 [0.0004229869]
6690 [0.0004229588]
6700 [0.0004229317]
6710 [0.00042290462]
6720 [0.00042287708]
6730 [0.0004228501]
6740 [0.0004228224]
6750 [0.00042279522]
6760 [0.00042276803]
6770 [0.00042274114]
6780 [0.00042271378]
6790 [0.00042268677]
6800 [0.00042265907]
6810 [0.00042263194]
6820 [0.0004226052]
6830 [0.00042257778]
6840 [0.00042255077]
6850 [0.00042252385]
6860 [0.00042249687]
6870 [0.0004224697]
6880 [0.000422443]
6890 [0.00042241617]
6900 [0.00042238928]
6910 [0.00042236224]
6920 [0.00042233508]
6930 [0.00042230825]
6940 [0.0004222815]
6950 [0.0004222545]
6960 [0.00042222804]
6970 [0.00042220097]
6980 [0.0004221744]
6990 [0.0004221475]
7000 [0.00042212105]
7010 [0.0004220943]
7020 [0.0004220679]
7030 [0.0004220412]
7040 [0.00042201445]
7050 [0.00042198808]
7060 [0.0004219614]
7070 [0.00042193473]
7080 [0.00042190813]
7090 [0.00042188144]
7100 [0.00042185537]
7110 [0.00042182862]
7120 [0.00042180225]
7130 [0.00042177536]
7140 [0.00042174905]
7150 [0.00042172277]
7160 [0.0004216962]
7170 [0.00042166957]
7180 [0.0004216437]
7190 [0.00042161753]
7200 [0.000421591]
7210 [0.00042156447]
7220 [0.0004215385]
7230 [0.00042151252]
7240 [0.00042148613]
7250 [0.00042145967]
7260 [0.00042143368]
7270 [0.00042140714]
7280 [0.000421381]
7290 [0.00042135452]
7300 [0.00042132847]
7310 [0.00042130263]
7320 [0.00042127617]
7330 [0.0004212501]
7340 [0.00042122416]
7350 [0.000421198]
7360 [0.00042117178]
7370 [0.0004211453]
7380 [0.00042111956]
7390 [0.00042109363]
7400 [0.00042106755]
7410 [0.00042104154]
7420 [0.00042101578]
7430 [0.0004209896]
7440 [0.00042096406]
7450 [0.00042093784]
7460 [0.0004209119]
7470 [0.0004208855]
7480 [0.00042085975]
7490 [0.0004208342]
7500 [0.00042080812]
7510 [0.00042078234]
7520 [0.00042075673]
7530 [0.00042073082]
7540 [0.00042070504]
7550 [0.0004206793]
7560 [0.0004206535]
7570 [0.00042062785]
7580 [0.00042060178]
7590 [0.00042057605]
7600 [0.00042055067]
7610 [0.00042052488]
7620 [0.00042049945]
7630 [0.00042047366]
7640 [0.00042044782]
7650 [0.00042042212]
7660 [0.00042039648]
7670 [0.0004203713]
7680 [0.00042034558]
7690 [0.00042032034]
7700 [0.0004202948]
7710 [0.000420269]
7720 [0.0004202435]
7730 [0.0004202179]
7740 [0.00042019263]
7750 [0.00042016726]
7760 [0.00042014162]
7770 [0.00042011598]
7780 [0.00042009092]
7790 [0.0004200651]
7800 [0.00042003996]
7810 [0.0004200144]
7820 [0.00041998937]
7830 [0.00041996394]
7840 [0.00041993853]
7850 [0.0004199131]
7860 [0.00041988798]
7870 [0.0004198628]
7880 [0.00041983754]
7890 [0.0004198123]
7900 [0.00041978716]
7910 [0.00041976175]
7920 [0.00041973646]
7930 [0.00041971155]
7940 [0.00041968632]
7950 [0.0004196614]
7960 [0.0004196359]
7970 [0.00041961123]
7980 [0.000419586]
7990 [0.00041956038]
8000 [0.0004195358]
8010 [0.0004195103]
8020 [0.00041948535]
8030 [0.00041946006]
8040 [0.00041943497]
8050 [0.00041941024]
8060 [0.00041938515]
8070 [0.00041936003]
8080 [0.00041933465]
8090 [0.00041931018]
8100 [0.00041928503]
8110 [0.0004192599]
8120 [0.00041923486]
8130 [0.00041921035]
8140 [0.0004191856]
8150 [0.00041916012]
8160 [0.0004191359]
