So it's not a conversation, every time it's just a prompt and a response, because the chat doesn't remember previous posts. So I added a few lines of code to upgrade this: " import openai openai.api_key = "paste your key" def chat_with_gpt(chat_log): response = openai.ChatCompletion.create(model='gpt-3.5-turbo', messages=chat_log ) return response.choices[0].message.content.strip() chat_log = [] # Remembering more posts is more expensive n_remembered_post = 2 if __name__ == "__main__": while True: user_input = input("You: ") if user_input.lower() in ['quit', "exit", "bye"]: break chat_log.append({'role': 'user', 'content': user_input}) if len(chat_log) > n_remembered_post: del chat_log[:len(chat_log)-n_remembered_post] response = chat_with_gpt(chat_log) print("Chatbot:", response) chat_log.append({'role': "assistant", 'content': response}) "
As beginner who has been struggling with different methods of having a user exit a program or loop, and having used various ways of doing so, I really loved the idea of defining options in a list like that. It might be super obvious now that I have seen it, but what a great little snippet of code - I am definitly bringing it with me going forwards.
@TechWithTim still worth mentioning. Especially with your last video on openai. Not everyone has that immediate opptunity. It's not fair for your audience to follow your tutorials only to be met with a pay wall because they weren't informed. Sure its cheap, but it only takes two seconds of your time to save others a lot of time.
Thank you very much for this video. It will make my work much easier. I am making a script that will take your voice, convert it to text, run it through ChatGPT, and then throw it into Text To Speech.
@STEFFMUSICOFFICIAL Uh, that was a whole ago, and the project kinda goofed up since I realized I have no knowledge in Python, until me and a friend got some motivation and used Llama api to create an AI discord bot. I honestly would recommend Llama if you don't mind rate limits, since it's absolutely free.
I've searched for stuff like this when I've discovered chatGPT and thinking that I can improve my virtual assistant. Now I can give my virtual assistant some new neurons to provide some information directly. Thanks a lot!!
Can I see your code in github? Or u can just say what functionalities u have added. Actually I have my own but it only uses if else statements with pyttsx3. I want to add much more features and make it a good project. So wanted to have some insight.
Hello, I have a problem because I have the key but it shows me that I have exceeded my current amount of available resources on OpenAI and I haven't even used it. Is it possible to have this without providing a credit card?
Really appreciate this video! The only problem I ran into is that the chatbot doesn't remember the context of the conversation. Its like creating a new chat every time you input something
You need to build up a list of responses in each direction and feed that to the openai api at each interation. But that would take more than five minutes....
Oh I have made my own chat both, I suggest doing some filtering(watch it on youtube) and divide the task into steps like Understand_userinput(user_input), Understanding = Understand_userinput(*User input from string*), Divide_into_steps(Understanding), Complete_tasks(You get it by now). For AI to remember get some SQL code from Github or RUclips and add a function like Send_to_database to feed into it. Then at each function feed it the database array. Or you can go complex and use a vector database.
@@TechWithTim just make a pinned comment and I’ll delete the comment, I didn’t expect you to see this lol it’s just that I’ve been going at this for the past 4 hours and gave up just 10 minutes ago
amazing video Tim, you always create the best and most useful python content online! Thank you so much! can you please make the same video but based on 2024. because now its not free, and you could be the first one to show us how to make this tech in 2024!!
I was considering designing a skill system using chat gpt to create random skills based on character actions. You could have chat gpt respond with .json style skills, titles, etc based on player actions and achievements.
Hello, I'm facing an issue where I can't import the OpenAI library. I have Python installed correctly and it's up-to-date. I also have the API key, but I'm unable to install the OpenAI library. Does anyone have a solution?
Why I get an exceeded quota limit if I have got just errors? Moreover it seemd ChatCompletion doesn't exist and I should turn back to openai version o.28. How I sort it?
