It would be useful to have functions that can run a full workflow, such as label rendering or prediction aggregation. This would help streamline experimentation and avoid boilerplate. These could be based off the entrypoint scripts, and then used to update those entrypoints so it's easier to keep them current, as described in #95.
It would be useful to have functions that can run a full workflow, such as label rendering or prediction aggregation. This would help streamline experimentation and avoid boilerplate. These could be based off the
entrypointscripts, and then used to update those entrypoints so it's easier to keep them current, as described in #95.