【TensorFlow】tf.nn.conv2d是怎样实现卷积的？
tf.nn.conv2d是TensorFlow里面实现卷积的函数，参考文档对它的介绍并不是很详细，实际上这是搭建卷积神经网络比较核心的一个方法，非常重要
tf.nn.conv2d(input, filter, strides, padding, use_cudnn_on_gpu=None, name=None)
除去name参数用以指定该操作的name，与方法有关的一共五个参数：
第一个参数input：指需要做卷积的输入图像，它要求是一个Tensor，具有[batch, in_height, in_width, in_channels]这样的shape，具体含义是[训练时一个batch的图片数量, 图片高度, 图片宽度, 图像通道数]，注意这是一个4维的Tensor，要求类型为float32和float64其中之一
第二个参数filter：相当于CNN中的卷积核，它要求是一个Tensor，具有[filter_height, filter_width, in_channels, out_channels]这样的shape，具体含义是[卷积核的高度，卷积核的宽度，图像通道数，卷积核个数in_channels
第三个参数strides：卷积时在图像每一维的步长，这是一个一维的向量，长度4
第四个参数padding：string类型的量，只能是"SAME","VALID"其中之一，这个值决定了不同的卷积方式（后面会介绍）
第五个参数：use_cudnn_on_gpu:bool类型，是否使用cudnn加速，默认为true
结果返回一个Tensor，这个输出，就是我们常说的feature map
那么TensorFlow的卷积具体是怎样实现的呢，用一些例子去解释它：
1.考虑一种最简单的情况，现在有一张3×3单通道的图像（对应的shape：[1，3，3，1]），用一个1×1的卷积核（对应的shape：[1，1，1，1]
2.增加图片的通道数，使用一张3×3五通道的图像（对应的shape：[1，3，3，5]），用一个1×1的卷积核（对应的shape：[1，1，1，1]的feature map，这就相当于每一个像素点，卷积核都与该像素点的每一个通道做点积
input = tf.Variable(tf.random_normal([1,3,3,5]))
filter = tf.Variable(tf.random_normal([1,1,5,1])) op = tf.nn.conv2d(input, filter, strides=[1, 1, 1, 1], padding='VALID')
3.把卷积核扩大，现在用3×3的卷积核做卷积，最后的输出是一个值，相当于情况2的feature map所有像素点的值求和
input = tf.Variable(tf.random_normal([1,3,3,5]))
filter = tf.Variable(tf.random_normal([3,3,5,1])) op = tf.nn.conv2d(input, filter, strides=[1, 1, 1, 1], padding='VALID')
4.使用更大的图片将情况2的图片扩大到5×5，仍然是3×3的卷积核，令步长为1，输出3×3的feature map
input = tf.Variable(tf.random_normal([1,5,5,5]))
filter = tf.Variable(tf.random_normal([3,3,5,1])) op = tf.nn.conv2d(input, filter, strides=[1, 1, 1, 1], padding='VALID')
注意我们可以把这种情况看成情况2和情况3的中间状态，卷积核以步长1滑动遍历全图，以下x表示的位置，表示卷积核停留的位置，每停留一个，输出feature map的一个像素
.....
.xxx.
.xxx.
.xxx.
.....
5.上面我们一直令参数padding的值为‘VALID’，当其为‘SAME’时，表示卷积核可以停留在图像边缘，如下，输出5×5的feature map
input = tf.Variable(tf.random_normal([1,5,5,5]))
filter = tf.Variable(tf.random_normal([3,3,5,1])) op = tf.nn.conv2d(input, filter, strides=[1, 1, 1, 1], padding='SAME')
xxxxx
xxxxx
xxxxx
xxxxx
xxxxx
6.如果卷积核有多个
input = tf.Variable(tf.random_normal([1,5,5,5]))
filter = tf.Variable(tf.random_normal([3,3,5,7])) op = tf.nn.conv2d(input, filter, strides=[1, 1, 1, 1], padding='SAME')
此时输出7张5×5的feature map
7.步长不为1的情况，文档里说了对于图片，因为只有两维，通常strides取[1，stride，stride，1]
input = tf.Variable(tf.random_normal([1,5,5,5])) filter = tf.Variable(tf.random_normal([3,3,5,7])) op = tf.nn.conv2d(input, filter, strides=[1, 2, 2, 1], padding='SAME')
此时，输出7张3×3的feature map
x.x.x
.....
