As the picture shows,
When I call the interface /v1/chat/completions, a socket hang up error occurs. This error happens more frequently when the model is set to gpt4. Also, this error is related to the request body. If the request body is not changed, this error will keep occurring.(temperature is 0)
Is there any solution for this issue?
Welcome to the forum, bcc.global.
Errors will occur with API calls of all kinds, it is the nature of interconnected systems. Part of your application must gracefully handle such errors, typically employing incremental backoff and retry systems to ensure timely recovery and prevention of swamping the API with calls while there is an outage.
Documentation can be found here : OpenAI Platform
