How to Create Jarvis AI Assistant | Like Iron Man
ฝัง
- เผยแพร่เมื่อ 23 ก.ค. 2024
- Creating Artificial Intelligence Assistant | Jarvis | Friday | EDITH | Ironman
Creating Jarvis powered by OpenAI and Python | ChatGPT
pycharm = www.jetbrains.com/pycharm/dow...
home brew = brew.sh/
Find the code at = github.com/crackallcode/AI-As...
0:00 Setting Up
3:25 Importing Modules
4:56 Windows Install pyaudio
5:24 Mac Install pyaudio
7:08 Mac/Windows Install
7:57 Writing Code
15:11 Testing The Project
16:30 Complete Code
I was thinking about having a Jarvis this week!
Great content man! :)
A big hug from Brazil!
tbm sou do Brasil
this is not a personal ai, if you use ChatGPT or any API provider then you do not have your own AI, for example you can teach it how to do illegal things and let it give you back a prompt about it or let even make bypass simple things like NSFW filters. Your personal AI will learn any dataset you want and will provide you with any feedback you want, and you will be able to modify its training and algorithm.
But to achieve that, how much computational power is needed, how much investment in components?
@alrimvt02 if you are going to download pre-trained sets and alter them, then all you need is 16GB RAM because language models that are commercially available don't have long term memory, though I suggest 32GB for smoother work. If you are going to do the training from scratch and expect replies similar to what you see in GPT 3.5, then you will need the following:
If you're programming through Python, such as using pytorch, then you will need GPU since Python requires GPU, and the best would to buy an Nvidea RTX 4090 GPU currently costs around $2000 USD.
If you are using Php or C++, go with CPU such as Intel Core i9-13900KS, which costs around $1000 USD.
You could always use plugins that translate your code behind the scenes to different languages so you can use the power of either GPU or CPU, whichever you have on hand.
Also, you can use TPU to increase tensor performance for as low as $150 USD, especially if you are using Tensorflow.
To begin with deep learning and I mean just learn and test AI a $100 computer will do the work, all needed software will be downloaded and will work but won't be a smart AI only a learning AI.
For something that runs models such as GPT turbo and learns from the internet, a full option computer will cost as low as $4000, surely the faster you want it to work the higher the cost up to a $100,000 for computers running Nvidea A100 GPU, 512 GB RAM, and a Ryzen Threadripper processor.
Again, $4000 would still be great and give you all you need.
Lastly, keep in mind it's the learning process for the AI that needs resources, but after it learns then the AI would run on any cheap computer would run it just have between 16 to 32 GB RAM. While saying this, since we need the resources just for training the AI then we have an option to rent powerful computers that are made specifically to train your AI, those are found on clouds such as Google cloud and AWS Amazon cloud, and they are very cheap because you pay only for time your AI is being trained. Once the AI is trained, you can have it installed on your personal computer.
If you never worked with AI before it might take you between one month to three months to learn how to do every by yourself on your personal computer and develop your own AI. Just expect to use platforms and plugins that are already made, because to create something like Pytorch yourself that needs a community of specialists, and neither chatgpt have their own platform they just use what's made already.
You can run something like KoboldAI or Oogabooga. Running them purely from your home PC obviously won't be as powerful as something running of billion dollar computer like ChatGPT, but these other versions are perfectly capable of holding (mostly) coherent conversations.
Can you show me how to do that to my computer I'm willing to pay also??
that how elite block you from growing
You can simply determine ur mic with sr.Microphone().list_microphone_names(). You could also just write ur mic name into a var like mic_name = 'Mikrofon (NVIDIA Broadcast)' and set the mic index directly like mic_index = sr.Microphone().list_microphone_names().index(mic_name)
What if we want to connect it to a visual interface to optimize the conversation experience
FINALLY THIS ITS THE BEST ND ONLY VIDEO IT EXPLAIN WITH DETAILS ❤🎉🎉🎉THANK YOU ❤
Great video! I did never do anything with python, but this video is really easy to understand.
