father-bot / chatgpt_telegram_bot

💬 Telegram bot with ChatGPT, Python-based, using OpenAI's API.
https://t.me/chatgpt_karfly_bot
MIT License
5.18k stars 1.81k forks source link

Something went wrong during completion. Reason: #397

Closed Vanylo closed 12 months ago

Vanylo commented 12 months ago

Already reinstalled it several times. Nothing works. At first the bot writes a response to the request, but in the process it aborts. image

troyxmccall commented 12 months ago

This bot uses the Open AI API. When there are issues, it's always wise to check the Open AI dashboard and down-detector first before opening an issue:

issues

Vanylo commented 12 months ago

I don't think that's the problem. Because whenever I use it, I always get an error. What a month ago for two weeks and now nothing has changed. The bot only works well if the length of the response is small. Okay, let's see.

davletsh1n commented 12 months ago

image Same thing for the last couple of weeks

Vanylo commented 12 months ago

This bot uses the Open AI API. When there are issues, it's always wise to check the Open AI dashboard and down-detector first before opening an issue:

issues

All stats are now normal, but the bot keeps giving an error on a relatively long response

Wixej commented 12 months ago

Try changing the value in the file \bot\openai_utils "request_timeout": 100.0, I had the same interruptions before, changed the value and switched to a faster server. I don’t know which of these helped, try it. You can also try switching to another GPT 3.5 model. Register this gpt-3.5-turbo-1106

mithew commented 12 months ago

Try changing the value in the file \bot\openai_utils "request_timeout": 100.0, I had the same interruptions before, changed the value and switched to a faster server. I don’t know which of these helped, try it. You can also try switching to another GPT 3.5 model. Register this gpt-3.5-turbo-1106

Very useful

Vanylo commented 12 months ago

Try changing the value in the file \bot\openai_utils "request_timeout": 100.0, I had the same interruptions before, changed the value and switched to a faster server. I don’t know which of these helped, try it. You can also try switching to another GPT 3.5 model. Register this gpt-3.5-turbo-1106

Changed "request_timeout" to 150. Really helped, thanks