x.x.x
.....
x.x.x
8.如果batch值不为1，同时输入10张图
input = tf.Variable(tf.random_normal([10,5,5,5]))
filter = tf.Variable(tf.random_normal([3,3,5,7])) op = tf.nn.conv2d(input, filter, strides=[1, 2, 2, 1], padding='SAME')
每张图，都有7张3×3的feature map，输出的shape就是[10，3，3，7]
最后，把程序总结一下：
import tensorflow as tf
#case 2
input = tf.Variable(tf.random_normal([1,3,3,5]))
filter = tf.Variable(tf.random_normal([1,1,5,1])) op2 = tf.nn.conv2d(input, filter, strides=[1, 1, 1, 1], padding='VALID')
#case 3
input = tf.Variable(tf.random_normal([1,3,3,5]))
filter = tf.Variable(tf.random_normal([3,3,5,1])) op3 = tf.nn.conv2d(input, filter, strides=[1, 1, 1, 1], padding='VALID')
#case 4
input = tf.Variable(tf.random_normal([1,5,5,5]))
filter = tf.Variable(tf.random_normal([3,3,5,1])) op4 = tf.nn.conv2d(input, filter, strides=[1, 1, 1, 1], padding='VALID')
#case 5
input = tf.Variable(tf.random_normal([1,5,5,5]))
filter = tf.Variable(tf.random_normal([3,3,5,1])) op5 = tf.nn.conv2d(input, filter, strides=[1, 1, 1, 1], padding='SAME')
#case 6
input = tf.Variable(tf.random_normal([1,5,5,5]))
filter = tf.Variable(tf.random_normal([3,3,5,7])) op6 = tf.nn.conv2d(input, filter, strides=[1, 1, 1, 1], padding='SAME')
#case 7
input = tf.Variable(tf.random_normal([1,5,5,5]))
filter = tf.Variable(tf.random_normal([3,3,5,7])) op7 = tf.nn.conv2d(input, filter, strides=[1, 2, 2, 1], padding='SAME')
#case 8
input = tf.Variable(tf.random_normal([10,5,5,5]))
filter = tf.Variable(tf.random_normal([3,3,5,7])) op8 = tf.nn.conv2d(input, filter, strides=[1, 2, 2, 1], padding='SAME') init = tf.initialize_all_variables()
with tf.Session() as sess:
sess.run(init)
print("case 2")
print(sess.run(op2))
print("case 3")
print(sess.run(op3))
print("case 4")
print(sess.run(op4))
print("case 5")
print(sess.run(op5))
print("case 6")
print(sess.run(op6))
print("case 7")
print(sess.run(op7))
print("case 8")
print(sess.run(op8))
因为是随机初始化，我的结果是这样的：
case 2
[[[[-0.64064658]
[-1.82183945]
[-2.63191342]] [[ 8.05008984]
[ 1.66023612]
[ 2.53465152]] [[-3.51703644]
[-5.92647743]
[ 0.55595356]]]]
case 3
[[[[ 10.53139973]]]]
case 4
[[[[ 10.45460224]
[ 6.23760509]
[ 4.97157574]] [[ 3.05653667]
[-11.43907833]
[ -2.05077457]] [[ -7.48340607]
[ -0.90697062]
[ 3.27171206]]]]
case 5
[[[[ 5.30279875]
[ -2.75329947]
[ 5.62432575]
[-10.24609661]
[ 0.12603235]] [[ 0.2113893 ]
[ 1.73748684]
[ -3.04372549]
[ -7.2625494 ]
[-12.76445198]] [[ -1.57414591]
[ -3.39802694]
[ -6.01582575]
[ -1.73042905]
[ -3.07183361]] [[ 1.41795194]
[ -2.02815866]
[-17.08983231]
[ 11.98958111]
[ 2.44879103]] [[ 0.29902667]
[ -3.19712877]
[ -2.84978414]
[ -2.71143317]
[ 5.99366283]]]]
case 6
[[[[ 12.02504349 4.35077286 2.67207813 5.77893162 6.98221684
-0.96858567 -8.1147871 ]
