Alex Wu

I suspect the issue here is related to explainability of the aws node provider's actions. When a request fails, it implicitly falls back to other AZs, then only reports the...

Would either a kahan_sum/fsum be a feature others think is worth implementing?

> My only concern is that this is not really an OpenAI-compatible API - but rather an OpenAI-compatible file format. I think this is fine, but perhaps we should make...

Example/docs are in. Let me know what you think.

🤦yes, thanks for catching that.

I think we're good here, Sang merged it already

cc @mfitton

I agree Closed by #44973

Yes, I tested this when creating the issue. What issues are you running into?

Don't have a strong opinion here. Just for historical context, IIRC the 10k actor limit and per-job cleanup were there to avoid leaking memory.