Have you experimented much with ElevenLabs in regards to this sort of thing? Wondering how easy it would be to combine. I'm trying to find an AI that'll let me set reminders and then specifically speak to me TO remind me, to jar me out of hyperfixation/help with memory issues.
Tell me if you find any news. I am seeking a similar AI tool that I can build that'll help me with my memory
If I were to do something like this I would probably use something like twilio or smtplib to send me a text or a email as a reminder... I would save my voice input in a variable and use os.system to create a file for the reminder... OR use an api to store my reminder and then then give a voice command (store it as a variable) for date and time to send me a text or email.
you can do that using the latest api of openai where you define the behavior and functionalities of the ai by writing an invisible system message. response = openai.ChatCompletion.create(
model="gpt-3.5-turbo",
messages=[
{"role": "system", "content": {"role": "system", "content": "You are a helpful assistant."},
{"role": "user", "content": user_input},
]
)
gpt_response = response['choices'][0]['message']['content'].strip()
Or try setting an alarm on your phone or PC
@@only1muppet that defeats the purpose of the ai
Any ideas on how to package this for deployment like a smart home speaker? Parts lists or anyone doing this commercially or simi commercially?
Can this be used on a phone with the app? I am wanting to make something like KITT on Knight Rider. I'll be using a Wireless mic connected under the hood (Similar to echo auto) to my phone so it can use the phones ChatGPT. Okay, so I currently have a 2016 Cadillac XTS that I have turned into Knight Riders KITT (Knightrider Cadillac on TikTok), I use echo auto but its so limited. I'd like to give it the ability to answer as something other then Alexa, and i'd like to give it some attitude. thx
Definition of quality content.❤
Edit:
Every single time i have no audio back, and the only response i get now is: "You exceeded your current quota, please check your plan and billing details."
Any ideas :) ?
You need a Open AI API billing account, you will need a CC and you can put $5 on the account to get it up and running.
yes, pay open ai.
That simply means it is not your own assistant that we are searching for😂
so now it wont work if i dont pay
I tried this, I was using a Windows PC
The first error I had was
"You exceeded your current quota, please check your plan and billing details."
I added my CC details to the API and gave them $5...
The next error I had was
"The specified device is not open or is not recognized by MCI"
To fix this I found the new version of playsound didn't work so removed it and installed 1.2.2
pip uninstall playsound
pip install playsound==1.2.2
Then I was able to hear the Voice.
Note that from asking the question to getting a response was about 30-60 seconds.
It does not do follow up questions
but you can keep asking questions at any time.
Until it stops answering.... and you have to reload the script.
Overall a fun learning experience
underrated comment. I also had to pay $5 :). And the current version of playsound is 1.3. Was fun indeed
i couldn't install playsound. but i fixed it with "pip install playsound==1.2.2". thanks dude
If you are having issues check out the troubleshooting portion of this video for "common issues" I see people having. th-cam.com/video/lCw7EGW5O-s/w-d-xo.html
Will the prompt continue to have context of previous parts of the conversation? For example if I tell it my favorite color is blue. And then ask it a few other things and then ask what my favorite color is. Is each prompt isolated from the rest? That would be a great next step. You'd have to keep track of the conversation in a string and send that along with the new prompt. And then also make sure to truncate that log so that it doesn't exceed the token limit.
Another mod would be to use the API to customize the personality of the bot.
You would need to set up the API for it to keep the context of the conversation...
@@ryan_phdsec do you simply set up the api to do that or do you have to submit the history as part of the prompt? Like is it just a switch to flip in the API?
The code is seemingly recognizing that I am speaking to it, but will not speak back and the terminal only says exception. Anyone else finding this error?
I am also having that issue, maybe its probably to do with the open ai part
Are you writing on kali linux or your host computer?
if you are using kali linux you may need to change the output settings manually
same here, Im on Mac... what could be the issue?
