How severe does this issue affect your experience of using Ray?
- Low: It annoys or frustrates me for a moment.
I am new to Ray but I have been able to extend the official Docker Ray image in the past to install my own dependencies and create a RayJob with the custom image.
Now, I am thinking about using a dependency that requires its own base image, and my understanding is that I would need to install Ray myself in that base image.
Has anyone gone about doing that in the past? I know the official Ray image uses conda to install Ray… would my case be as simple as exporting that conda env from the official image and loading it into my own, or is it more complicated than that?