-
Notifications
You must be signed in to change notification settings - Fork 448
Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
Signed-off-by: Ce Gao <gaoce@caicloud.io>
- Loading branch information
Showing
1 changed file
with
9 additions
and
38 deletions.
There are no files selected for viewing
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -1,43 +1,14 @@ | ||
| # Katib 2020 Roadmap | ||
|
|
||
| This document provides a high level view of where Katib will grow in 2020. | ||
| ## New Features | ||
|
|
||
| The original Katib design document can be found [here](https://docs.google.com/document/d/1ZEKhou4z1utFTOgjzhSsnvysJFNEJmygllgDCBnYvm8/edit#heading=h.7fzqir88ovr). | ||
| - Support Early Stopping [#692](https://github.com/kubeflow/katib/issues/692) | ||
| - Support Advanced NAS Algorithms like DARTs, ProxylessNAS [#461](https://github.com/kubeflow/katib/issues/461) | ||
| - Support Auto Model Compression [#460](https://github.com/kubeflow/katib/issues/460) | ||
| - Support Auto Feature Engineering [#475](https://github.com/kubeflow/katib/issues/475) | ||
|
|
||
| # Katib 1.0 Readiness | ||
| ## Enhancements | ||
|
|
||
| * Stabilize APIs for Experiments | ||
| * Reconsider the design of Trial Template [#906](https://github.com/kubeflow/katib/issues/906) | ||
| * Early Stopping [#692](https://github.com/kubeflow/katib/issues/692) | ||
| * Resuming Experiment [#1061](https://github.com/kubeflow/katib/issues/1061), [#1062](https://github.com/kubeflow/katib/issues/1062) | ||
| * Fully integrate Katib with existing E2E examples: | ||
| * Xgboost | ||
| * Mnist | ||
| * GitHub issue summarization | ||
| * Publish API documentation, best practices, tutorials | ||
| * [Issues list](https://github.com/kubeflow/katib/issues) | ||
|
|
||
| # Enhance HP Tuning Experience | ||
|
|
||
| The objectives here are organized around the three stages defined in the CUJ: | ||
|
|
||
| ## 1. Defining Model and Parameters | ||
|
|
||
| Integration with KF distributed training components | ||
| * TFJob | ||
| * PyTorch | ||
| * Allow Katib to support other operator types generically [#341](https://github.com/kubeflow/katib/issues/341) | ||
|
|
||
| ## 2. Configuring a Experiment | ||
| * Supporting additional suggestion algorithms [#15](https://github.com/kubeflow/katib/issues/15) | ||
|
|
||
| ## 3. Tracking Model Performance | ||
| * UI enhancements: allowing data scientists to visualize results easier | ||
| * Support for persistent model and metadata storage | ||
| * Ideally users should be able to export and reuse trained models from a common storage | ||
|
|
||
| # Test and Release Infrastructure | ||
|
|
||
| * Improve e2e test coverage | ||
| * Improve test harness | ||
| * Enhance release process; adding automation (see https://bit.ly/2F7o4gM) | ||
| - Delete Suggestion deployment after Experiment is finished [#1061](https://github.com/kubeflow/katib/issues/1061) | ||
| - Save Suggestion state after deployment is deleted [#1062](https://github.com/kubeflow/katib/issues/1062) | ||
| - Reconsider the design of Trial Template [#906](https://github.com/kubeflow/katib/issues/906) |