ChatGPT-英语练习第一天-2023.3.24

Posted 2023-04-03 _腾飞

tags:

篇首语：本文由小常识网(cha138.com)小编为大家整理，主要介绍了ChatGPT-英语练习第一天-2023.3.24相关的知识，希望对你有一定的参考价值。

学习目标

根据英语学习计划，第一周主要是单词和语法，安装好插件后和ChatGPT聊起来！

基本对话包含常用单词和语法

学习内容

Hello can you please act as a English teacher and teach me how to?Speak English.

Of course, I'd be happy to help you improve your English speaking skills! Here are some tips to get you started:

Listen to English: One of the best ways to improve your speaking skills is to listen to English as much as possible. You can do this by watching English movies or TV shows, listening to English music, or even just listening to people speaking English around you。
Speak as much as possible: Practice makes perfect, so try to speak English as much as possible. You can practice with friends or family members who speak English, or even by speaking to yourself in English.
Focus on pronunciation: English pronunciation can be difficult, but it's important to focus on getting it right. Listen to native English speakers and try to imitate their pronunciation.
Learn new vocabulary: The more words you know, the easier it will be to express yourself in English. Try to learn new words every day and use them in conversation.
Think in English: When you're speaking English, try to think in English as well. This will help you to form sentences more quickly and fluently.
Be confident: Don't be afraid to make mistakes – everyone makes mistakes when they're learning a new language. Just keep practicing and your English will improve over time.

I hope these tips are helpful! Let me know if you have any other questions or if there's anything else I can do to help you improve your English speaking skills.

just keep practicing

Yes, practicing regularly is key to improving your English speaking skills. The more you practice, the more comfortable you will become with the language, and the more your skills will improve. Don't get discouraged

OK, let's start at.Talking.

Sure, let's start by having a conversation! Is there anything in particular you would like to talk about or any specific topic you are interested in?

Yes, I'm very interested in how to introduce myself.

Great! Introducing yourself is an important skill in English, especially when you are meeting new people. Here's an example of how to introduce yourself:

"Hi, my name is [Your Name]. I'm from [Your Country or City]. Nice to meet you!"

You can also add some additional information about yourself, such as your occupation, hobbies or interests:

"Hi, my name is [Your Name]. I'm a software engineer and I'm from [Your Country or City]. In my free time, I enjoy hiking and playing guitar. Nice to meet you!"

When introducing yourself, it's important to speak clearly and confidently. You can also try to make some small talk by asking the other person a question, such as "How about you? Where are you from?" or "What do you do for a living?"

Remember, introducing yourself is a great way to start a conversation and make new friends. Good luck!

Hi, my name is Smith. I'm from China. Nice to meet you.

Hi Smith, nice to meet you too! How are you doing today? Is there anything you'd like to talk about or any questions you have about English or anything else?

Yes, and.I'm very good, yes. And uh, let me say I will talk about my country, China. It's very beautiful.And comfortable.in my hometown

That's great to hear, Smith! China is a very beautiful country with a rich history and culture. Can you tell me a little more about your hometown and what makes it special?

Yes, let me introduce you about my hometown.My hometown is.Chairman yes, it's a.Hello.In Hubei province.And I.Live layer.In my childhood.I think it's a very beautiful.Place.You can see many kinds of lake.You can.If.Play.Many kinds of.What game?Such as swimming.Uh.

It sounds like your hometown in Hubei province is a wonderful place with a lot of natural beauty! It's great that you have fond memories of it from your childhood. I believe you were saying that there are many lakes and opportunities for activities like swimming there. That sounds like a great way to enjoy the outdoors and have fun with friends and family. Do you have a favorite lake or activity that you like to do in your hometown?

Yes, I very like. I'm like to swimming in the childhood in my hometown.And the wave of swimming is not is not seemed seemed to.The professional.Jim.Yes, it's like the.Weld the egg and.The little boy and girl to play water in there.It's very happy, but what so a little dangerous?Because the water is not very clear and safe.Maybe there are some snake there are.Sam.Umm.Animals around.Hometown.

It sounds like swimming was a very fun activity for you in your hometown, Smith! It's great to have happy childhood memories like that. However, it's important to always be careful when swimming or doing any water activities, especially in natural bodies of water where there may be wildlife like snakes or other potential hazards. It's always a good idea to swim with a buddy, stay within your comfort level, and follow any safety rules or guidelines in place. Have you continued to swim or participate in other water activities since moving away from your hometown?