Hi, in the meantime congratulations for the video, I wanted to ask you some questions for a project that I would like to realize. 1) I would like to create an AI always with python, but customize it, give it a name for example, 2) How could I make sure that he remembers the information I give him? For example, that they remember my name or my date of birth? 3) All this of course I would like to pass in a text to talk to talk to him, and turn him into a friend, Thank you in advance if you will answer me :)
Hey Tim, I wanted to make a version of this where the response would be read out bye a custom tts voice using pytts. What websites do you recommend using for creating this custom tts voice (based on manual training)?
I understand that there is a payment to be done in order to get the API key. Since my account was created maybe a year ago, my 3 months free ran out. So would I have to buy the $20 a month purchase, or is there another way to just buy the API keys itself so it works?
My openai account is more than 3 months old so when I create a new API key, it doesn't working saying quota exceeded :( I dont wanna create a new account ;-;
While it’s pretty cool to be able to do this, it’s easy enough to have a browser open and do it via there. Where I think this is useful, is if you pass things to gpt to summarise or do stuff with. Still got to be careful with what you feed any of these AI’s, but still being able to interact in python is useful.
Hi Tim, this tutorial has been really helpful and I wanted to make this myself too, can you give me the name of the software you are working on? Thank you.
hey Tim, knowing what you know about advancements in AI, would you still learn how to code if you wouldn't remember anything tomorrow, and had to start completely from the beginning again? i just need a yes or a no , I'd be so happy for your answer. thank you !
I just added this to my discord bot and ran my first prompt, it says that I am rate limited cause it exceeds my quota limit. is the api paywalled or am I missing something?
Correct me if I'm wrong, but this code probably doesn't use history, so every new request is a new request for GPT and GPT doesn't remember what you asked before. You need to save all history and send it with every request. Isn't it? I didn't use this library with python, I worked with GPT in PHP.
instead of response = openai.ChatCompletion.create use response = openai.completions.create OpenAi removed that on the newer update, I hope this helps.
how would you get it to output on a chatbot UI on a web page instead of the output going into the terminal so for example your using html,css and javascript for the front end
It is very simple to make it indeed. But it is not working for me. It is only giving me option to write by myself. "You : " is showing only and no response is generated by the chatbot.
Very good tutorial. I been trying to make it work with different versions of code but everytime I get the same error: " You exceeded your current quota, please check your plan and billing details" I have tried several models and keep getting the same error
Didn't work, and how are we supposed to make these work if we actually make a human or robot or something and depend on these API keys and then it no longer offers the service or something you're pretty much screwed at that point.
yes, for the API you have to pay. it's very low price based on how much you use it ($0.0015 / 1K tokens). for a quick calculation you can consider 1 word as approximately one token. if you stay under a few dollars they will not deduct anything from the card.
Thanks Tim! Can you make a video about pushing at chatbot like this to a unique website URL with Flask? (I have a domain with dreamhost and just can't seem to get it to work. Any tips would be greatly appreciated!)
You could use it for whatever you'd like. The first thing that came to my mind, was a support bot. Customize the prompt to correlate with a particular company and prevent specific texts, then it could be used to help end users with whatever it's set up for. For instance, IT chat bot would troubleshoot a user's hardware/software issues via text before routing to a person. We see this often on websites if you start a chat. Personally, I could use this for my own website that would help users understand that the data my site uses only supports content from the US, or why even some of that content is missing. This way, they could just ask the bot instead of sending me messages/emails. Maybe create a desktop application for yourself, so you don't have to open up the webpage anytime you want to communicate with it. Implement voice, so you can do talk to text. Maybe add a button that generates a random fact from it. haha. You could do whatever you'd like! AI is a great tool. Utilize it. The rate it's been growing over the past year, any and every developer should probably be working with AI. Many companies, small and large have already started implementation.
I’m planning on creating something just like this. But instead of the AI being able to give me information from all over the internet. I want to feed it a specific group of knowledge for a specific purpose. How can I do that? For example ( not what I’m doing ) creating an AI which explains a certain book in detail for example Harry Potter. How would you limit the AI to just Harry Potter and make sure it gives accurate data and not false, basically restricting the information input from the AI.