Same with me, I think the issue is with mac
At 12:52 in July 2024. I can't seem it find the Documentation Tab anywhere. Could anyone help?
it won't let me type anything in the terminal on pycharm, I'm using a pi4
00:00 Create an AI voice assistant using open AI
02:01 Learn how to create and use an API key
04:17 Installing Pi audio can be challenging
06:24 Install necessary packages for Pi audio and speech recognition
08:49 Creating a while loop and setting up microphone
10:58 Creating an always-listening AI program
13:01 Using OpenAI's GPT 3.5 model with Google speech to text and MP3 file creation
15:07 Learn how to carry on an audible conversation with chat GPT
Crafted by Merlin AI.
Any suggestion on a bot to be develop to keep learning from a cloud documente? For example, if the user prompt new questions and the admin join the chat to answer, this answer could be sync on real-time to the bot similar as ChatGPT learn from your inputs? The purpose of this is to create a more specialized virtual assistant to suggest answers for customers
Have u found anything yet lol I wanna add this too
@@tonyplayesgames5316 I’ve learned so far using botpress to create this knowledge base chatbot, I think is possible to create a variables to add this new information added by the user so the bot can use this information in the future with new customers and so on
hi, i tried following the steps and it only shows the file path in the console. i changed the (device index) part to 1, 2 and deleting it entirely. all gave the same result. i copy pasted your code and it still gave the same result. is this an error or do i just have bad pronunciation and it is working?
mas, colocando o API da openai do chatgpt, oque for fazer utilizando essa API não vai ficar armazenado no banco de dados da OpenAI não?
Thanks for this, I was able to get mine to work. Now I'm trying to add skills to it
yo is there been any improvements?
Can you help me with it.? mine is not working
Is there a way to make like an visual jarvis? Like the sphere, when it speaks it gets bigger and so on.
bhai badhiya
it was very helpfull
Ryan, its possible to use VSCode and use the Python extension to create this AI?
Yes. When originally created this it was on kali linux with vscode. However, it is possible you will have to install pip on your host computer.
when i run the code it only prints out what I say and "Exception"
My man! Can’t wait to try
Hi, so I have wanted to make a python AI assistant for quite some time now but I've never found a decent tutorial until this one I followed along with everything and was a bit disappointed when my code wasn't functioning as I had intended or in anyway as yours is but I thought it's okay as I am able to just copy your code directly but still it won't do anything it just keeps saying "Exception" I have no idea why
Check out the Trouble Shooting part of this video at the end th-cam.com/video/lCw7EGW5O-s/w-d-xo.html
what's the error?
Mine always writes what I say but I dont get answers and no sound can somebody help me please?
i honestly cant complete it due to not having pip is there a way to download pip
Am having an issue to install play sound import it gives a long error
Import speech rocognition as sr is showing module not found error
You are the best guy in whole TH-cam lot's and lot's of love from India ❤❤❤
Bhai bhai❤️✌️😂
@@ractor6175 tari ksm bhai aisa bandey miltay nahi hain bohot talented h bnda ❤️
can I get the output voice that way that someone who Im talking to would hear that answer too?
how do i turn it into an app or a feature on my device
Can we make the type of AI that can be really attached with my computer and my personal mobile phone , to use it in both of devices ( Android + Computer) that can assist me with my voice like Jarvis or FRIDAY ?
im using window but icant see intalled all of the import i ask
i can do all this with my visual studio code no?
my program is printing Exception what may be the error. please help
can we put Cortana voice int it?
Is there a way to change the voice of the ai at all? 🤔
Sooo many errors :( Didn't work and couldn't come close to get it working. no formatting errors copy and pasted from doc too and changed key .. nothing i do works :(
it says there is no module for import playsound and openai
Homebrew link is not working, parameter cannot be found that matches parameter name 'fsSL' and '/bin/bash' not recognized. Please help!
I have a question, can i do this with another kind of voice?, like, a more human like voice?