No, I think I'm not good at.Anna water activities. I'm more like to play.Grant activities.

That's okay, Smith! Everyone has their own interests and preferences when it comes to activities. It's great that you enjoy playing sports or other outdoor activities. There are many great options for outdoor activities that you can enjoy in different parts of the world. Have you tried any new outdoor activities since moving away from your hometown?

Yes, I like tennis very much. It's a very interesting sport.In my life I.Have none.And the training the tennis for nearly six year, but my skill is not very good.

That's great that you enjoy playing tennis, Smith! It's a very fun and challenging sport that requires a lot of practice and skill. It's impressive that you have been training for six years, even if you don't feel like your skill level is where you would like it to be. Remember that improving in any sport or skill takes time and consistent effort. You might consider finding a coach or mentor who can help you work on your technique and develop your skills further. Keep practicing and don't give up, and I'm sure you'll continue to improve over time!

Yes, I would like.Tool agreed, but I'm very busy in my.Uh.Course, yeah, I need to.Oh.I need to.Solve the difficulty in my work? Yeah, I'm reading for a PhD.in computer science

I understand, Smith! Pursuing a PhD in computer science can be very demanding and time-consuming. It's important to prioritize your studies and work, but it's also important to take care of yourself and make time for things you enjoy, such as playing tennis or other hobbies. Finding a balance between your academic work and personal life can be a challenge, but it's important for your overall well-being. You might consider setting aside some dedicated time each week for your hobbies or interests, and try to use that time as a way to recharge and refresh yourself. Good luck with your studies, and remember to take care of yourself too!

OK. Thank you very much. I will take care of myself.Especially my body unhealthy.

You're welcome, Smith! Taking care of your physical and mental health is very important, especially when you have a demanding academic workload. Make sure to prioritize getting enough sleep, eating a healthy diet, and getting regular exercise. If you're feeling stressed or overwhelmed, try to take breaks and do things that help you relax, such as reading a book or listening to music. Remember, taking care of yourself is an investment in your overall well-being, and it can also help you perform better in your studies and other areas of your life. Good luck!

Oh, thank you very much. I think you must be a nice man. And in.Encourage me to take care of myself.

You're welcome, Smith! I'm glad that I could help encourage you to take care of yourself. It's important to prioritize our well-being and take care of our bodies and minds, especially during busy and demanding times. Remember, self-care is not selfish, it's necessary for our overall health and happiness. If you have any other questions or concerns, feel free to ask!

I also have a.Potent questions about the.Men.Tell Halsey.Yes.How could I?Keep healthy in mental.Do you have some?Umm.Have some.Suggestion.Yes, maybe you can.Uh.Tell for me to me.

Sure, Smith! Taking care of your mental health is very important, and there are many things you can do to help maintain your mental well-being. Here are some suggestions:

Practice self-care: Make time for yourself to do things you enjoy, such as reading, listening to music, or spending time with friends and family.
Exercise regularly: Exercise is a great way to relieve stress and improve your mood. Try to get at least 30 minutes of exercise each day.
Get enough sleep: Getting enough sleep is important for your mental health. Aim for 7-8 hours of sleep per night.
Eat a healthy diet: Eating a balanced and nutritious diet can help improve your mood and overall well-being.
Practice mindfulness: Mindfulness practices, such as meditation or deep breathing exercises, can help you manage stress and improve your mental clarity.
Seek support: If you're feeling overwhelmed or struggling with your mental health, don't hesitate to seek support from friends, family, or a mental health professional.

Remember, taking care of your mental health is just as important as taking care of your physical health. By making small changes to your daily routine, you can help maintain your mental well-being and improve your overall quality of life.

学习时间

周五早 7 .30- 8.10

学习总结

- 反应非常快，对话很流畅，但语音合成和输出稍慢

- 它的表述非常相关和准确，就是我的提问不太流畅，输出的提问也有识别不太准确的，我需要强化一些基础表达

- 它的回答过于长，还老是要给出一堆意见，下次对话前可以加上约束，每次三句以内进行对话

- 这样表达可能并不能有效提高英语表达能力，还需要找些B站视频学习和应用

- 每日的日记改为英文写作，还挺有趣！

基于ChatGPT制作的一款英语口语练习应用SpokenAi

本文介绍了一款基于ChatGPT的英语口语练习应用SpokenAi，包括PortAudio的安装流程和核心代码，以及语音合成TextToSpeech的实现。同时提供了配置文件和部署运行示例。 --由ChatGPT总结生成

