how to configure ideal value for MaxConcurrentTransforms in setting up a sagemaker batch transform ?

Question

based on the documentation , https://docs.aws.amazon.com/sagemaker/latest/dg/batch-transform.html, it states that " The ideal value for MaxConcurrentTransforms is equal to the number of compute workers in the batch transform job." how to figure out what the number of compute workers is , i assume this depends on the instance type. also what about the instance count parameter we can set , do we have to take that into account as well?

Accepted Answer

The ideal value for MaxConcurrentTransforms varies based on instance type as well as based on your specific model.

it could make sense to increase MaxConcurrentTransforms up to the core count of the instance you are using (for cpu based transform), however, you should also take into account the **memory** utilisation by your model.

The ultimate answer is it "depends" and I would recommend that you experiment with increasing this number gradually from 1 up to instance core count, while monitoring RAM/cpu utilisation to find the optimal.

how to configure ideal value for MaxConcurrentTransforms in setting up a sagemaker batch transform ?

相关内容