Its saying:
File "[The File name here]", line 38
SyntaxError: expected 'except' or 'finally' block
Process finished with exit code 1
what do I do in this case?
what if I want to make my own custom voice or other voice as a assistant what i have to do ?
Ryan, why she didn't talk back? i already right the code correctly and test it without api but with other alternative codes, she talked, but if i input the api she did not answer, it seems i need to pay? is that what she want? do i need to pay using cc to activate her talkback? please enlighten me
I get the error
AttributeError: module 'pkgutil' has no attribute 'ImpImporter'. Did you mean: 'zipimporter'?
while trying to install the pip libraries on windows, apreciate the help
Is there a way to download pip without using pycharm?
Can we run in the many devices
Can I add this code with the API into my esp32 and would it work?
I cant get playsound to download with pip!
where is the tamplate from the documentation
Bro are this code voice asistence can work like when we ask like 'open you tube and play this video' is this query work?
homebrew error when trying A parameter cannot be found that matches parameter name 'fsSL'.
At line:1 char:22 / The term '/bin/bash' is not recognized as the name of a cmdlet, function
i have a problem in creating Ai can you help me?
in pip install
can i build it into a website?
Hello, Thanks for making this video. I copied your code from github, installed the packages, and ran the code. I can start the program by saying the word "Friday" and I can see it transcribed in the output, but couple seconds after that I am getting an error as "Exception". I double checked my API key and checked my Open AI account to see if it has been accessed. It says that the API key is not accessed ("never"). How can I fix this?
Thanks!
This is going to sound really silly 🫣 But I have had this issue and I restarted my computer or Virtual Machine and it started working... If that does work I know a few people had to change Friday to something else... Just make sure to read the output of your word so you can make sure to change the if statement "if "Friday" in said:" to the new word.
@@ryan_phdsec Got the same error will try restarting thanks!
@@ryan_phdsec I restarted it but it didn't work I also changed Friday to Saturday it now gives me the second output that it sends to OpenAI but then just Exception
@@Celeste108butbetter check out the trouble shooting part of this video th-cam.com/video/lCw7EGW5O-s/w-d-xo.html
@@ryan_phdsec thank you
super confused looked over this video and notice that he makes corrections off screen but thought I had them all but mine doesnt respond to me it can hear me but has no reply
help?
i am using pycharm professional 2022.2 and i am getting an error "ModuleNotFoundError: No module named 'speech_recognition'" even though i installed the package using "pip install SpeechRecognition" same error is ocurring can u pls debug this
Hey everything work fine to me, but i cannot heat a single sound coming from it. Even delete mine and try your from Github and i can see it print message but I hear nothing. Any idea how to fix this ?
Hi, I am getting this error "def get_audio():
KeyboardInterrupt". Please help
how do i integrate this with a front end ive designed and make an actual project ?
Can you add skills to it, or os this just a “my parameters dont allow it” overly fancy bs way to use gpt on an alternative software window?
how can i turn this into an app so i can gift it to my nephew ?
what apllication is this
i keep getting the exception error but i have no idea whats the issue
hi there! Interesting video this seems a lot more straight forward than I thought. I got it to understand what I'm saying but I didn't hear anything playing back to me. Also wondering how we might be able to use a custom voice or something..
I have same issue :/. were you able to fix this?
Same issue. It understands me... types it out, but there is no response...
I was able to get it to make an MP3 file of it talking but can't actually hear the mp3 file
I think it would be nice if you try to use llm models from local pc by ollama
project name , do you replace whats already there? or do add your project name next to pythonProject?
can we do this using arduino? i mean, connect it with chatgpt and make a personal assistant ? maybe plug a small screen too?
i don't know why if i grab your code all the way to the pycharm and wen its run,but terminal shows expecting and if i say someting it did not speak back to me 😶
Im getting this error = def get_adio()
^
SyntaxError: invalid syntax
please help
Traceback (most recent call last):
File "C:\Users\Heerad\IdeaProjects\untitled3\venv\Scripts\activate_this.py", line 5, in
import playsound
ModuleNotFoundError: No module named 'playsound'
Process finished with exit code 1
this is my error HELP
pip install --upgrade setuptools wheel then reinstall playground
i tried to pip install playsound and did not install.