[ -0.02988982 -2.52141953 15.24755192 6.39476395 -4.36355495
-2.34515095 5.55743504]
[ -2.74448752 -1.62703776 -6.84849405 10.12248802 3.7408421
4.71439075 6.13722801]
[ 0.82365227 -1.00546622 -3.29460764 5.12690163 -0.75699937
-2.60097408 -8.33882809]
[ 0.76171923 -0.86230004 -6.30558443 -5.58426857 2.70478535
8.98232937 -2.45504045]] [[ 3.13419819 -13.96483231 0.42031103 2.97559547 6.86646557
-3.44916964 -0.10199898]
[ 11.65359879 -5.2145977 4.28352737 2.68335319 3.21993709
-6.77338028 8.08918095]
[ 0.91533852 -0.31835344 -1.06122255 -9.11237717 5.05267143
5.6913228 -5.23855162]
[ -0.58775592 -5.03531456 14.70254898 9.78966522 -11.00562763
-4.08925819 -3.29650426]
[ -2.23447251 -0.18028721 -4.80610704 11.2093544 -6.72472
-2.67547607 1.68422937]] [[ -3.40548897 -9.70355129 -1.05640507 -2.55293012 -2.78455877
-15.05377483 -4.16571808]
[ 13.66925812 2.87588191 8.29056358 6.71941566 2.56558466
10.10329056 2.88392687]
[ -6.30473804 -3.3073864 12.43273926 -0.66088223 2.94875336
0.06056046 -2.78857946]
[ -7.14735603 -1.44281793 3.3629775 -7.87305021 2.00383091
-2.50426936 -6.93097973]
[ -3.15817571 1.85821593 0.60049552 -0.43315536 -4.43284273
0.54264796 1.54882073]] [[ 2.19440389 -0.21308756 -4.35629082 -3.62100363 -0.08513772
-0.80940366 7.57606506]
[ -2.65713739 0.45524287 -16.04298019 -5.19629049 -0.63200498
1.13256514 -6.70045137]
[ 8.00792599 4.09538221 -6.16250181 8.35843849 -4.25959206
-1.5945878 -7.60996151]
[ 8.56787586 5.85663748 -4.38656425 0.12728286 -6.53928804
2.3200655 9.47253895]
[ -6.62967777 2.88872099 -2.76913023 -0.86287498 -1.4262073
-6.59967232 5.97229099]] [[ -3.59423327 4.60458899 -5.08300591 1.32078576 3.27156973
0.5302844 -5.27635145]
[ -0.87793881 1.79624665 1.66793108 -4.70763969 -2.87593603
-1.26820421 -7.72825718]
[ -1.49699068 -3.40959787 -1.21225107 -1.11641395 -8.50123024
-0.59399474 3.18010235]
[ -4.4249506 -0.73349547 -1.49064219 -6.09967899 5.18624878
-3.80284953 -0.55285597]
[ -1.42934585 2.76053572 -5.19795799 0.83952439 -0.15203482
0.28564462 2.66513705]]]]
case 7
[[[[ 2.66223097 2.64498258 -2.93302107 3.50935125 4.62247562
2.04241085 -2.65325522]
[ -0.03272867 -1.00103927 -4.3691597 2.16724801 7.75251007
-4.6788125 -0.89318085]
[ 4.74175072 -0.80443329 -1.02710629 -6.68772554 4.57605314
-3.72993755 4.79951382]] [[ 5.249547 8.92288399 7.10703182 -9.10498428 -7.43814278
-8.69616318 1.78862095]
[ 7.53669024 -14.52316284 -2.55870199 -1.11976743 3.81035042
2.45559502 -2.35436153]
[ 3.93275881 5.11939669 -4.7114296 -11.96386623 2.11866689
0.57433248 -7.19815397]] [[ 0.25111672 1.40801668 1.28818977 -2.64093828 0.98182392
3.69512987 4.78833389]
[ 0.30391204 -10.26406097 6.05877018 -6.04775047 8.95922089
0.80235004 -5.4520669 ]