Does this still work for anyone? I'm trying to get it to work, but it says that I exceeded my current quota for API calls. Is this service only paid now?
exact error message: pip : The term 'pip' is not recognized as the name of a cmdlet, function, script file, or operable program. Check the spelling of the name, or if a path was included, verify that the path is correct and try again. At line:1 char:1 + pip install openai + ~~~ + CategoryInfo : ObjectNotFound: (pip:String) [], CommandNotFoundException + FullyQualifiedErrorId : CommandNotFoundException
Great, now I just need to turn this into a VOICE chatbot that can manipulate my code in real time, so I only need to think about the modifications/improvements instead of understanding the reasoning behind it
To learn programming and Python - check out Datacamp!
💻 Learn Python - datacamp.pxf.io/oqm3ao
💻 Learn Programming - datacamp.pxf.io/Jze0Y2
I didn't even imagine building a Chatbot, however simple it is, is so easy. Thanks Tim for opening up my mind.
This isnt building one this is important one
You are just importing a chatbot using API. You are not building it practically.
you're not building one , take an API as a bridge , so you're just connecting to the backend of OpenAI using the gpt 3.5 model
@@amanshaw7121 while making API you can add much more so that your chatbot is smarter than ChatGPT gives you
So it's not a conversation, every time it's just a prompt and a response, because the chat doesn't remember previous posts. So I added a few lines of code to upgrade this:
"
import openai
openai.api_key = "paste your key"
def chat_with_gpt(chat_log):
response = openai.ChatCompletion.create(model='gpt-3.5-turbo',
messages=chat_log
)
return response.choices[0].message.content.strip()
chat_log = []
# Remembering more posts is more expensive
n_remembered_post = 2
if __name__ == "__main__":
while True:
user_input = input("You: ")
if user_input.lower() in ['quit', "exit", "bye"]:
break
chat_log.append({'role': 'user', 'content': user_input})
if len(chat_log) > n_remembered_post:
del chat_log[:len(chat_log)-n_remembered_post]
response = chat_with_gpt(chat_log)
print("Chatbot:", response)
chat_log.append({'role': "assistant", 'content': response})
"
Love this! Would love to see more in-depth vids about function calling, deployment and chatbot use-cases for businesses. Keep it up!
As beginner who has been struggling with different methods of having a user exit a program or loop, and having used various ways of doing so, I really loved the idea of defining options in a list like that. It might be super obvious now that I have seen it, but what a great little snippet of code - I am definitly bringing it with me going forwards.
Thanks for breaking it down in under 4 minutes. This tutorial is incredibly helpful!
But this requires a paid plan. Probably worth mentioning.
Yes, but it costs fractions of a cent!
@TechWithTim still worth mentioning. Especially with your last video on openai. Not everyone has that immediate opptunity. It's not fair for your audience to follow your tutorials only to be met with a pay wall because they weren't informed. Sure its cheap, but it only takes two seconds of your time to save others a lot of time.
You're right, an oversight by me, I appreciate the feedback and will be more careful in future videos! @@maxfarouk3407
This is usually a nightmare for devs in 3rd world countries
@@TechWithTim Sweet man, Its not a huge deal.
Thank you very much for this video. It will make my work much easier. I am making a script that will take your voice, convert it to text, run it through ChatGPT, and then throw it into Text To Speech.
any update on this? i'm actually looking to do something like this. I'd be curious to hear how you've managed to accomplish this.
@STEFFMUSICOFFICIAL Uh, that was a whole ago, and the project kinda goofed up since I realized I have no knowledge in Python, until me and a friend got some motivation and used Llama api to create an AI discord bot. I honestly would recommend Llama if you don't mind rate limits, since it's absolutely free.
@@STEFFMUSICOFFICIALSame
Yooo dude this is sick 🎉
Thanks Tim for making everything easier for us
for Canada as well
@@msvcpdll3735 😂
I've searched for stuff like this when I've discovered chatGPT and thinking that I can improve my virtual assistant.
Now I can give my virtual assistant some new neurons to provide some information directly.
Thanks a lot!!
Can I see your code in github? Or u can just say what functionalities u have added. Actually I have my own but it only uses if else statements with pyttsx3. I want to add much more features and make it a good project. So wanted to have some insight.