8170 [0.0004191105]
8180 [0.0004190857]
8190 [0.00041906128]
8200 [0.00041903622]
8210 [0.0004190112]
8220 [0.0004189867]
8230 [0.0004189619]
8240 [0.00041893742]
8250 [0.0004189124]
8260 [0.00041888765]
8270 [0.00041886323]
8280 [0.00041883835]
8290 [0.0004188135]
8300 [0.00041878893]
8310 [0.00041876428]
8320 [0.00041873954]
8330 [0.0004187148]
8340 [0.00041869044]
8350 [0.0004186657]
8360 [0.00041864076]
8370 [0.0004186163]
8380 [0.00041859187]
8390 [0.00041856675]
8400 [0.00041854221]
8410 [0.0004185176]
8420 [0.00041849338]
8430 [0.00041846873]
8440 [0.00041844416]
8450 [0.00041841972]
8460 [0.0004183951]
8470 [0.0004183705]
8480 [0.00041834655]
8490 [0.00041832202]
8500 [0.00041829728]
8510 [0.00041827292]
8520 [0.00041824867]
8530 [0.00041822388]
8540 [0.00041819952]
8550 [0.0004181754]
8560 [0.00041815094]
8570 [0.00041812644]
8580 [0.00041810234]
8590 [0.00041807737]
8600 [0.00041805336]
8610 [0.00041802882]
8620 [0.00041800435]
8630 [0.00041798034]
8640 [0.00041795612]
8650 [0.00041793176]
8660 [0.00041790743]
8670 [0.00041788348]
8680 [0.00041785912]
8690 [0.00041783496]
8700 [0.00041781052]
8710 [0.0004177863]
8720 [0.00041776203]
8730 [0.000417738]
8740 [0.00041771392]
8750 [0.00041768976]
8760 [0.00041766555]
8770 [0.00041764113]
8780 [0.00041761698]
8790 [0.00041759296]
8800 [0.00041756884]
8810 [0.00041754477]
8820 [0.00041752064]
8830 [0.00041749678]
8840 [0.00041747288]
8850 [0.00041744896]
8860 [0.00041742477]
8870 [0.00041740085]
8880 [0.00041737672]
8890 [0.00041735228]
8900 [0.00041732888]
8910 [0.0004173053]
8920 [0.0004172811]
8930 [0.00041725716]
8940 [0.00041723327]
8950 [0.00041720938]
8960 [0.0004171857]
8970 [0.00041716182]
8980 [0.00041713775]
8990 [0.00041711383]
9000 [0.00041709005]
9010 [0.00041706616]
9020 [0.0004170424]
9030 [0.0004170186]
9040 [0.00041699485]
9050 [0.00041697107]
9060 [0.00041694744]
9070 [0.0004169238]
9080 [0.00041689991]
9090 [0.0004168761]
9100 [0.0004168523]
9110 [0.00041682873]
9120 [0.00041680504]
9130 [0.00041678135]
9140 [0.0004167578]
9150 [0.00041673443]
9160 [0.00041671086]
9170 [0.0004166869]
9180 [0.00041666385]
9190 [0.0004166398]
9200 [0.00041661653]
9210 [0.00041659304]
9220 [0.00041656964]
9230 [0.0004165461]
9240 [0.0004165226]
9250 [0.00041649875]
9260 [0.00041647532]
9270 [0.00041645198]
9280 [0.0004164286]
9290 [0.0004164054]
9300 [0.0004163818]
9310 [0.0004163587]
9320 [0.00041633527]
9330 [0.0004163122]
9340 [0.00041628856]
9350 [0.00041626545]
9360 [0.00041624167]
9370 [0.00041621833]
9380 [0.0004161954]
9390 [0.00041617185]
9400 [0.00041614834]
9410 [0.00041612543]
9420 [0.0004161023]
9430 [0.00041607887]
9440 [0.00041605576]
9450 [0.00041603262]
9460 [0.00041600957]
9470 [0.00041598617]
9480 [0.00041596298]
9490 [0.0004159397]
9500 [0.0004159168]
9510 [0.00041589345]
9520 [0.00041587008]
9530 [0.00041584714]
9540 [0.0004158242]
9550 [0.0004158009]
9560 [0.0004157777]
9570 [0.0004157549]
9580 [0.00041573215]
9590 [0.00041570902]
9600 [0.00041568608]
9610 [0.0004156631]
9620 [0.0004156398]
9630 [0.00041561693]
9640 [0.00041559397]