[ -7.24697018 -2.33498096 -10.20039558 -1.24307609 3.99351597
-8.1029129 2.44411373]]]]
case 8
[[[[ -6.84037447e+00 1.33321762e-01 -5.09891272e+00 5.55682087e+00
8.22002888e+00 -4.94586229e-02 4.19012117e+00]
[ 6.79884481e+00 1.21652853e+00 -5.69557810e+00 -1.33555794e+00
3.24849486e-01 4.88868570e+00 -3.90220714e+00]
[ -3.53190374e+00 -4.11765718e+00 4.54340839e+00 1.85549557e+00
-3.38682461e+00 2.62719369e+00 -4.98658371e+00]] [[ -9.86354351e+00 -6.76713943e+00 3.62617874e+00 -6.16720629e+00
1.96754158e+00 -4.54203081e+00 -1.37485743e+00]
[ -1.76783955e+00 2.35163045e+00 -2.21175838e+00 3.83091879e+00
3.16964531e+00 -7.58307219e+00 4.71943617e+00]
[ 1.20776439e+00 4.86006308e+00 1.04233503e+01 -7.82327271e+00
5.39195156e+00 -6.31672382e+00 1.35577369e+00]] [[ -3.65947580e+00 -1.98961139e+00 7.53771305e+00 2.79224634e-01
-2.90050888e+00 -3.57466817e+00 -6.33232594e-01]
[ 5.89931488e-01 2.83219159e-01 -1.65850735e+00 -6.45545387e+00
-1.17044592e+00 1.40343285e+00 5.74970901e-01]
[ -8.58810043e+00 -1.25172977e+01 6.84177876e-01 3.80004168e+00
-1.54420209e+00 -3.32161427e+00 -1.05423713e+00]]] [[[ -4.82677078e+00 3.11167526e+00 -4.32694483e+00 -4.77198696e+00
2.32186103e+00 1.65402293e-01 -5.32707453e+00]
[ 3.91779566e+00 6.27949667e+00 2.32975650e+00 -1.06336937e+01
4.44044876e+00 8.08288479e+00 -5.83346319e+00]
[ -2.82141399e+00 -9.16103745e+00 6.98908520e+00 -5.66505909e+00
-2.11039782e+00 2.27499461e+00 -5.74120235e+00]] [[ 6.71680808e-01 -4.01104212e+00 -4.61760712e+00 1.02667952e+01
-8.21200657e+00 -8.57054043e+00 1.71461976e+00]
[ 2.40794683e+00 -2.63071585e+00 9.68963623e+00 -4.51778412e+00
-3.91073084e+00 -5.91874409e+00 9.96273613e+00]
[ 2.67705870e+00 2.85607010e-01 2.45853162e+00 4.44810390e+00
-2.11300468e+00 -5.77583075e+00 2.83322239e+00]] [[ -8.21949577e+00 -7.57754421e+00 3.93484974e+00 2.26189137e+00
-3.49395227e+00 -6.40283823e+00 -6.00450039e-01]
[ 2.95964479e-02 -1.19976890e+00 5.38537979e+00 4.62369967e+00
3.89780998e+00 -6.36872959e+00 7.12107182e+00]
[ -8.85006547e-01 1.92706418e+00 3.26668215e+00 2.03566647e+00
1.44209075e+00 -6.48463774e+00 -8.33671093e-02]]] [[[ -2.64583921e+00 3.86011934e+00 4.18198538e+00 3.50338411e+00
6.35944796e+00 -4.28423309e+00 4.87355423e+00]
[ 4.42271233e+00 3.92883778e+00 -5.59371090e+00 4.98251200e+00
-3.45068884e+00 2.91921115e+00 1.03779554e+00]
[ 1.36162388e+00 -1.06808968e+01 -3.92534947e+00 1.85111761e-01
-4.87255526e+00 1.66666222e+01 -1.04918976e+01]] [[ -4.34632540e+00 1.74614882e+00 -2.89012527e+00 -8.74067783e+00
5.06610107e+00 1.24989772e+00 -3.06433105e+00]
[ 2.49973416e+00 2.14041996e+00 -4.71008825e+00 7.39326143e+00
3.94770741e+00 8.23049164e+00 -1.67046225e+00]
[ -2.94665837e+00 -4.58543825e+00 7.21219683e+00 1.09780006e+01
