Hi,
Looking into PostgresML and really impressed by what I'm seeing with the examples.
My use cases are mainly to run regression with (many) scikit-learn models on tabular or TS data.
Do you see anything preventing me from building a custom docker image based on your Dockerfile but not based on Cuda ?

Starting from debian bookworm or ubuntu, installing pgml and pgml dashboard and forgoing cuda and pgvector for example

Thanks,