9650 [0.00041557132]
9660 [0.00041554854]
9670 [0.00041552522]
9680 [0.00041550235]
9690 [0.00041547898]
9700 [0.00041545648]
9710 [0.00041543337]
9720 [0.00041541076]
9730 [0.000415388]
9740 [0.00041536536]
9750 [0.0004153423]
9760 [0.0004153193]
9770 [0.0004152967]
9780 [0.0004152741]
9790 [0.0004152511]
9800 [0.00041522834]
9810 [0.00041520514]
9820 [0.0004151827]
9830 [0.0004151602]
9840 [0.00041513704]
9850 [0.00041511454]
9860 [0.00041509198]
9870 [0.00041506908]
9880 [0.0004150464]
9890 [0.00041502382]
9900 [0.00041500098]
9910 [0.00041497863]
9920 [0.00041495578]
9930 [0.0004149331]
9940 [0.00041491084]
9950 [0.00041488826]
9960 [0.0004148658]
9970 [0.00041484297]
9980 [0.00041482053]
9990 [0.00041479804]

画图

prediction_value

array([[ 2.25063935e-01],
       [ 2.21927479e-01],
       [ 2.18783051e-01],
       [ 2.15631843e-01],
       [ 2.12473899e-01],
       [ 2.09309980e-01],
       [ 2.06140459e-01],
       [ 2.02966228e-01],
       [ 1.99787661e-01],
       [ 1.96605504e-01],
       [ 1.93419963e-01],
       [ 1.90231815e-01],
       [ 1.87042430e-01],
       [ 1.83851272e-01],
       [ 1.80659994e-01],
       [ 1.77468449e-01],
       [ 1.74277812e-01],
       [ 1.71088740e-01],
       [ 1.67901978e-01],
       [ 1.64718151e-01],
       [ 1.61538050e-01],
       [ 1.58362180e-01],
       [ 1.55191839e-01],
       [ 1.52027786e-01],
       [ 1.48870036e-01],
       [ 1.45719916e-01],
       [ 1.42578259e-01],
       [ 1.39446199e-01],
       [ 1.36324152e-01],
       [ 1.33212686e-01],
       [ 1.30113572e-01],
       [ 1.27026588e-01],
       [ 1.23953380e-01],
       [ 1.20894611e-01],
       [ 1.17851257e-01],
       [ 1.14823900e-01],
       [ 1.11813806e-01],
       [ 1.08821817e-01],
       [ 1.05848670e-01],
       [ 1.02895431e-01],
       [ 9.99631658e-02],
       [ 9.70522910e-02],
       [ 9.41644013e-02],
       [ 9.13001820e-02],
       [ 8.84601995e-02],
       [ 8.56461227e-02],
       [ 8.28583241e-02],
       [ 8.00981522e-02],
       [ 7.73662925e-02],
       [ 7.46635869e-02],
       [ 7.19912201e-02],
       [ 6.93499595e-02],
       [ 6.67411834e-02],
       [ 6.41653091e-02],
       [ 6.16234466e-02],
       [ 5.91165014e-02],
       [ 5.66457808e-02],
       [ 5.42116649e-02],
       [ 5.18151335e-02],
       [ 4.94576171e-02],
       [ 4.71394025e-02],
       [ 4.48619798e-02],
       [ 4.26254123e-02],
       [ 4.04314846e-02],
       [ 3.82805653e-02],
       [ 3.61734591e-02],
       [ 3.41115110e-02],
       [ 3.20949145e-02],
       [ 3.01249549e-02],
       [ 2.82023363e-02],
       [ 2.63276398e-02],
       [ 2.45018788e-02],
       [ 2.27260180e-02],
       [ 2.10004635e-02],
       [ 1.93258710e-02],
       [ 1.77035891e-02],
       [ 1.61337107e-02],
       [ 1.46172345e-02],
       [ 1.31547088e-02],
       [ 1.17471023e-02],
       [ 1.03947800e-02],
       [ 9.09853820e-03],
       [ 7.85879977e-03],
       [ 6.67633582e-03],
       [ 5.55169769e-03],
       [ 4.48533101e-03],
       [ 3.47829680e-03],
       [ 2.53045000e-03],
       [ 1.64265034e-03],
       [ 8.15197593e-04],
