mckaywrigley/wait-but-why-gpt

OpenAI Rate Limit

Opened this issue · 3 comments

Hi!
Thanks for posting about this project on twitter! I am trying to run your project. Scraping works fine, but when I run the embedding, I get the following error:
Rate limit reached for default-global-with-image-limits in organization <ORG_ID> on requests per min. Limit: 60 / min. Current: 70 / min. Contact support@openai.com if you continue to have issues. Please add a payment method to your account to increase your rate limit. Visit https://platform.openai.com/account/billing to add a payment method.

I assume that is simply a function of further rate limiting by OpenAI. If I figure out how to increase the time between requests, to avoid this... can I simply re-run or do I need to "delete existing embeddings" on the OpenAI side? Or is the script smart enough to not embed already embedded items?

Thanks!

Nevermind, after reading about embeddings I now understand how it works. And I am running it in batches to avoid the rate limit, should be done in a few hours. (even with setting it to 2 seconds I am still getting the error)

Did you just create your account? I know the rate limits greatly improve after 48hrs and after you’ve entered a credit card.

RPM should be 3,500/min, so it throwing an a error at >60/min is curious.

I had my account for a while, at least a month or two, but yes, I have not entered a credit card yet, trying to get by with the free version. But I guess the error messages suggests that simply adding a payment method will improve the rate limit... it doesn't say they will use it to charge for the rate limit increase. ;-)