Voice + GPT | Build your own AI assistant connected to GPT | Python | 2023

Поделиться
HTML-код
  • Опубликовано: 1 дек 2024

Комментарии • 57

  • @ShaneTrace
    @ShaneTrace Год назад +1

    you're a saint my guy! This gives my Star Trek inches a scratch . Good on you mate!

  • @alexxenaosas2416
    @alexxenaosas2416 Год назад

    Hi there , super awesome ! - I will build my personal assistent for sure :)

  • @joaocarlosleme691
    @joaocarlosleme691 Год назад +1

    Awesome. Instead of using either bing api or chatgpt, I wish to use something like AutoGPT, and let GPT decide if it needs to access the internet or not. Hopefully it's a good idea for another video. Thanks a lot.

    • @mikaelcodes
      @mikaelcodes  Год назад +1

      It's the future... But I see a flood of AutoGPT videos too, I wonder if I can add value.

    • @joaocarlosleme691
      @joaocarlosleme691 Год назад

      @@mikaelcodes yes, but haven't found a single one that implemented speech recognition.

  • @comeonwindows7
    @comeonwindows7 Год назад

    Hi mikael, coild you please tell me where you get that theme for vsc? Thanks

    • @comeonwindows7
      @comeonwindows7 Год назад

      Could*

    • @mikaelcodes
      @mikaelcodes  Год назад +1

      It's available through VS Code extensions and called Cyberpunk Umbra Protocol

  • @yaseenakh6225
    @yaseenakh6225 10 месяцев назад

    Hi mikael, in the new version of this davinci is now under a deprecation and can no longer be used in the python code as for the other fine tuning models I don't believe they work either because they have also been depreciated, how do I solve this issue to use GPT in python please respond 🙏

  • @gianni4302
    @gianni4302 Год назад

    would it be possible to get this running solo on a pi?

  • @alexbarton9012
    @alexbarton9012 Год назад +1

    UNDER
    04:34 Coding the AI assistant what is my Organization ID ?

    • @alexbarton9012
      @alexbarton9012 Год назад

      i dont want to give up on this project . SOmeone help me get this audio AI turned on . I FOLLOWED STEP BY STEP AND WAS LEFT WITH ( PROLEMS

    • @mikaelcodes
      @mikaelcodes  Год назад +1

      You can get the org ID from the OpenAI settings page once you've created an account.

  • @류연석-x9b
    @류연석-x9b Год назад +1

    Amazing!
    ...But I have some questions.
    If I understood it correctly, this code is not using the "Chat GPT". Can't we use it in our code yet? If I'm accepted in the openai chat GPT test something, will it be able for me to use in in the code?
    How should I change the code to make it respond to keybord input(ex: shift+a+b) + "Sara", not "Computer, insite ~~"?
    Lastly, if I only need GPT for my assistant, will it be okay for me to just wipe out all the other parts like wikipedia, woldframalpha, webbrowser?

    • @mikaelcodes
      @mikaelcodes  Год назад +1

      Hey, that's right, ChatGPT only colloquially but it's tapping the text-davinci-003 model, but as I understand ChatGPT is also fine-tuned further towards dialogue. I'm sure ChatGPT will be available through an API eventually, and I think there are unofficial ways of accessing it anyway. I mean if you are able to do something through a browser like post a text and get a response, then it can be automated. For key commands try the 'keyboard' package www.geeksforgeeks.org/how-to-detect-if-a-specific-key-pressed-using-python/. And yes, you can definitely wipe out all the other methods to simplify the code to just GPT. I'll put such a variant of the code up soon.

    • @류연석-x9b
      @류연석-x9b Год назад +1

      @@mikaelcodes thank you so much!!

    • @mikaelcodes
      @mikaelcodes  Год назад +1

      @@류연석-x9b No worries :)

  • @minhscollection6209
    @minhscollection6209 Год назад

    Nice video, I did as you have done, the code runs smoothly, but it seems like my AI does not hear what I say. The TERMINAL only shows 'Listening for a command,' and nothing happens.

    • @mikaelcodes
      @mikaelcodes  Год назад

      Hey, thank you. If you are building it in Windows, there seems to be a bug in PyAudio, so that could explain this. I'm working on a quick video on how to fix this but it's basically the same approach as this: ruclips.net/video/T33yvP64WuI/видео.html

  • @Kshitiz457
    @Kshitiz457 Год назад

    Bro plz tell me how to add translate Command
    When I say give me answer hindi but I will ask in english

  • @SweetChillieBeat
    @SweetChillieBeat Год назад +1

    I've never coded a thing in my life, could I follow this and make this happen?

    • @mikaelcodes
      @mikaelcodes  Год назад +1

      It should be possible if you start with the earlier video on creating your own AI assistant. All the code is there. Getting to grips with Python and installing the relevant dependencies might be something where it's worth watching an intro video somewhere else if you get stuck, I cover it relatively rapidly.

  • @m9wshideout613
    @m9wshideout613 Год назад

    Hello sir i've watched both of your videos on this AI series i loved every bit of it altho i've had a problem this time because when i ask the AI a question it seems to respond with something random for example i have asked it "what is 3+1" and it has responded back with a description of the British television series A touch of frost.
    also what are other engines that i can use besides Davinci?

    • @mikaelcodes
      @mikaelcodes  Год назад

      Glad you're liking them, I think I'll make some more in the series. That's a strange response indeed - have you set the temperature quite high? That hyperparameter can lead to some very random results. New engines with APIs are coming out all the time. I haven't had time to try them out, but here's one for example www.databricks.com/blog/2023/04/12/dolly-first-open-commercially-viable-instruction-tuned-llm

  • @michaelmccenna1378
    @michaelmccenna1378 Год назад

    Hey man you're a genius. I've been waiting for a video like this I've been thinking about something like this for a few days and watched some of your other videos earlier.
    Can I use this ai system to connect to different websites and do tasks for me?
    Thanks again for the videos really.