This was awesome I would love to see some content about deploying the ai's on different platforms
Hello, I have a problem because I have the key but it shows me that I have exceeded my current amount of available resources on OpenAI and I haven't even used it. Is it possible to have this without providing a credit card?
you cant
Oh! So fast! :) Thanks for sharing!
Really appreciate this video! The only problem I ran into is that the chatbot doesn't remember the context of the conversation. Its like creating a new chat every time you input something
You need to build up a list of responses in each direction and feed that to the openai api at each interation. But that would take more than five minutes....
Oh I have made my own chat both, I suggest doing some filtering(watch it on youtube) and divide the task into steps like Understand_userinput(user_input), Understanding = Understand_userinput(*User input from string*), Divide_into_steps(Understanding), Complete_tasks(You get it by now). For AI to remember get some SQL code from Github or RUclips and add a function like Send_to_database to feed into it. Then at each function feed it the database array. Or you can go complex and use a vector database.
It helps me a lot! Thank you.
God bless you and your family Tim, you are a hero.
Tim upload more videos like this,really it is very easy to create a chatbot
This is brilliant, good work...
You make great content, thanks!
Its not working
Can't wait to see more tutorials about AI!!!🎉🎉
You gotta love it when only outdated youtube tutorials are your only source of doing something and the creator never says it outdated
Well at the time of filming it wasn’t outdated?
@@TechWithTim just make a pinned comment and I’ll delete the comment, I didn’t expect you to see this lol it’s just that I’ve been going at this for the past 4 hours and gave up just 10 minutes ago
@@TechWithTim this is pure bait. you never told us anything about the RateLimitError.
you are freakin awesome bruhh, thank u so much mate 🔥
amazing video Tim, you always create the best and most useful python content online! Thank you so much! can you please make the same video but based on 2024. because now its not free, and you could be the first one to show us how to make this tech in 2024!!
I was considering designing a skill system using chat gpt to create random skills based on character actions. You could have chat gpt respond with .json style skills, titles, etc based on player actions and achievements.
Check out my recent choose your own adventure game using ai, it does a similar thing!
This is great... Thanks man
Wow .. Excelent !!! . Thank you very much
kau sungguh hebat tim......good job
Hello, I'm facing an issue where I can't import the OpenAI library. I have Python installed correctly and it's up-to-date. I also have the API key, but I'm unable to install the OpenAI library. Does anyone have a solution?
Why I get an exceeded quota limit if I have got just errors? Moreover it seemd ChatCompletion doesn't exist and I should turn back to openai version o.28. How I sort it?
i have the same problem with the quota limit, did you fix it yet?
Simply explayned and superbe quality content, would love to see more
Can you please tell the openai version as it doesn't seem to be working for me on the latest version
Did you find an updated code that works with the current version?
@@YouKnowWhoIAm118 up, we have to use the old version ? openai==0.28 ?
Hi, in the meantime congratulations for the video, I wanted to ask you some questions for a project that I would like to realize. 1) I would like to create an AI always with python, but customize it, give it a name for example, 2) How could I make sure that he remembers the information I give him? For example, that they remember my name or my date of birth? 3) All this of course I would like to pass in a text to talk to talk to him, and turn him into a friend, Thank you in advance if you will answer me :)
Now there's an update, it's not working anymore 😢.
Can you please make a video based & working it work with the updates.
Did you figure it out?
@@connerrecoskie Facing the same issue. Did you figure it out ?
@@Zeus56552Yes
@@Zeus56552 Same here
same problem here too
Thank you. That was awesome. Can you help us with a video about training the AI for specific tasks?
Hey Tim, I wanted to make a version of this where the response would be read out bye a custom tts voice using pytts. What websites do you recommend using for creating this custom tts voice (based on manual training)?
I understand that there is a payment to be done in order to get the API key. Since my account was created maybe a year ago, my 3 months free ran out. So would I have to buy the $20 a month purchase, or is there another way to just buy the API keys itself so it works?
My openai account is more than 3 months old so when I create a new API key, it doesn't working saying quota exceeded :(
I dont wanna create a new account ;-;
Much love Tim
My chatbot throws an error stating there is no attribute called 'ChatCompletion'
mine too. did u find a solution?