一.前言

Hi，大家好，我是Baird。最近几个月大火大热的ChatGPT已经发布到ChatGPT4版本了，我也一直在关注ChatGPT的发展，一直在思考能基于ChatGPT或着说openai的能力能做出点什么应用，解决一些问题。

💥在仔细看过openai的API文档后，发现openai不止提供了Chat的能力，还提供了如语音转文字，图片生成等能力。虽然没有ChatGPT那么火，但经过一番试用后，发现和ChatGPT效果一样让人惊艳。索性就直接来一个openai全家桶，通过openai的能力开发一款应用试试。

开发什么呢？ 🤔

ChatGPT对英语的语言能力自然不必说，而我们国人当下英语学习面临的一大问题就是哑巴英语，市面上的提供的英语对话机器人和ChatGPT比起来差得不是一星半点。只能请老师一对一真人教学？ 🤨 拜托，现在都2023年了，还需要花钱请口语老师么？

来造一款Ai口语练习应用解决这个问题 💯

ChatGPT4是由OpenAI开发的自然语言处理模型，采用了大规模无监督学习的方式进行训练，可用于生成文本、回答问题和聊天等任务。OpenAI是人工智能领域的一家公司，其API文档提供了多种功能，包括语音转文字、图片生成等。

二.需求清单

首选先列出我们需求清单，这个是一个简单版本的英语口语练习功能，先不打算造一个大app，我们只需要解决如下问题

读取语音输入
语音转文字
通过ChatGPT沟通交流
文字转音频
音频播放
持续进行上述1～5步骤

基于上述功能，第一期我打算先做一个终端版本的应用-SpokenAi，看看后续发展再考虑做一个Web或APP版本的程序（实际上是缺人手缺时间 🤧）

三.系统架构

来，先设计一下我们SpokenAi的系统架构

portaudio: PortAudio是一个跨平台的音频I/O库，提供了简单的API，使得开发人员可以在不同的平台上以相似的方式访问音频硬件。它支持Windows、Mac OS X、Linux和其他主要的操作系统。PortAudio支持多种音频API，包括ASIO、Core Audio、DirectSound、MME / WDM、ALSA和OSS。PortAudio还包括一个流接口，允许开发人员以相同的方式使用不同的音频API和硬件。
openai-api: 提供了多种API服务，包括但不限于自然语言处理、语音转文字、文字转语音、图像生成等。在OpenAI的API文档中，用户可以申请API密钥，并使用API进行开发和测试。
tensorflowtts: TensorflowTTS 是一个基于 TensorFlow 的语音合成工具包，它包含了多种语音合成模型和前处理工具，并且支持多种语音合成任务，例如有人声合成（Vocoder）、语音转换、语音增强等。它可以帮助开发者快速搭建语音合成模型，定制自己的语音合成系统。
Docker: Docker是一种容器化技术，可以将应用程序及其依赖项打包在一个容器中，以便在任何地方运行。容器是一种轻量级的虚拟化技术，可以提供与传统虚拟机相似的隔离性和安全性，但占用的资源更少。Docker还提供了一套工具和平台，使得容器的构建、部署和管理变得更加容易。
SpokenAi:整体应用层，其中有三个库分别是praudio、rocket、ttsclient，对应如下作用
1. praudio: 封装portaudio，对外提供音频录制和音频播放等接口
2. rocket: 封装openai-api, 对外提供Chat接口、音频转文字等接口
3. ttsclient: 提供调用容器化运行的tensorflowtts的接口
Console: 终端交互层，用户按提示进行操作，输入信息和进行相关操作

Tips:tensorflowtts 依赖较多，为方便完整，这里采用Docker部署

四.流程设计

接下来，我们设计一下交互流程

简单概括总体流程有三个步骤，一是输入个人信息二是录音转文字，三是发送消息，进行对话交互

由于是通过终端访问，主要监听键盘事件，进行不同操作

如按W键会开始录音，录音过程中按Q停止录音，待录音翻译完成后，按Ctrl+shift+enter发送消息

五.开发细节

编程语言: Go 1.6+

SpokenAi使用依赖于portaudio，需要提前安装好portaudio否则编译无法通过

PortAudio

一.PortAudio下载地址:

Windows: http://www.portaudio.com/download.html
MacOS: brew install portaudio
Linux: apt-get install portaudio19-dev

