Yang

> Was running into this as well. Need to explicitly declare the provider in each module... Yes, I got it working by repeating the declaration in the root of the...

@MikhailKravets I'm also wondering the same.

https://github.com/SeldonIO/MLServer/issues/1134#issuecomment-1542381655 sounds like a possible fix on Kserve side.

@louis-huang can you please run the above test please?

related but not exactly the same: https://github.com/danielgtaylor/python-betterproto/issues/554

@tjohnson31415 Hey thanks for the reply. > To handle this load, the system probably needs to be scaled up. Do you have any suggestions? What configurations can I tweak? I...

@tjohnson31415 > How many different models are in active use? Can they all fit into the memory available in the Runtime or is there a lot of churn in the...

@tjohnson31415 No we don't have monitoring set up yet, will set it up soon, thanks for the suggestion. We've set up 4 replicas of the serving runtime pod, with each...

@MLHafizur scikit-learn from MLServer

@tjohnson31415 thanks so much for the in-depth explanations. We will do some testing and get back to you!