使用JavaScript构建语音识别

Today in this article, we will discuss how to build a speech recognition tool using TensorFlow.js. Actually when we are going to build an AI, Deep Learning, Machine Learning often we think of using Python language. But what about using JavaScript? It is one of the powerful languages we can make use of a framework called TensorFlow.js.

今天，在本文中，我们将讨论如何使用TensorFlow.js构建语音识别工具。 实际上，当我们要构建AI，深度学习，机器学习时，我们常常会考虑使用Python语言。 但是使用JavaScript呢？ 它是我们可以使用称为TensorFlow.js的框架的强大语言之一。

Before getting deep into the topic, let’s see talk about TensorFlow.js.

在深入探讨该主题之前，让我们看一下TensorFlow.js 。

TensorFlow.js is a JavaScript library developed by Google for developing and deploying machine learning projects in the browser and in Node.js.

TensorFlow.js是Google开发JavaScript库，用于在浏览器和Node.js中开发和部署机器学习项目。

TensorFlow.js is not just a library. It doesn’t mean that you should not use TensorFlow.js, but I think it is a great library for deploying and developing ML applications, let’s see about in the rest of the article.

TensorFlow.js不仅仅是一个库。 这并不意味着您不应该使用TensorFlow.js，但我认为它是用于部署和开发ML应用程序的出色库，让我们在本文的其余部分中进行介绍。

使用TensorFlow.js开发项目 (Developing a project with TensorFlow.js)

As I said before, it is a powerful library, and we can work on a lot of different functions like image capturing, video manipulation, and speech recognition. For today I have decided to work on a basic speech recognition project.

如前所述，它是一个功能强大的库，我们可以处理许多不同的功能，例如图像捕获，视频处理和语音识别。 今天，我决定从事一个基本的语音识别项目。

In this code we can able to listen through the microphone and identify what the user is talking, up to a few words as we have some limitations on this sample model.

在此代码中，由于我们对该示例模型有一些限制，因此我们可以通过麦克风收听并识别用户在说什么，最多可以输入几个单词。

The first step we need to do is install the library. In order to install TensorFlow.js, a few options is available. In our project we can import it from CDN.

我们需要做的第一步是安装库。 为了安装TensorFlow.js，可以使用一些选项。 在我们的项目中，我们可以从CDN导入它。

<