Python TensorFlow教程

TensorFlow was developed for Google’s internal use by Google Brain team, but the system is general enough to be applied to a wide variety of domains. On November 9, 2015, they decided to open source it, and release it under Apache 2.0 open source license. Today we will look into TensorFlow basics and then some Python TensorFlow example programs.

TensorFlow是由Google Brain团队开发的，供Google内部使用，但该系统具有足够的通用性，可以应用于各种领域。 2015年11月9日，他们决定将其开源，并在Apache 2.0开源许可下发布。 今天，我们将研究TensorFlow基础知识，然后介绍一些Python TensorFlow示例程序。

TensorFlow (TensorFlow)

TensorFlow is a library for dataflow programming. It’s a symbolic math library and is also used for application of machine learning such as neural network.

TensorFlow是用于数据流编程的库。 它是一个符号数学库，还用于机器学习（例如神经网络）的应用。

TensorFlow uses various optimization techniques to make the calculations more efficient. When it comes to computation across several machines and involves huge data sets, TensorFlow offers high scalability. These features put together make TensorFlow a perfect framework for machine learning.

TensorFlow使用各种优化技术来提高计算效率。 在涉及多台机器的计算并涉及庞大的数据集时，TensorFlow具有很高的可扩展性。 这些特性组合在一起使TensorFlow成为机器学习的理想框架。

TensorFlow术语 (TensorFlow Terminologies)

In this post, we will learn more about TensorFlow and work out some examples to see how TensorFlow can be used to visualize, tweak and debug the libraries created with it, but before we get working with TensorFlow we must know what are Tensors and Tensor Processing Units.

在本文中，我们将学习更多关于TensorFlow的信息，并通过一些示例来了解如何使用TensorFlow可视化，调整和调试使用它创建的库，但是在使用TensorFlow之前，我们必须了解什么是Tensor和Tensor处理单位。

张量 (Tensors)

A tensor is central unit of data in TensorFlow. It consists of primitive values stored in shape of a multidimensional array. The number of dimensions a tensor has is called its rank.

张量是TensorFlow中数据的中心单元。 它由存储为多维数组形状的原始值组成。 张量具有的维数称为其秩 。

A rank 0 tensor is just a scalar. To keep things simple, we can say that a tensor is rather a fancy name of an array and now we call dimension number as rank. Let’s have a look at what tensors look like:

0级张量只是一个标量 。 为了简单起见，我们可以说张量是数组的一个奇特的名称，现在我们将维数称为等级。 让我们看一下张量是什么样的：

A rank 1 tensor:

1级张量：

[1,2,3,4]

A rank 2 tensor:

等级2张量：

[[1,2,3,4,],[5,6,7,8]]

张量处理单元（TPU） (Tensor Processing Units (TPUs))

TPU is a programmable AI accelerator designed to provide high throughput