-
Notifications
You must be signed in to change notification settings - Fork 6.5k
[air] Add _max_cpu_fraction_per_node to ScalingConfig and documentation #26634
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
Conversation
tune.run(trainer.as_trainable(), num_samples=4) | ||
|
||
|
||
# TODO(ekl/sang) this currently fails. |
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.
@rkooo567 , it seems this fails since all CPUs end up excluded. Can we ensure at least 1 CPU is available on either side no matter how aggressive the fraction is?
Btw, I think we should disallow 1.0 and 0.0 as values (raise ValueError).
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.
Filed #26635
doc/source/data/key-concepts.rst
Outdated
``_max_cpu_fraction_per_node`` is experimental and not recommended for use with | ||
autoscaling clusters. |
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.
Should we be more explicit and say that the reason is that doing so may cause deadlock?
…on (ray-project#26634) Signed-off-by: Ubuntu <ubuntu@ip-172-31-32-136.us-west-2.compute.internal>
…on (ray-project#26634) Signed-off-by: Xiaowei Jiang <xwjiang2010@gmail.com>
…on (ray-project#26634) Signed-off-by: Stefan van der Kleij <s.vanderkleij@viroteq.com>
Why are these changes needed?
As a followup for #26397, add this to the docs and API as an experimental feature.