5.17258358e+00 7.90257788e+00 -2.13929534e+00]] [[ 4.20402241e+00 -2.98926830e+00 -3.89006615e-01 -8.16001511e+00
-2.38355541e+00 1.42584383e+00 -5.46632290e+00]
[ 5.52395058e+00 5.09255171e+00 -1.08742390e+01 -4.96262169e+00
-1.35298109e+00 3.65663052e-01 -3.40589857e+00]
[ -6.95647061e-01 -4.12855625e+00 2.66609401e-01 -9.39565372e+00
-3.85058141e+00 2.51248240e-01 -5.77149725e+00]]] [[[ 1.22103825e+01 5.72040796e+00 -3.56989503e+00 -1.02248180e+00
-5.20942688e-01 7.15008640e+00 3.43482435e-01]
[ 6.01409674e+00 -1.59511256e+00 -6.48080063e+00 -1.82889538e+01
-1.03537569e+01 -1.48270035e+01 -5.26662111e+00]
[ 5.51758146e+00 -2.91831636e+00 3.75461340e-01 -9.23893452e-02
-9.22101116e+00 7.16952372e+00 -6.86479330e-01]] [[ -3.03645611e+00 6.68620300e+00 -3.31973934e+00 -4.91346550e+00
9.20719814e+00 -2.55552864e+00 -2.16087699e-02]
[ -3.02986956e+00 -1.29726543e+01 1.53023469e+00 -8.19733238e+00
5.68085670e+00 -1.72856820e+00 -4.69369221e+00]
[ -6.67176056e+00 8.76355553e+00 2.18996063e-01 -4.38777208e+00
-6.35764122e-01 -1.37812555e+00 -4.41474581e+00]] [[ 2.25345469e+00 1.02142305e+01 -1.71714854e+00 -5.29060185e-01
2.27982092e+00 -8.75302982e+00 7.13998675e-02]
[ -6.67547846e+00 3.67722750e+00 -3.44172812e+00 5.69674826e+00
-2.28723526e+00 5.92991543e+00 5.53608060e-01]
[ -1.01174891e-01 -2.73731589e+00 -4.06187654e-01 6.54158068e+00
2.59603882e+00 2.99202776e+00 -2.22350287e+00]]] [[[ -1.81271315e+00 2.47674489e+00 -2.90284491e+00 1.34291325e+01
7.69864845e+00 -1.27134466e+00 3.02233839e+00]
[ -2.08135307e-01 1.03206539e+00 1.90775347e+00 9.01517391e+00
-3.52140331e+00 9.05393791e+00 -9.12732124e-01]
[ 1.12128162e+00 5.98179293e+00 -2.27206993e+00 -5.21281779e-01
6.20835352e+00 3.73474598e+00 1.18961644e+00]] [[ 3.17242837e+00 -6.00571585e+00 2.37661076e+00 -5.64483738e+00
-6.45412731e+00 8.75251675e+00 7.33790398e-02]
[ 3.08957529e+00 -1.06855690e-01 -5.16810894e-01 -9.41085911e+00
8.23878098e+00 6.79738426e+00 -1.23478663e+00]
[ -9.20640087e+00 -6.82801771e+00 -5.96975613e+00 7.61030674e-01
-4.35995817e+00 -3.54818010e+00 -2.56281614e+00]] [[ 4.69872713e-01 8.36402321e+00 5.37103415e-01 -1.68033957e-01
-3.21731424e+00 -7.34270859e+00 -3.14253521e+00]
[ 6.69656086e+00 -5.27954197e+00 -8.57314682e+00 4.84328842e+00
-2.96387672e+00 2.47114658e+00 2.85376692e+00]
[ -7.86032295e+00 -7.18845367e+00 -3.27161223e-01 9.27330971e+00
-6.14093494e+00 -4.49041557e+00 3.47160912e+00]]] [[[ -1.89188433e+00 5.43082857e+00 6.04252160e-01 6.92894220e+00
8.59178162e+00 1.02003086e+00 5.31300211e+00]
[ -8.97491455e-01 6.52438164e+00 -4.43710327e+00 7.10509634e+00
8.84234428e+00 3.08552694e+00 2.78152227e+00]
[ -9.40537453e-02 2.34666920e+00 -5.57496691e+00 -8.62346458e+00