       [ 4.89205086e-05],
       [-6.55896845e-04],
       [-1.29903771e-03],
       [-1.87968975e-03],
       [-2.39771116e-03],
       [-2.85275211e-03],
       [-3.24420957e-03],
       [-3.57212056e-03],
       [-3.83575587e-03],
       [-4.03512269e-03],
       [-4.16994607e-03],
       [-4.23978642e-03],
       [-4.24454780e-03],
       [-4.18393081e-03],
       [-4.05778736e-03],
       [-3.86620569e-03],
       [-3.60879209e-03],
       [-3.28545552e-03],
       [-2.89607700e-03],
       [-2.44076783e-03],
       [-1.91937119e-03],
       [-1.33186474e-03],
       [-6.78390148e-04],
       [ 4.11421024e-05],
       [ 8.26611940e-04],
       [ 1.67785399e-03],
       [ 2.59482255e-03],
       [ 3.57741071e-03],
       [ 4.62519750e-03],
       [ 5.73831424e-03],
       [ 6.91618863e-03],
       [ 8.15872569e-03],
       [ 9.46560502e-03],
       [ 1.08364448e-02],
       [ 1.22709060e-02],
       [ 1.37685742e-02],
       [ 1.53289940e-02],
       [ 1.69517230e-02],
       [ 1.86364278e-02],
       [ 2.03822404e-02],
       [ 2.21888181e-02],
       [ 2.40555219e-02],
       [ 2.59817354e-02],
       [ 2.79667657e-02],
       [ 3.00099235e-02],
       [ 3.21105421e-02],
       [ 3.42677012e-02],
       [ 3.64807025e-02],
       [ 3.87487151e-02],
       [ 4.10707146e-02],
       [ 4.34460826e-02],
       [ 4.58736718e-02],
       [ 4.83524986e-02],
       [ 5.08816428e-02],
       [ 5.34600057e-02],
       [ 5.60865737e-02],
       [ 5.87602034e-02],
       [ 6.14798255e-02],
       [ 6.42441139e-02],
       [ 6.70520291e-02],
       [ 6.99022636e-02],
       [ 7.27936625e-02],
       [ 7.57248625e-02],
       [ 7.86945596e-02],
       [ 8.17015022e-02],
       [ 8.47441480e-02],
       [ 8.78213495e-02],
       [ 9.09314603e-02],
       [ 9.40733626e-02],
       [ 9.72453579e-02],
       [ 1.00446045e-01],
       [ 1.03674121e-01],
       [ 1.06927760e-01],
       [ 1.10205807e-01],
       [ 1.13506474e-01],
       [ 1.16828367e-01],
       [ 1.20169982e-01],
       [ 1.23529576e-01],
       [ 1.26905933e-01],
       [ 1.30297109e-01],
       [ 1.33701801e-01],
       [ 1.37118548e-01],
       [ 1.40545383e-01],
       [ 1.43980935e-01],
       [ 1.47423998e-01],
       [ 1.50872558e-01],
       [ 1.54325172e-01],
       [ 1.57780424e-01],
       [ 1.61236838e-01],
       [ 1.64692804e-01],
       [ 1.68146804e-01],
       [ 1.71597198e-01],
       [ 1.75043017e-01],
       [ 1.78482369e-01],
       [ 1.81913778e-01],
       [ 1.85336232e-01],
       [ 1.88747972e-01],
       [ 1.92147866e-01],
       [ 1.95534274e-01],
       [ 1.98906153e-01],
       [ 2.02262357e-01],
       [ 2.05601111e-01],
       [ 2.08921447e-01],
       [ 2.12222338e-01],
       [ 2.15502277e-01],
       [ 2.18760476e-01],
       [ 2.21995428e-01],
       [ 2.25206286e-01],
       [ 2.28391871e-01],
       [ 2.31551319e-01]], dtype=float32)

plt.figure()
plt.scatter(x_data,y_data)
plt.plot(x_data,prediction_value,"r-",lw = 3)
plt.show()

注意：

当使用tf.nn.relu()替换tf.nn.tanh()就会绘制的图像就会出现一条折线，并且loss也不会降到很低。这可能是由于relu激活函数不光滑（不连续可导）导致的。