Mac和Linux比较好安装，Windows只能源码编译安装，具体流程如下：

PortAudio Window 安装流程

下载PortAudio源码：http://www.portaudio.com/download.html
安装Visual Studio 2019和CMake
将源代码解压缩到 C:\\PortAudio 文件夹中
打开 CMake，选择 C:\\PortAudio 作为源代码文件夹，选择一个输出文件夹，点击 Configure 按钮
在生成选项中，选择 Visual Studio 16 2019，点击 Finish 按钮
点击 Generate 按钮
打开 Visual Studio，选择 Open a project or solution，选择 C:\\PortAudio\\build\\portaudio.sln 文件，点击打开
在 Visual Studio 中，右键点击 portaudio_static 项目，选择 Build，等待编译完成
在 Visual Studio 中，右键点击 portaudio_static 项目，选择 Install，等待安装完成
在环境变量中添加 PortAudio 的路径，例如 C:\\Program Files (x86)\\PortAudio\\bin
完成安装

PortAudio Window DLLs 安装流程

除此之外，可以直接选择网上编译好的库，可以从这里https://github.com/spatialaudio/portaudio-binaries下载。

在后续编译链接过程中，还是需要安装gcc, 推荐安装MinGW-w64。

安装完成MinGW后，将下载PortAudio库文件libportaudio-x86_64-w64-mingw32.static.dll 更名成portaudio.dll，放到MinGW ld能搜索到的库路径。

将PortAudio库文件libportaudio-x86_64-w64-mingw32.static.dll 放到Window System32目录下，在运行时程序需要找到该动态库

调用PortAudio的核心代码

录音程序

func RecordAndSaveWithContext(ctx context.Context, filename string) error 
	portaudio.Initialize()
	defer portaudio.Terminate()
	done := make(chan struct)

	// 初始化音频录制
	recorder, err := NewAudioRecorder()
	if err != nil 
		return fmt.Errorf("failed to initialize audio recorder: %v", err)
	
	defer recorder.Stop()

	// 初始化进度条
	progressBar := pb.Full.Start(maxRecordSize)
	progressBar.SetRefreshRate(time.Millisecond * 200) // 设置刷新率
	progressBar.Set(pb.Bytes, true)                    // 显示录制音频的数据量

	// 记录开始时间
	startTime := time.Now()

	// 创建Context，用于取消录音
	ctxnew, cancel := context.WithCancel(ctx)

	// 开启协程进行录音
	samples := make([]int32, 0)
	go func() 
		defer close(done)
		for 
			select 
			case <-ctxnew.Done():
				return
			case data := <-recorder.dataCh:
				// 将音频数据追加到samples
				samples = append(samples, data...)

				// 当达到最长录音时间时，取消录音
				if time.Since(startTime) >= maxRecordDuration 
					cancel()
					break
				

				// 更新录制音频的数据量
				dataSize := int64(len(samples)) * int64(reflect.TypeOf(samples).Elem().Size())
				progressBar.Add(int(dataSize)) // 更新进度条
				progressBar.SetCurrent(dataSize)
			
		
	()

	// 等待录音完成或接收到中断信号
	<-done

	progressBar.Finish()

	// 保存音频数据到WAV文件
	if err := saveToWavFile(filename, int32SliceToIntSlice(samples)); err != nil 
		return fmt.Errorf("failed to save audio to file: %v", err)
	

	return nil

主要实现功能有：

支持开始录音和停止录音
最长录制60s的音频，到时取消

2.播放音频

// PlayWavFile 播放指定的WAV文件
func PlayWavFile(filename string) error 
	// 打开WAV文件
	f, err := os.Open(filename)
	if err != nil 
		return fmt.Errorf("failed to open WAV file: %v", err)
	
	defer f.Close()

	// 解码WAV文件
	s, format, err := wav.Decode(f)
	if err != nil 
		return fmt.Errorf("failed to decode WAV file: %v", err)
	
	defer s.Close()

	// 初始化扬声器
	err = speaker.Init(format.SampleRate, format.SampleRate.N(time.Second/10))
	if err != nil 
		return fmt.Errorf("failed to initialize speaker: %v", err)
	

	// 播放音频
	done := make(chan struct)
	speaker.Play(beep.Seq(s, beep.Callback(func() 
		close(done)
	)))