-1.32807600e+00 -8.12027454e-02 -9.00946975e-01]] [[ -3.53673506e+00 8.93675327e+00 3.27456236e-01 -3.41519475e+00
7.69804525e+00 -5.18698692e+00 -3.96991730e+00]
[ 1.99988627e+00 -9.16149998e+00 -7.49944544e+00 5.02162695e-01
3.57059622e+00 9.17566013e+00 -1.77589107e+00]
[ -1.18147678e+01 -7.68992901e+00 1.88449645e+00 2.77643538e+00
-1.11342735e+01 -3.12916255e+00 -3.34161663e+00]] [[ -3.62668943e+00 -3.10993242e+00 3.60834384e+00 4.69678783e+00
-1.73794723e+00 -1.27035933e+01 3.65882218e-01]
[ -8.97550106e+00 -4.33533072e-01 4.41743970e-01 -5.83433771e+00
-4.85818958e+00 9.56629372e+00 3.56375504e+00]
[ -6.87092066e+00 1.96412420e+00 5.14182663e+00 -8.97769547e+00
3.61136627e+00 5.91387987e-01 -2.95224571e+00]]] [[[ -1.11802626e+00 3.24175072e+00 5.94067669e+00 9.29727936e+00
9.28199863e+00 -4.80889034e+00 6.96202660e+00]
[ 7.23959684e+00 3.11182523e+00 1.84116721e+00 5.12095928e-01
-7.65049171e+00 -4.05325556e+00 5.38544941e+00]
[ 4.66621685e+00 -1.61665392e+00 9.76448345e+00 2.38519001e+00
-2.06760812e+00 -6.03633642e-01 3.66192675e+00]] [[ 1.52149725e+00 -1.84441996e+00 4.87877655e+00 2.96750760e+00
2.37311172e+00 -2.98487616e+00 9.98114228e-01]
[ 9.20035839e+00 5.24396753e+00 -2.57312679e+00 -7.26040459e+00
-1.17509928e+01 6.85688591e+00 3.37383580e+00]
[ 6.17629957e+00 -5.15294194e-01 -1.64212489e+00 -5.70274448e+00
-2.36294913e+00 2.60432816e+00 2.63957453e+00]] [[ 7.91168213e-03 -1.15018034e+00 3.05471039e+00 3.31086922e+00
5.35744762e+00 1.14832592e+00 9.56500292e-01]
[ 4.86464739e+00 5.37348413e+00 1.42920148e+00 1.62809372e+00
2.61656570e+00 7.88479471e+00 -6.09324336e-01]
[ 7.71319962e+00 -1.73930550e+00 -2.99925613e+00 -3.14857435e+00
3.19194889e+00 1.70928288e+00 4.90955710e-01]]] [[[ -1.79046512e+00 8.54369068e+00 1.85044312e+00 -9.88471413e+00
9.52995300e-01 -1.34820042e+01 -1.13713551e+01]
[ 8.37582207e+00 6.64692163e+00 -3.22429276e+00 3.37997460e+00
3.91468263e+00 6.96061993e+00 -1.18029404e+00]
[ -2.13278866e+00 4.36152029e+00 -4.14593410e+00 -2.15160155e+00
1.90767622e+00 1.16321917e+01 -3.72644544e+00]] [[ -5.03508925e-01 -6.33426476e+00 -1.06393566e+01 -6.49301624e+00
-6.31036520e+00 3.13485146e+00 -5.77433109e-01]
[ 7.41444230e-01 -4.87326956e+00 -5.98253345e+00 -9.14121056e+00
-8.64077091e-01 2.06696177e+00 -7.59688473e+00]
[ 1.38767815e+00 1.84418947e-01 5.72539902e+00 -2.07557893e+00
9.70911503e-01 1.16765432e+01 -1.40111232e+00]] [[ -1.21869087e+00 2.44499159e+00 -1.65706706e+00 -6.19807529e+00
-5.56950712e+00 -1.72372568e+00 3.62687564e+00]
[ 2.23708963e+00 -2.87862611e+00 2.71666467e-01 4.35115099e+00
-8.85548592e-01 2.91860628e+00 8.10848951e-01]
[ -5.33635712e+00 7.15072036e-01 5.21240902e+00 -3.11152220e+00
