Skip to content
This repository was archived by the owner on Jan 31, 2022. It is now read-only.

Clean up notebooks and rebuild fairing images.#54

Merged
k8s-ci-robot merged 1 commit intokubeflow:masterfrom
abcdefgs0324:clean-nb
Sep 13, 2019
Merged

Clean up notebooks and rebuild fairing images.#54
k8s-ci-robot merged 1 commit intokubeflow:masterfrom
abcdefgs0324:clean-nb

Conversation

@abcdefgs0324
Copy link
Copy Markdown
Member

@abcdefgs0324 abcdefgs0324 commented Sep 13, 2019 •
edited
Loading

  • Change the default values instead of the original basic settings, such as default size for hidden layers, learning rate strategy and max iteration, for classifiers. The original basic setting is too hard to have a look at the initial results because they are too "basic". They are still able to be specified in the pipeline if someone likes.
  • Update the repo configs in two notebooks, which was the changes from Clean up RepoConfig #43.
  • Clean up the notebooks to make them use the latest /py.
  • Rebuild the fairing image for the training pipeline.

/assign @jlewi
/assign @zhenghuiwang
/assign @hamelsmu

This change is Reviewable

@k8s-ci-robot k8s-ci-robot requested review from inc0 and jlewi September 13, 2019 01:14
@review-notebook-app
Copy link
Copy Markdown

review-notebook-app Bot commented Sep 13, 2019

Check out this pull request on  ReviewNB

You'll be able to see Jupyter notebook diff and discuss changes. Powered by ReviewNB.

@jlewi
Copy link
Copy Markdown
Contributor

jlewi commented Sep 13, 2019

@abcdefgs0324 It looks like you are making some model changes

  • Changing default size for hidden layers
  • Using "adaptive"

Can you explain these changes in the PR description?

@abcdefgs0324
Copy link
Copy Markdown
Member Author

abcdefgs0324 commented Sep 13, 2019 •
edited
Loading

@jlewi I revised the PR description, PTAL. Thanks!

@zhenghuiwang
Copy link
Copy Markdown
Contributor

zhenghuiwang commented Sep 13, 2019

/lgtm

Do the new default parameters yield models with better precision and recall?

@abcdefgs0324
Copy link
Copy Markdown
Member Author

abcdefgs0324 commented Sep 13, 2019

@zhenghuiwang Yes, I think so. The old default parameters are pretty bad on multiple datasets in my experiments.

@jlewi
Copy link
Copy Markdown
Contributor

jlewi commented Sep 13, 2019

/lgtm
/approve

@k8s-ci-robot
Copy link
Copy Markdown

k8s-ci-robot commented Sep 13, 2019

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: jlewi

The full list of commands accepted by this bot can be found here.

The pull request process is described here

Details Needs approval from an approver in each of these files:

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment

@k8s-ci-robot k8s-ci-robot merged commit 3b80598 into kubeflow:master Sep 13, 2019
@abcdefgs0324 abcdefgs0324 deleted the clean-nb branch September 13, 2019 19:51
Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.

Reviewers

@inc0 inc0 Awaiting requested review from inc0

@jlewi jlewi Awaiting requested review from jlewi

Assignees

@jlewi jlewi

@hamelsmu hamelsmu

@zhenghuiwang zhenghuiwang

Labels

approved cla: yes lgtm size/L

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

6 participants

@abcdefgs0324 @jlewi @zhenghuiwang @k8s-ci-robot @hamelsmu @googlebot