openai-python
openai-python copied to clipboard
Extend reqeust timeout
To fix the async API hang issue during message sending, allow setting a sock_read
timeout to prevent program unresponsiveness from incomplete transmissions.
I think with gpt-4, and the rest api running slower, if you send a big request - it's quite slow, and blows past the default 5 minute timeout.
I even think you need an async method, send a job, get a ticket, check the ticket. Or send a job, poll the queue.
much needed feature
This is resolved in v1, which was released earlier this week 🎉