z2539329562的博客

最近在做项目时，总是会有程序崩溃的问题，系统也没有任何提示。最后通过监控系统发现是内存溢出造成的。追查下去，发现一段类似这样的代码，在session中调用tensorflow的api进行运算：import tensorflow as tfX = tf.constant([[1,2,3], [3,2,4]], dtype=tf.float32)W = tf.constant([[1,1...

本文代码来自于其他人的博客：import tensorflow as tfimport numpy as nptf.reset_default_graph() x = np.random.randn(2,3, 4)x[1, 2:] = 0x_length = [3, 2]cell = tf.nn.rnn_cell.LSTMCell(num_units=2, state_is...

import tensorflow as tfa = tf.placeholder(tf.float32, shape=[])b = tf.constant(1, dtype=tf.int32)tf.summary.scalar("a", a)tf.summary.scalar("b", b)sess = tf.Session()init_op = tf.global_va...

import tensorflow as tfpath = "train.csv"record_defaults = [[0.0], [0.0], [0], [""], [""], [0.0], [0.0], [0.0], [""], [0.0], [""], [""]]batch_size = 10filename_queue = tf.train.string_inpu

tensorflow 执行代码时出现以下错误 OutOfRangeError: RandomShuffleQueue '_83_shuffle_batch_7/random_shuffle_queue' is closed and has insufficient elements (requested 100, current size 0)def read_csv(batch_si...

本文主要包括tf.losses.softmax_cross_entropy()、tf.nn.softmax_cross_entropy_with_logits_v2()、tf.losses.softmax_cross_entropy()、以及通过交叉熵公式实现cross_entropy  '''tf.losses.softmax_cross_entropy( onehot...

 抄袭一段：checkpoint是一个内部事件，该事件激活后会触发数据库写进程将数据缓冲中的脏数据写到数据文件中。checkpoint主要2个作用： 保证数据库的一致性缩短实例恢复时间通俗的讲，checkpoint像word的自动保存一样。tensorflow模型包含  meta图（网络结构图） 和 checkpoint文件（网络结构里的参数值，现已经被分拆为3个文件）即总...

import tensorflow as tfa = tf.Variable(1)with tf.Session() as sess: sess.run(tf.global_variables_initializer) print(sess.run(2))TypeError: Fetch argument <function global_variables_...

tf.square( x, name=None)Computes square of x element-wise.I.e., y = x*x = x^2.Returns:A Tensor or SparseTensor. Has the same type as x.tf.math.squared_difference( x, y, ...

运行tensorflow后，程序报错： FailedPreconditionError (see above for traceback): Attempting to use uninitialized value beta1_power     [[Node: beta1_power/read = Identity[T=DT_FLOAT, _class=["loc:@Adam/Ass...

import tensorflow as tfgraph = tf.Graph()arr = [i for i in range(20)]with graph.as_default(): sess = tf.Session(graph=graph) with tf.name_scope("variables"): global_steps = ...

'''tf.reduce_prod( input_tensor, axis=None, keepdims=None, name=None, reduction_indices=None, keep_dims=None)此函数计算一个张量的各个维度上元素的乘积（张量沿着某一维度计算乘积）。 Computes the product of e...

"""eval()其实就是tf.Tensor的session.run()的另一种写法，1、eval()也是启动计算的一种方式。基于tensorflow基本原理，首先需要定义图，然后计算图，其中计算图的函数有常见的run()函数，如sess.run()，eval()也是类似。2、eval()只能用于tf.tensor类对象，也就是有输出的operaton。没有输出的operation，使用...

之前一直弄混张量的维度和shape的关系，认为通过tf.shape()获得的就是维度，现在发现错误，记下来。tf.shap()用来获取的是张量的各个维度上的元素数目。1 #维度为0的标量[1, 2, 3] #维度为1，包含3个元素[[1, 2], [...

 最近看《面向机器智能tensorflow实践》这本书，里面提到在一份代码中实现多个graph。下面是一种实现方式，适合较短代码段import tensorflow as tf#c = tf.add(a, b, "add")g1 = tf.Graph()g2 = tf.Graph()with g1.as_default(): sess = tf.Session() ...

# tf.assign(# ref, # ref: A mutable Tensor. Should be from a Variable node. May be uninitialized. 待赋值的变量# value,# value: A Tensor. Must have the same type as ref. The value to be assigned ...

# tf.cond(# pred,# true_fn=None,# false_fn=None,# strict=False,# name=None,# fn1=None,# fn2=None# )# tensorflow下的三目运算符import tensorflow as tfx = tf.constant(1...

# tf.Print(# input_, # input_: A tensor passed through this op. 输入单元# data, # data: A list of tensors to print out when op is evaluated 输出单元# message=None, # mess...

from __future__ import print_functionimport numpy as npimport tensorflow as tffrom tensorflow.contrib.factorization import KMeansfrom tensorflow.examples.tutorials.mnist import input_data#mni...

在tensorflow中存在一种情况，如果数据本身是通过迭代产生的，这时无法拿到数据集进行tensorboard显示，这时可以先把计算的结果存到列表中，然后使用列表进行显示操作。代码如下：import tensorflow as tfimport numpy as npx = np.array([2*i*i + 3*i + 4 for i in range(-10, 15, 1)]...

tf.data.Dataset.from_tensor_slices(array)获取分片数据，将输入的array按照第0维度分片dataset.make_initializable_iterator() 创建dataset迭代器，需要进行初始化one_element = iterator.get_next() 从迭代器中获取一个iter的数据iterator.initializer 迭代器初始化...

import tensorflow as tfimport numpy as npfrom sklearn import datasetsfrom sklearn.cluster import KMeansfrom sklearn import svmimport mathfrom sklearn.metrics import precision_scoreimport matplo...

本方法从stack overflow获得You can create a tf.Summary object in your Python program and write it to the same tf.summary.FileWriter object that takes your TensorFlow-produced summaries using the SummaryWri...

import tensorflow as tffrom numpy.random import RandomStateimport math#hyperparametersNUM_BATCH = 100MAX_STEPS = 9hidden_size = 8n_inputs = 2n_steps = 3batch_size = 5dataset_size = n_steps ...

import tensorflow as tffrom numpy.random import RandomStateimport mathlearning_rate = 0.01MAX_STEPS = 1001rdm = RandomState(1)dataset_size = 256validation_rate = 0.2validation_size = math.flo...

今天看tensorflow时遇到 tf.TensorArray，写一个见到的教程指导使用import tensorflow as tfsess = tf.Session()x = np.arange(20) input_ta = tf.TensorArray(size=0, dtype=tf.int32, dynamic_size=True)input_ta = input_ta.uns...

程序主要包括两部分：程序的保存程序变量的重载第一部分：程序的保存import tensorflow as tfimport numpy as npimport osfrom tensorflow.examples.tutorials.mnist import input_datamnist = input_data.read_data_sets("MNIST_data/", one_ho...

学着用tensorboard在一个模型里面有训练模型和测试模型的损失函数曲线对比，上网找了好多，但是还都是一个曲线，即自己画的是这样的但是想要的是这样的：到底应该怎么样呢？简单操作：tensorboard --logdir=run1:"/home/.../summary",run2:"/home/.../summary" 其实只要在终端同时运行几个events即可，events就是summary生...