怎么用TensorFlow实现卷积神经网络

这篇文章主要介绍"怎么用TensorFlow实现卷积神经网络"，在日常操作中，相信很多人在怎么用TensorFlow实现卷积神经网络问题上存在疑惑，小编查阅了各式资料，整理出简单好用的操作方法，希望对大家解答"怎么用TensorFlow实现卷积神经网络"的疑惑有所帮助！接下来，请跟着小编一起来学习吧！

import tensorflow as tfimport numpy as npimport input_datamnist = input_data.read_data_sets('data/', one_hot=True)print("MNIST ready")n_input  = 784 # 28*28的灰度图，像素个数784n_output = 10  # 是10分类问题# 权重项weights = {    # conv1,参数[3, 3, 1, 32]分别指定了filter的h、w、所连接输入的维度、filter的个数即产生特征图个数    'wc1': tf.Variable(tf.random_normal([3, 3, 1, 32], stddev=0.1)),       # conv2，这里参数3，3同上，32是当前连接的深度是32，即前面特征图的个数，64为输出的特征图的个数    'wc2': tf.Variable(tf.random_normal([3, 3, 32, 64], stddev=0.1)),     # fc1，将特征图转换为向量，1024由自己定义    'wd1': tf.Variable(tf.random_normal([7*7*64, 1024], stddev=0.1)),     # fc2，做10分类任务，前面连1024，输出10分类    'wd2': tf.Variable(tf.random_normal([1024, n_output], stddev=0.1)) }"""特征图大小计算：f_w = (w-f+2*pad)/s + 1 = (28-3+2*1)/1 + 1 = 28 # 说明经过卷积层并没有改变图片的大小f_h = (h-f+2*pad)/s + 1 = (28-3+2*1)/1 + 1 = 28# 特征图的大小是经过池化层后改变的第一次pooling后28*28变为14*14第二次pooling后14*14变为7*7，即最终是一个7*7*64的特征图"""# 偏置项biases = {    'bc1': tf.Variable(tf.random_normal([32], stddev=0.1)),      # conv1，对应32个特征图    'bc2': tf.Variable(tf.random_normal([64], stddev=0.1)),      # conv2，对应64个特征图    'bd1': tf.Variable(tf.random_normal([1024], stddev=0.1)),    # fc1，对应1024个向量    'bd2': tf.Variable(tf.random_normal([n_output], stddev=0.1)) # fc2，对应10个输出}def conv_basic(_input, _w, _b, _keep_prob):    # INPUT    # 对图像做预处理，转换为tf支持的格式，即[n, h, w, c],-1是确定好其它3维后，让tf去推断剩下的1维    _input_r = tf.reshape(_input, shape=[-1, 28, 28, 1])     # CONV LAYER 1    _conv1 = tf.nn.conv2d(_input_r, _w['wc1'], strides=[1, 1, 1, 1], padding='SAME')     # [1, 1, 1, 1]分别代表batch_size、h、w、c的stride    # padding有两种选择：'SAME'（窗口滑动时，像素不够会自动补0）或'VALID'（不够就跳过）两种选择    _conv1 = tf.nn.relu(tf.nn.bias_add(_conv1, _b['bc1'])) # 卷积层后连激活函数    # 最大值池化，[1, 2, 2, 1]其中1,1对应batch_size和channel，2,2对应2*2的池化    _pool1 = tf.nn.max_pool(_conv1, ksize=[1, 2, 2, 1], strides=[1, 2, 2, 1], padding='SAME')    # 随机杀死一些神经元,_keepratio为保留神经元比例，如0.6     _pool_dr1 = tf.nn.dropout(_pool1, _keep_prob)     # CONV LAYER 2    _conv2 = tf.nn.conv2d(_pool_dr1, _w['wc2'], strides=[1, 1, 1, 1], padding='SAME')    _conv2 = tf.nn.relu(tf.nn.bias_add(_conv2, _b['bc2']))    _pool2 = tf.nn.max_pool(_conv2, ksize=[1, 2, 2, 1], strides=[1, 2, 2, 1], padding='SAME')    _pool_dr2 = tf.nn.dropout(_pool2, _keep_prob) # dropout    # VECTORIZE向量化    # 定义全连接层的输入，把pool2的输出做一个reshape，变为向量的形式    _densel = tf.reshape(_pool_dr2, [-1, _w['wd1'].get_shape().as_list()[0]])     # FULLY CONNECTED LAYER 1    _fc1 = tf.nn.relu(tf.add(tf.matmul(_densel, _w['wd1']), _b['bd1'])) # w*x+b，再通过relu    _fc_dr1 = tf.nn.dropout(_fc1, _keep_prob) # dropout    # FULLY CONNECTED LAYER 2    _out = tf.add(tf.matmul(_fc_dr1, _w['wd2']), _b['bd2']) # w*x+b，得到结果    # RETURN    out = {'input_r': _input_r, 'conv1': _conv1, 'pool1': _pool1, 'pool_dr1': _pool_dr1,           'conv2': _conv2, 'pool2': _pool2, 'pool_dr2': _pool_dr2, 'densel': _densel,           'fc1': _fc1, 'fc_dr1': _fc_dr1, 'out': _out           }    return outprint("CNN READY")x = tf.placeholder(tf.float32, [None, n_input]) # 用placeholder先占地方，样本个数不确定为Noney = tf.placeholder(tf.float32, [None, n_output]) # 用placeholder先占地方，样本个数不确定为Nonekeep_prob = tf.placeholder(tf.float32)_pred = conv_basic(x, weights, biases, keep_prob)['out'] # 前向传播的预测值cost = tf.reduce_mean(tf.nn.softmax_cross_entropy_with_logits(_pred, y)) # 交叉熵损失函数optm = tf.train.AdamOptimizer(0.001).minimize(cost) # 梯度下降优化器_corr = tf.equal(tf.argmax(_pred, 1), tf.argmax(y, 1)) # 对比预测值索引和实际label索引，相同返回True，不同返回Falseaccr = tf.reduce_mean(tf.cast(_corr, tf.float32)) # 将True或False转换为1或0,并对所有的判断结果求均值init = tf.global_variables_initializer()print("FUNCTIONS READY")# 上面神经网络结构定义好之后，下面定义一些超参数training_epochs = 1000 # 所有样本迭代1000次batch_size = 100 # 每进行一次迭代选择100个样本display_step = 1# LAUNCH THE GRAPHsess = tf.Session() # 定义一个Sessionsess.run(init) # 在sess里run一下初始化操作# OPTIMIZEfor epoch in range(training_epochs):    avg_cost = 0.    total_batch = int(mnist.train.num_examples/batch_size)    for i in range(total_batch):        batch_xs, batch_ys = mnist.train.next_batch(batch_size) # 逐个batch的去取数据        sess.run(optm, feed_dict={x: batch_xs, y: batch_ys, keep_prob:0.5})        avg_cost += sess.run(cost, feed_dict={x: batch_xs, y: batch_ys, keep_prob:1.0})/total_batch    if epoch % display_step == 0:        train_accuracy = sess.run(accr, feed_dict={x: batch_xs, y: batch_ys, keep_prob: 1.0})        test_accuracy = sess.run(accr, feed_dict={x: mnist.test.images, y: mnist.test.labels, keep_prob:1.0})        print("Epoch: d/d cost: %.9f TRAIN ACCURACY: %.3f TEST ACCURACY: %.3f"              % (epoch, training_epochs, avg_cost, train_accuracy, test_accuracy))print("DONE")