While it’s pretty cool to be able to do this, it’s easy enough to have a browser open and do it via there. Where I think this is useful, is if you pass things to gpt to summarise or do stuff with. Still got to be careful with what you feed any of these AI’s, but still being able to interact in python is useful.
AttributeError: module 'openai' has no attribute 'ChatCompletion'
The error should link the documentation telling you to change 'ChatCompletion' to 'chat.completions'
this is the error i am getting right now . What should i do ?
You tried to access openai.Completion, but this is no longer supported in openai>=1.0.0
migrate to openai v1
This might have worked when vid was recorded. There's a safer way to avoid including the API KEY in the code. Anyway, thanks for sharing.
Can you share? I'm facing the same problem.
Hi Tim, this tutorial has been really helpful and I wanted to make this myself too, can you give me the name of the software you are working on? Thank you.
You did an amazing job, similar to this i have made a Ai powered Chatbot using Gemini Api 🔑, which you can customise for your own need. Its totally 🆓
As of April 12th - 2024' it say's these packages are no longer available.
hey Tim, knowing what you know about advancements in AI, would you still learn how to code if you wouldn't remember anything tomorrow, and had to start completely from the beginning again? i just need a yes or a no , I'd be so happy for your answer. thank you !
Hi I got an error like this -"import openai" could not be resolved pylance(reportmissingimports)[ln1,col8]
Pip install openai
I just added this to my discord bot and ran my first prompt, it says that I am rate limited cause it exceeds my quota limit.
is the api paywalled or am I missing something?
U can use it for free but u need to pay if you over the free usage - that was how it worked last time I used it
Correct me if I'm wrong, but this code probably doesn't use history, so every new request is a new request for GPT and GPT doesn't remember what you asked before. You need to save all history and send it with every request. Isn't it?
I didn't use this library with python, I worked with GPT in PHP.
i'm Still faeing openai=0.28 type error please help
instead of response = openai.ChatCompletion.create use response = openai.completions.create OpenAi removed that on the newer update, I hope this helps.
@@brk.aep7 it didnt work, can u instead give me a full script of what you made?
عالي و آموزنده
how would you get it to output on a chatbot UI on a web page instead of the output going into the terminal so for example your using html,css and javascript for the front end
It is very simple to make it indeed. But it is not working for me. It is only giving me option to write by myself. "You : " is showing only and no response is generated by the chatbot.
Thanks!
2:09 can someone explain how _name_ was executed bcz mine is giving name error
Undefined
not working it shows that chatcompletion is no more there according to open ai
Great video! I would love to see how to make this work on iOS or Android.
i wrote the same code but it keeps showing me the error
The model `gtp-3.5-turbo` does not exist any help plz
It's a spelling error bro. "gpt" not "gtp"
Thank you so much
Is there a way to give the api some rules so that the ai responds as if it were a predefined character ? Like a person with a specific name, age etc
Mine is so cool. when i talk to it it says error at line 20
its so cool
Very good tutorial. I been trying to make it work with different versions of code but everytime I get the same error:
" You exceeded your current quota, please check your plan and billing details" I have tried several models and keep getting the same error
It's not a free service. It's about $0.0015 per word I think.
same
Didn't work, and how are we supposed to make these work if we actually make a human or robot or something and depend on these API keys and then it no longer offers the service or something you're pretty much screwed at that point.
good video: super new to this, how did he do def in italics?
Thanks Tim.
Can u explain please how to do the same thing but with the Google Bard
I'm getting the error: You exeeded your current quota, please check your plan and billing details.
Do I need to pay for it to work?
yes, for the API you have to pay. it's very low price based on how much you use it ($0.0015 / 1K tokens). for a quick calculation you can consider 1 word as approximately one token. if you stay under a few dollars they will not deduct anything from the card.
Yes :(
Hello sir, my command prompt isn’t functioning correctly, even though my code matches yours. How can I achieve the same output as you? Help me !
Could you please make a video on using llama2 as well?
