Hitting rate limits with Inference Providers

Inference Endpoints on the Hub
1

Hi

We are using HuggingFace Inference Providers (with Novita AI) and getting “code”:429, “reason”:“RATE_LIMIT_EXCEEDED” API responses.

We’re calling the API ~300 times per minute but would like to call it 400 times per minute. Is it possible to increase our rate limits?

Thanks

1 Like
2

“reason”:“RATE_LIMIT_EXCEEDED” API responses.

Based on the error message, it seems to be getting caught on an error on Novita’s side. In this case, it looks like the only option is to consult Novita’s support…

To raise your rate limit, contact support or use a verified account.

3

Thanks so much, I contacted them. It wasn’t clear though whether it was due to HuggingFace or Novita.

1 Like
4

There’s an Inference Providers channel on the HF Discord, so consulting there might be a good option. At the very least, your voice will reach the maintenance team.

5

Thanks @John6666 - You’re very helpful. I can’t manage to validate my account on Discord, it redirects me to the HF login page and says I’m not logged in (even though I am). I’ll send a direct message to lunarflu on Discord as you’ve advised in other threads.

1 Like