我用的显卡是GTX960，在跑这个卷积神经网络的时候，第一次filter分别设的是64和128，结果报蜜汁错误了，反正就是我显存不足，所以改成了32和64，让特征图少一点。所以，是让我换1080的意思喽

I c:\tf_jenkins\home\workspace\release-win\device\gpu\os\windows\tensorflow\core\common_runtime\gpu\gpu_device.cc:885] Found device 0 with properties: name: GeForce GTX 960major: 5 minor: 2 memoryClockRate (GHz) 1.304pciBusID 0000:01:00.0Total memory: 4.00GiBFree memory: 3.33GiBI c:\tf_jenkins\home\workspace\release-win\device\gpu\os\windows\tensorflow\core\common_runtime\gpu\gpu_device.cc:906] DMA: 0 I c:\tf_jenkins\home\workspace\release-win\device\gpu\os\windows\tensorflow\core\common_runtime\gpu\gpu_device.cc:916] 0:   Y I c:\tf_jenkins\home\workspace\release-win\device\gpu\os\windows\tensorflow\core\common_runtime\gpu\gpu_device.cc:975] Creating TensorFlow device (/gpu:0) -> (device: 0, name: GeForce GTX 960, pci bus id: 0000:01:00.0)W c:\tf_jenkins\home\workspace\release-win\device\gpu\os\windows\tensorflow\core\common_runtime\bfc_allocator.cc:217] Ran out of memory trying to allocate 2.59GiB. The caller indicates that this is not a failure, but may mean that there could be performance gains if more memory is available.W c:\tf_jenkins\home\workspace\release-win\device\gpu\os\windows\tensorflow\core\common_runtime\bfc_allocator.cc:217] Ran out of memory trying to allocate 1.34GiB. The caller indicates that this is not a failure, but may mean that there could be performance gains if more memory is available.W c:\tf_jenkins\home\workspace\release-win\device\gpu\os\windows\tensorflow\core\common_runtime\bfc_allocator.cc:217] Ran out of memory trying to allocate 2.10GiB. The caller indicates that this is not a failure, but may mean that there could be performance gains if more memory is available.W c:\tf_jenkins\home\workspace\release-win\device\gpu\os\windows\tensorflow\core\common_runtime\bfc_allocator.cc:217] Ran out of memory trying to allocate 3.90GiB. The caller indicates that this is not a failure, but may mean that there could be performance gains if more memory is available.Epoch: 000/1000 cost: 0.517761162 TRAIN ACCURACY: 0.970 TEST ACCURACY: 0.967Epoch: 001/1000 cost: 0.093012387 TRAIN ACCURACY: 0.960 TEST ACCURACY: 0.979...省略

到此，关于"怎么用TensorFlow实现卷积神经网络"的学习就结束了，希望能够解决大家的疑惑。理论与实践的搭配能更好的帮助大家学习，快去试试吧！若想继续学习更多相关知识，请继续关注网站，小编会继续努力为大家带来更多实用的文章！