-
Notifications
You must be signed in to change notification settings - Fork 341
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
cloud-runner is fused with base image #159
Conversation
Co-authored-by: Restyled.io <commits@restyled.io>
I like it a lot. See no problems on my end, once merged I'll update docs. |
What do you think about adding |
@andronovhopf py3 would be more difficult since people would tend to use the version they want. I mean python ...
pip install ... which version? I have always thought that the runner is unfortunately not well resolved from a GH point of view. A GH user would maybe use the python action. GPU is super doable we would just end up with a fatter image. What do you think @andronovhopf and @dmpetrov ? I think that this way it would less confusing, with just py3 we would have only:
|
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Let's go for it!
I see some disadvantages - we hide a bit too much under our dockers. But let's prioritize the simplicity.
@dmpetrov only runner right? no gpu yet as far I understand |
Having too many images makes a bit confusing the message, thats specially true when you have to setup a self-hosted runner where you specify the image in the yaml and launch another image with docker.
To simplify it a bit I proposee to merge the CML cloud runner into the base image being
dvcorg/cml
equal todvcorg/cml-cloud-runner
dvcorg/cml-gpu
equal todvcorg/cml-gpu-cloud-runner
dvcorg/cml-py3
equal todvcorg/cml-py3-cloud-runner
dvcorg/cml-gpu-py3
equal todvcorg/cml-gpu-py3-cloud-runner
we still generate the images to as tags to not break workflows that may now be using the
-cloud-runner
tag. In the near future we can deprecate them totally.@andronovhopf this PR implies also update the docs in cml.dev later on.
I wish we could do the same with py3 and gpu somehow.