	// 等待音频播放完成
	<-done

	return nil

语音合成TextToSpeech

目前比较通用的TextToSpeech（TTS）方案有以下几种：

Google Cloud Text-to-Speech
Amazon Polly
Microsoft Azure Text-to-Speech
IBM Watson Text to Speech
Mozilla TTS
Tacotron 2
WaveNet
TensorFlowTTS

有简单调用云服务的，也有自己安装环境的。调用云服务需要注册账号，按量收费，比较费钱 🙂。决定自己部署，采用TensorFlowTTS，通过构建Docker容器运行该服务。

Dockerfile如下，第一次构建速度会比较慢。

FROM tensorflow/tensorflow:2.6.0

# 安装必要的依赖
RUN apt-get update  &&\\
    apt-get install -y libsndfile1

# 安装TensorFlowTTS
RUN pip install -i https://mirrors.aliyun.com/pypi/simple/ TensorFlowTTS flask

# 安装
RUN apt-get install -y git
RUN pip install git+https://github.com/repodiac/german_transliterate.git#egg=german_transliterate
RUN pip install --upgrade numpy numba

# 安装 
ADD tts-server-api.py /app/tts-server-api.py

# 运行REST API服务器
CMD python /app/tts-server-api.py --host 0.0.0.0 --port 5000

如果不想自己自定义，可以直接下载我构建好的服务

docker pull ptonlix/tensorflowtts:1.0.9
docker run -itd -p 5000:5000 --name spokenai-tts ptonlix/tensorflowtts:1.0.9

tts-server-api.py 是启动Flask的API服务的脚本，对外提供/api/tts 文字转语音接口

@app.route('/api/tts', methods=['POST'])
def tts():
    data = request.get_json()
    text = data['text']

    # fastspeech inference
    input_ids = processor.text_to_sequence(text)
    mel_before, mel_after, duration_outputs, _, _ = fastspeech2.inference(
        input_ids=tf.expand_dims(tf.convert_to_tensor(input_ids, dtype=tf.int32), 0),
        speaker_ids=tf.convert_to_tensor([0], dtype=tf.int32),
        speed_ratios=tf.convert_to_tensor([1.0], dtype=tf.float32),
        f0_ratios=tf.convert_to_tensor([1.0], dtype=tf.float32),
        energy_ratios=tf.convert_to_tensor([1.0], dtype=tf.float32),
    )

    # melgan inference
    audio_before = mb_melgan.inference(mel_before)[0, :, 0]
    audio_after = mb_melgan.inference(mel_after)[0, :, 0]

    # save to file

    # Convert audio data to byte stream
    buffer = io.BytesIO()
    sf.write(buffer, audio_after, 22050, format='WAV', subtype='PCM_16')
    audio_bytes = buffer.getvalue()

    # Return audio data as a response with MIME type audio/wav
    return Response(audio_bytes, mimetype='audio/wav')

解决完上述两个关键问题，剩下就是业务代码的编写

配置文件

[openai]

  [openai.base]
    apikey = ""
    apihost = "https://api.openai.com/v1" 

  [openai.chat]
    chatmodel =  "gpt-3.5-turbo"
    chatmaxtoken = 2048
    chattemperature = 0.7
    chattopp = 1

  [openai.audio]
    audiomodel = "whisper-1"
  
[file]
  [file.history]
    path = "./data/history/"
  [file.audio]
    [file.audio.record]
      path = "./data/audio/record/"
    [file.audio.play]
      path = "./data/audio/play/"
      enable =  0
      ttshost = "http://127.0.0.1:5000"

采用toml配置文件格式，主要分为两部分

openai配置，主要需要填写自己的apikey和如果走代理则修改apihost地址。其他都是模型配置按需修改即可
file配置，由于是终端版本，采用文件存储的形式较为方便
- history为聊天上下文存储
- audio为音频存储
  - record为录音文件存储目录
  - play为语言合成文件存储目录
  - enable 可以选为是否开启语音合成，默认不开启,开启需要运行tensorflowtts。
  - ttshost 为tts api服务地址

六.部署运行

🚀项目地址:https://github.com/ptonlix/spokenai

🌟欢迎Star 、PR 、 Issue 、交流

编译运行

安装portaudio, 参考上节PortAudio流程

# 下载源码
git clone https://github.com/ptonlix/spokenai.git
cd spokenai

# 修改配置文件
edit fat_config.toml

# 编译
go build

# 查看命令
./spokenai -h

# 运行
./spokenai