I need to know a thing, how can I chaing the voice? For example whit the voice of, Idk, Madara? Kratos? Obama?
So I'm not getting any audio to play out, just a message that says Exception: module 'openai' has no attribute 'ChatCompletion" what can I do to fix that?
In the terminal type “openai migrate”
Does anybody know how to make this a link?
Hi Ryan, great video and tutorial! I'm trying to run the code but get the below error. WHy is this package being called and how do I resolve the issue?
ImportError: cannot import name 'override' from 'typing_extensions' (C:\Users\foldername\anaconda3\lib\site-packages\typing_extensions.py)
So, I am ready with the code but when I run it, it just shows me the path to the File. Any idea why?
error on line 7 " import openai " no module named 'openai' and thats straight from your git hub,
pip install openai did the trick for me
Would there be a way to change the voice to something else other than the standard google TTS voice?
Not without using an external library.
There are some apis like elevan labs you could try
Is it possible to make the audio not from GTTS? a custom audio for example
Yeah, example with elevenlabs
I cannot install playsound, error: subprocess-exited-with-error. Please help!
pip install --upgrade setuptools wheel then reinstall playground
how can you give it a 3d body and face that lip syncs the speech perfectly and also has accurate body language gestures?
im not a 100% on this but you could likely use services VTubers use and connect it to the AI
Hi, I keep getting a yellow exclamation point when it says say = r.recognize_google(audio) and audio playback won't work, I still get a red exclamation point when playing audio. I need help
:(
I could also send a screen shot
I got mine to work so if u want I can send you the code ?
can you send it to me?@@Haha-yes22353
yes please :)
@@Haha-yes22353
I think this is a Great First Project, now would it be Possible to use Replika's API instead of ChatGPT?
And for all the Programmers out there, I am looking for an Interface I can Install on my Linux Desktop (ZORIN) That will allow me to Input a Picture for the Avatar, and it would make the Avatar Talk, and this Sits on the Desk Top and so I could Talk to my Computer STT/TTS and the Picture would talk back to me using Like Replika API instead of ChatGPT. Replika's Capabilities are much more Versatile than ChatGPT, ChatGPT has too many Restrictions on what you can and cannot say. Replika you can say anything you want. Also there are more and more AI Friends popping up on the Web, but none of them are making an Interface that Installs on your Computer and will actually Talk to you. Most of them are just Texting. So if someone was to Create an Interface that was Pre-Configured to Connect to Replika or any of the other AI Friends out there, I think it would be very Popular.
does anyone have the open ai documentation code for the if blank is in said part?
Can I use Bard instead? What would I need to change in the script? And can you maybe make a part 2 to this video?
bard is dumb, go with gpt
After install brew I could pip install pyaudio. I don't get any of this.
Use the following commands to get pyaudio to work with homebrew.
install homebrew: --> /bin/bash -c "$(curl -fsSL raw.githubusercontent.com/Homebrew/install/HEAD/install.sh)"
install portaudio: --> brew install portaudio
install pyaudio --> pip install pyaudio
It doesn't run but have a syntax error at line 16
wats da os?
Thank you so much for the tutorial! Im almost done with mine, but it says " import openai
ModuleNotFoundError: No module named 'openai'" when I try to run it, what should I do?
Go to the terminal and run import openai in it
If your using pycharms free community version there should the be package finder
Okay, so everything else works, but i can't install playsound. as it says error 1...
pip install --upgrade setuptools wheel then reinstall playground
how do i run it,
i did this on a steam deck lol
how do you done the last run of program
sir i am having an error of execption. for every input its showing exception
How to make a Website with this coding
can you just run it in normal python or in a terminal ?
homebrew key i have copied and paste in code. but is it not recognised is coming. error is comig. so plz tell me what should i do.