This message was deleted.

I'm using this conda configuration:

conda:
    environment_yml: "./environment.yml"

I think currently in BentoML, conda dependencies doesn’t have its own cache layer. Did you change any of the python packages between builds? Or does it always take 5 mins?

When is the conda environment unchanged everything is cached and the

containerize

takes few seconds. When is a part of the conda environment changed, all the packages are downloaded and installed.