thewickedkarma / amadeus

A Telegram Chatbot written in Python3 making use of OpenAI's GPT3 model (Da Vinci Engine). You can fork and deploy this repo on heroku.

Amadeus

A Telegram Chatbot written in Python3, making use of OpenAI's GPT3 model (Da Vinci Engine). You can fork and deploy this repo on heroku.

Steps to get your python chatbot working

• Fork this repository
• Make your forked repository private in order to keep your openAI and Telegram Keys safe.
• Now edit the file 'amadeus.py' and replace the keys with the ones you had copied.
• Save the file
• Open heroku, register or login into your account. If you don't know how to do it, you can watch this video.
• Create a new application
• Go to settings, add new buildpack i.e python as we have created the chatbot in python
• Now, under deploy tab, link your github account and deploy the repo you just forked.
• And now coming to overview tab, configure the dyno and turn it on.
• That's it, your chatbot is ready. Enjoy.

amadeus's Issues

Message text is empty

when asked complex question, bot responses with this reply: "Message text is empty"
also sometime it doesnt print the whole response. for example if response has multiple paragraph it would only print the first line.

P.S.; i tried text-davinci-003 model but same problems

Error "You exceeded your current quota, please check your plan and billing details."

Not sure what's the error. The quote is not exceeded yet. Probably API request is outdated?

Not Getting multiline reply

When the API is replying with multiline text, the bot only showing first line.

error filter

Traceback (most recent call last):
File "/Users/ethicdeveloper/Desktop/ChatGPT-Telegram-Bot/harshitethic.py", line 1, in
from telegram.ext import Updater, CommandHandler, MessageHandler, Filters
ImportError: cannot import name 'Filters' from 'telegram.ext' (/Users/ethicdeveloper/Library/Python/3.11/lib/python/site-packages/telegram/ext/init.py)

"I'm sorry, I don't know how to help you with that."

I borrowed some of your code for my own TG bot (thanks for making this btw!!) to do some experiments with GPT-3. I'm wondering how I can adjust the parameters for temperature, frequency, etc. in the code itself

2 questions

1. i notice Amadeus often responds with "I'm sorry, I don't know how to help you with that" a lot more than inputting the same question into openai's playground. Do you know why this might be the case?

2. and this may solve my first question, but specifically I'm asking Amadeus a lot of data-related questions, so I'd like to adjust the temperature setting to be lower. Using the playground, I can grab this code but unsure where to include it in my local version of Amadeus:

response = openai.Completion.create(
  model="text-davinci-002",
  prompt="prompt input",
  temperature=0,
  max_tokens=100,
  top_p=1,
  frequency_penalty=0,
  presence_penalty=0
)

I think turning temperature way lower would help solve my issues, but curious if you've experienced this yourself