2.01623154e+00 -2.28398323e-01 -3.23233747e+00]]] [[[ 3.77991509e+00 5.53513861e+00 -1.82022047e+00 4.22430277e+00
5.60331726e+00 -4.28308249e+00 4.54524136e+00]
[ -5.30983162e+00 -3.45605731e+00 2.69374561e+00 -6.16836596e+00
-9.18601036e+00 -1.58697796e+00 -5.73809910e+00]
[ 2.18868661e+00 6.96338892e-01 1.88057957e+01 -4.21353197e+00
1.20818818e+00 2.85108542e+00 6.62180042e+00]] [[ 1.01285219e+01 -4.86819077e+00 -2.45067930e+00 7.50106812e-01
4.37201977e+00 4.78472042e+00 1.19103444e+00]
[ -3.26395583e+00 -5.59358537e-01 1.52001972e+01 -5.93994498e-01
-1.49040818e+00 -7.02547312e+00 -1.29268813e+00]
[ 1.02763653e+01 1.31108007e+01 -2.91605043e+00 -1.37688947e+00
3.33029580e+00 1.96966705e+01 2.55259371e+00]] [[ 4.58397627e+00 -3.19160700e+00 -6.51985502e+00 1.02908373e+01
-4.17618275e+00 -9.69347239e-01 7.46259832e+00]
[ 6.09876537e+00 1.33044279e+00 5.04027081e+00 -6.87740147e-01
4.14770365e+00 -2.26751328e-01 1.54876924e+00]
[ 2.70127630e+00 -1.59834003e+00 -1.82587504e+00 -5.92888784e+00
-5.65038967e+00 -6.46078014e+00 -1.80765367e+00]]] [[[ -1.57899165e+00 3.39969063e+00 1.02308102e+01 -7.77082300e+00
-8.02129686e-01 -3.67387819e+00 -1.37204361e+00]
[ 3.93093729e+00 6.17498016e+00 -1.41695750e+00 -1.26903206e-01
2.18985319e+00 5.83657503e-01 7.39725351e-01]
[ 5.53898287e+00 2.22283316e+00 -1.10478985e+00 2.68644023e+00
-2.59913635e+00 3.74231935e+00 4.85016155e+00]] [[ 4.05368614e+00 -3.74058294e+00 7.32348633e+00 -1.17656231e+00
3.71810269e+00 -1.63957381e+00 9.91670132e-01]
[ -1.29317007e+01 1.12296543e+01 -1.13844347e+01 -7.13933802e+00
-8.65884399e+00 -5.56065178e+00 -1.46718264e+00]
[ -8.08718109e+00 -1.98826480e+00 -4.07488203e+00 2.06440473e+00
1.13524094e+01 5.68703651e+00 -2.18706942e+00]] [[ 1.51166654e+00 -6.84034204e+00 9.33474350e+00 -4.80931902e+00
-6.24172688e-02 -4.21381521e+00 -5.73313046e+00]
[ -1.35943902e+00 5.27799511e+00 -3.77813816e+00 6.88291168e+00
4.35068893e+00 -1.02540245e+01 8.86861205e-01]
[ -4.49999619e+00 -2.97630525e+00 -6.18604183e-01 -2.49702692e+00
-6.76169348e+00 -2.55930996e+00 -2.71291423e+00]]]]
【TensorFlow】tf.nn.conv2d是怎样实现卷积的？的更多相关文章
- tf入门-tf.nn.conv2d是怎样实现卷积的？
转自:https://blog.csdn.net/mao_xiao_feng/article/details/78004522 实验环境:tensorflow版本1.2.0,python2.7 介绍 ...
- 【Tensorflow】tf.nn.atrous_conv2d如何实现空洞卷积？膨胀卷积
介绍关于空洞卷积的理论可以查看以下链接,这里我们不详细讲理论: 1.Long J, Shelhamer E, Darrell T, et al. Fully convolutional network ...
- 深度学习原理与框架-Tensorflow卷积神经网络-卷积神经网络mnist分类 1.tf.nn.conv2d(卷积操作) 2.tf.nn.max_pool(最大池化操作) 3.tf.nn.dropout(执行dropout操作) 4.tf.nn.softmax_cross_entropy_with_logits(交叉熵损失) 5.tf.truncated_normal(两个标准差内的正态分布)
1. tf.nn.conv2d(x, w, strides=[1, 1, 1, 1], padding='SAME') # 对数据进行卷积操作 参数说明:x表示输入数据,w表示卷积核, stride ...
