I have a pretty basic chatbot that uses OpenAI to respond to messages. I updated my call to openai.createChatCompletion and made the other necessary changes. Last night it was working great, rapidly answering questions and even providing much better code suggestions.
I woke up this morning to find every request since about 10PM EST last night resulted in a 400 error.
After some initial testing this morning, I decided to revert back to text-davinci-003 and everything is working as expected right now.
My question is, what could be causing these 400 errors on gpt3.5 that are working fine on davinci?
400 is not listed as an error code in OpenAI’s API documentation. But it’s a client error: the server got the request but this one is invalid. There are a ton of reasons why this might be happening to you. You’ll need to add logs to your application propagating the whole error traceback (with the request metadata) to get more meaningful insights on what might be going on. Without it, it’s not possible to help you further.
I encountered the similiar problem of ChatGPT. two hours ago had been working good for 9 hours. But now failed. Here is log from python
: ============================= RESTART: D:/DaoooTop/openai/firstchat.py ============================
An error occurred: Error communicating with OpenAI: HTTPSConnectionPool(host=‘api.openai.com’, port=443): Max retries exceeded with url: /v1/engines/gpt-3.5-turbo/completions (Caused by SSLError(SSLEOFError(8, ‘EOF occurred in violation of protocol (_ssl.c:1122)’)))
So I went back to my original code that I was working with last night and retested and it’s working fine now, now 400 errors. Wasn’t able to get it to produce a 400 error again, but I will submit logs the next time it happens.