Thanks Tim! Can you make a video about pushing at chatbot like this to a unique website URL with Flask? (I have a domain with dreamhost and just can't seem to get it to work. Any tips would be greatly appreciated!)
Not a bad idea! Thanks :)
Can you connect this to a custom GPT instead of just 3.5 Turbo?
Very well done. Is there any way to implement this with one of the customized "myGPT's"?
what a day to watching this video perfect 1 year later
Use this command in the terminal box
"pip3 install openai"
how can i make it to have a defined prompt? like, alr act a certain way and have certain information
How do you got that python? My python is different and i cant do that.
can you run this offline like how dependent is is it when comes to the future like if open ai shuts down would mine that i made still work?
Is the first software you use your terminal or another IDE? Thanks for clarifying
another one, imo PyCharm
Chatbot: An error occurred: module 'openai' has no attribute 'ChatCompletion'
what should I do remove the error
python won't work on my Mac in viscose, idk why. when I click the run button. no code shows on screen. please help
well, seems that you don't know what are you doing at all, have you set environment variables? have you tested the Python IDE?
for some reasons it says file cant open , no such files or directories during last step
Look forward to more tiny videos
Not sure I get how to use this. Prompt the chatbot in your code to get some data that you would use in your code later?
You could use it for whatever you'd like. The first thing that came to my mind, was a support bot. Customize the prompt to correlate with a particular company and prevent specific texts, then it could be used to help end users with whatever it's set up for. For instance, IT chat bot would troubleshoot a user's hardware/software issues via text before routing to a person. We see this often on websites if you start a chat. Personally, I could use this for my own website that would help users understand that the data my site uses only supports content from the US, or why even some of that content is missing. This way, they could just ask the bot instead of sending me messages/emails.
Maybe create a desktop application for yourself, so you don't have to open up the webpage anytime you want to communicate with it. Implement voice, so you can do talk to text. Maybe add a button that generates a random fact from it. haha.
You could do whatever you'd like! AI is a great tool. Utilize it. The rate it's been growing over the past year, any and every developer should probably be working with AI. Many companies, small and large have already started implementation.
I’m planning on creating something just like this. But instead of the AI being able to give me information from all over the internet. I want to feed it a specific group of knowledge for a specific purpose. How can I do that? For example ( not what I’m doing ) creating an AI which explains a certain book in detail for example Harry Potter. How would you limit the AI to just Harry Potter and make sure it gives accurate data and not false, basically restricting the information input from the AI.
Does this still work for anyone? I'm trying to get it to work, but it says that I exceeded my current quota for API calls. Is this service only paid now?
exact error message:
pip : The term 'pip' is not recognized as the name of a cmdlet, function, script file, or operable program. Check the spelling of the name, or if a
path was included, verify that the path is correct and try again.
At line:1 char:1
+ pip install openai
+ ~~~
+ CategoryInfo : ObjectNotFound: (pip:String) [], CommandNotFoundException
+ FullyQualifiedErrorId : CommandNotFoundException
Hi Tim, really enjoy your videos but on a Mac with Python 3.12 getting 'openai' to install seems to be an issue. Any ideas ?
it didn't work idk why the cod is great but there is no result i tried to change the api key and it still not working :(
I got this message: "ModuleNotFoundError: No module named 'openai'". How to solve this? please. thanks
is there a way that i can use a custom model created with gpt creator and use it in this code?
Alas the way OpenAI API keys works changed in November 2023 and so the approach given in this video no longer works :(
Hi, i'm totally new to programming. How can i do this on google collab? Which platform should i use?
On giving user input, I keep getting the rate limit error. Please help
Do I need to add token to my account to make it works or is model gpt-3.5-turbo gives you a amount of free token to use to practice? Thanks!
Nevermind, I made it work by putting the api_key in the client = OpenAi(). Thanks!
Could you elaborate on this please @@marie-pierneault1035
Great, now I just need to turn this into a VOICE chatbot that can manipulate my code in real time, so I only need to think about the modifications/improvements instead of understanding the reasoning behind it
did not work for me
but my problem is with API
@@junenkingo1098same
hello, on mine it says invalid systax on line 8, can you please fix that for me