- tf.nn.conv2d。卷积函数
tf.nn.conv2d是TensorFlow里面实现卷积的函数,参考文档对它的介绍并不是很详细,实际上这是搭建卷积神经网络比较核心的一个方法,非常重要 tf.nn.conv2d(input, fil ...
- tf.nn.conv2d实现卷积的过程
#coding=utf-8 import tensorflow as tf #case 2 input = tf.Variable(tf.round(10 * tf.random_normal([1, ...
- 【Tensorflow】tf.nn.depthwise_conv2d如何实现深度卷积?
版权声明:本文为博主原创文章,遵循CC 4.0 BY-SA版权协议,转载请附上原文出处链接和本声明. 本文链接:https://blog.csdn.net/mao_xiao_feng/article/ ...
- TensorFlow使用记录 (二）： 理解tf.nn.conv2d方法
方法定义 tf.nn.conv2d(input, filter, strides, padding, use_cudnn_on_gpu=True, data_format="NHWC&quo ...
- tf.nn.conv2d卷积函数之图片轮廓提取
一.tensorflow中二维卷积函数的参数含义:def conv2d(input, filter, strides, padding, use_cudnn_on_gpu=True, data_for ...
- 【TensorFlow】理解tf.nn.conv2d方法 ( 附代码详解注释 )
最近在研究学习TensorFlow,在做识别手写数字的demo时,遇到了tf.nn.conv2d这个方法,查阅了官网的API 发现讲得比较简略,还是没理解.google了一下,参考了网上一些朋友写得博 ...
随机推荐
- myeclipse之完全破解
并不是所有的破解都是成功的,就如并不是所有的战争都会胜利一样,我们在做事情的时候,总会遇到些问题,比如Activate不成功,需要手动激活. 激活不成功就是不成功,来回的破解.卸载.重装,都还是不可能 ...
- Freemarker自定义方法
在项目中有一个需求,每个物品有一个guid,存在数据库中,而在页面上需要显示一个对应的业务数据值,暂且叫做serverId,serverId是通过guid移位计算得来.serverId只需要显示,后台 ...
- JS对象序列化为JSON对象的方法
var $ = $ || {}; /** * 将JS对象序列化为JSON字符串 * @param {Mixed} o The variable to decode * @return {String} ...
- UVALive 3942 Remember the Word 字典树+dp
/** 题目:UVALive 3942 Remember the Word 链接:https://vjudge.net/problem/UVALive-3942 题意:给定一个字符串(长度最多3e5) ...
- 编写可维护的JavaScript----笔记（三）
1.块语句花括号的使用 在JavaScript中,注入if和for语句有两种写法,使用花括号包裹的多行代码或者不使用花括号的单行代码.但强烈建议不论是单行还是多行代码,都应该使用花括号. 2.花括号的 ...
- JXTA 2: 具有高性能、海量伸缩性的 P2P 网络
这是来自developerWorks的一篇文章,地址是 https://www.ibm.com/developerworks/cn/java/j-jxta2/ ———————————————————— ...
- 解决Deepin每次打开Chome都提示解锁登录密钥环的麻烦
密钥环是linux系统用于安全保存程序私密数据的模块,可以用于加密保存密码.证书.密钥等安全数据.chrome的密钥环用于保存本地访问站点密码或缓存从google服务器同步下来的访问站点的密码. De ...
- EasyUI获取DataGrid中某一列的所有值
function count() { var rows = $('#dg'').datagrid('getRows')//获取当前页的数据行 var total = 0; for (var i = 0 ...
- android 开发积累
1.ListView滚动黑屏问题 ListView滚动时,数据项变成黑色 问题解决办法:通过添加 android:cacheColorHint = "#00000000" 将背景设 ...
- Redis ＂MISCONF Redis is configured to save RDB snapshots, but is currently not able to persist on disk＂问题
今天在程序中,jedis put数据到redis过程中,“MISCONF Redis is configured to save RDB snapshots, but is currently not ...