    • @mikaelcodes
      @mikaelcodes  Год назад +1

      Thanks! And sure, Python is pretty great for automating website tasks. Take a look at Selenium. Or if you have an example that would be interesting for a general audience, I could make a video.

    • @michaelmccenna1378
      @michaelmccenna1378 Год назад

      @@mikaelcodes Cool I'll check it out.
      One of my freelance jobs is transcribing videos on different websites. Is possible to fully automate that kind of work?

    • @michaelmccenna1378
      @michaelmccenna1378 Год назад

      @@mikaelcodes I obviously download the audio from the videos first and then turn generate into text

    • @mikaelcodes
      @mikaelcodes  Год назад +1

      @@michaelmccenna1378 Downloading the audio and automatically transcribing sound like things that can be automated these days. Various Python webscraping tools can make it pretty easy to grab any videos on sites.

  • @amirulmukminin614
    @amirulmukminin614 Год назад

    hey anyone please help, i try this for my project but its not talking back to me or answer my question, its only write out my speech into text. can anyone tell me where is the problem? i need to fix it asap for my project. If this works then i need to implement this into website that has some interface.

  • @medivir6823
    @medivir6823 Год назад

    Hey man, great video! I am having a error when trying to open a website, this is the error message it is giving. I am quite new to coding so i dont really know how to fix it:string index out of range
    Exception has occurred: IndexError
    string index out of range
    File "ai assistant\Assistant.py", line 223, in
    if query[0] == 'Wikipedia':

    • @mikaelcodes
      @mikaelcodes  Год назад

      That error means that when the code gets to that point the query list variable is empty. So when it's trying to reach the first element of the list, 0, it finds nothing. So something's not yet working right in getting the query to be fillled earlier on in the code. Probably something in the speech recognition part.

  • @coderxgamer2
    @coderxgamer2 Год назад

    hi mam I don't know why when I did everything its saying dontev its not working the module is not recognized please bro can you help

    • @mikaelcodes
      @mikaelcodes  Год назад

      Did something go wrong with the installation? Try running 'pip install python-dotenv' again to see that it's available for you, and check the docs here: pypi.org/project/python-dotenv/

  • @ibrahimjaballa2706
    @ibrahimjaballa2706 Год назад

    I am a beginner
    When using gpt chat, do I have to get the API from it, or can I get it from anywhere?

    • @mikaelcodes
      @mikaelcodes  Год назад

      You may find projects like proxy APIs that allow you to query ChatGPT for free, but I haven't explored them: github.com/ayaka14732/ChatGPTAPIFree

  • @matteodeamicis
    @matteodeamicis Год назад +1

    Thanks for the great video, it helped me a lot. Is it possible to create a personal voice assistant that uses pdf files with langchain? Thanks again :)

    • @mikaelcodes
      @mikaelcodes  Год назад +2

      Glad you like it! Short answer... yes. Worth doing a video for? So you'd load your own pdf and then ask questions about it using your voice?

    • @matteodeamicis
      @matteodeamicis Год назад

      @@mikaelcodes Exactly! A video about it would be great.

    • @mikaelcodes
      @mikaelcodes  Год назад +2

      @@matteodeamicis alrightie, I think I'll work on it 👌

    • @yrior6221
      @yrior6221 Год назад

      This will be perfect. I tried it with langchain but only text answers and it was with a csv file.would love if we can get a tutorial on how to integrate it with this. Thank you!

  • @darkerdoll
    @darkerdoll Год назад

    Noice

  • @Systemm00007
    @Systemm00007 Год назад

    error is there that import speech recognition cannot be resolved

    • @Systemm00007
      @Systemm00007 Год назад

      help me please

    • @mikaelcodes
      @mikaelcodes  Год назад

      There can be problems in installing PyAudio for Windows. You can try the steps outlined here stackoverflow.com/questions/66641177/python-speech-recognition-is-not-working-using-speechrecognition-3-8-1-library

  • @snappesisnappes
    @snappesisnappes Год назад

    what theme you are using

    • @mikaelcodes
      @mikaelcodes  Год назад

      Cyberpunk Umbra protocol marketplace.visualstudio.com/items?itemName=max-SS.cyberpunk

  • @aamirm5117
    @aamirm5117 Год назад

    What is that dope ass vscode theme?

    • @mikaelcodes
      @mikaelcodes  Год назад

      Cyberpunk UMBRA protocol marketplace.visualstudio.com/items?itemName=max-SS.cyberpunk

  • @harshtiwari5853
    @harshtiwari5853 Год назад

    it can be done in any language that we want 🤔🤔🤔🤔

    • @mikaelcodes
      @mikaelcodes  Год назад

      Ultimately, yes. But check out the speech recognition and voice output options offered by Google, for example: cloud.google.com/speech-to-text/docs/speech-to-text-supported-languages

  • @ashishdhakal1612
    @ashishdhakal1612 Год назад

    Hello brother how can i get organization id ?

    • @mikaelcodes
      @mikaelcodes  Год назад

      platform.openai.com/account/org-settings You can go to the Settings page when you're looking at your OpenAI account. So be sure to create an account.

  • @wolfcompany2
    @wolfcompany2 Год назад +1

    what your visual studio theem

    • @mikaelcodes
      @mikaelcodes  Год назад +2

      Cyberpunk UMBRA protocol marketplace.visualstudio.com/items?itemName=max-SS.cyberpunk