-
Notifications
You must be signed in to change notification settings - Fork 443
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
[GSOC] Project 8: Support various Parameter Distribution
in Katib
#2374
Comments
/area gsoc |
/reopen |
@tenzen-y: Reopened this issue. In response to this:
Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes-sigs/prow repository. |
/reopen |
@tenzen-y: Reopened this issue. In response to this:
Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository. |
/assign |
Parameter Distribution
in KatibParameter Distribution
in Katib
Motivation
Currently, Katib is limited to supporting only uniform distribution for integer, float, and categorical hyperparameters. By introducing additional distributions, Katib will become more flexible and powerful in conducting hyperparameter optimization tasks.
A Data Scientist requires Katib to support multiple hyperparameter distributions, such as log-uniform, normal, and log-normal, in addition to the existing uniform distribution. For instance, learning rates often benefit from a log-uniform distribution because small values can significantly impact performance. Similarly, normal distributions are useful for parameters that are expected to vary around a central value.
Goal
The project aims to enhance Katib's Experiment APIs by introducing support for various hyperparameter distributions beyond the current uniform distribution.
Currently, the Experiment APIs support
parameterType
andfeasibleSpace
. To provide greater flexibility, we expanded thefeasibleSpace
to include aDistribution
field, allowing users to specify the type of distribution for each parameter.This Distribution field is designed to be mutually exclusive, meaning it can be left empty, enabling users to tune hyperparameters without specifying a distribution if desired.
What I Did in GSoC Project & Ongoing Works
Preparation
Development
Distribution
field to theFeasibleSpace
object: [GSoC] AddedDistributionType
to Experiment API #2377Distribution
field: addedDistribution
field to feasibleSpace inapi.proto
#2397hyperopt
suggestion service: [GSOC]hyperopt
suggestion service logic update #2412optuna
suggestion servicegoptuna
suggestion serviceDocumentation and Ongoing Work
normal
andlognormal
distributions inhyperopt
: [GSOC]hyperopt
suggestion service logic update #2412Distribution
field.What I learnt during the project
One of the key hurdles was integrating multiple new distributions into the Katib Experiment APIs while ensuring backward compatibility. Modifying the
feasibleSpace
to include aDistribution
field required careful consideration of API design and how different suggestion services, like Optuna and Hyperopt, would handle these changes.Effective communication is key. The weekly syncs with @andreyvelich and @tenzen-y were extremely valuable, and having a clear agenda beforehand helped me get my questions answered efficiently and stay on track.
I learned the importance of end-to-end (E2E) tests. By implementing these tests, I gained valuable insights into how the different components of Katib interact and how critical it is to verify that new features, like the distribution support.
In the End
I am grateful to Google for the opportunity to work on this project and to the Kubeflow community for their support and guidance. Special thanks to my project mentors, @andreyvelich, @tenzen-y, and @terrytangyuan for your invaluable insights and encouragement. Your mentorship was pivotal in achieving the project's goals.
I hope the enhancements made will significantly benefit users and advance Katib’s capabilities in hyperparameter optimization. Looking forward to continue to contribute to Kubeflow.
The text was updated successfully, but these errors were encountered: