Skip to content
This repository has been archived by the owner on Sep 9, 2020. It is now read-only.

Releases: datarobot/batch-scoring

v1.16.5

12 Dec 10:32
3df2b16
Compare
Choose a tag to compare
  • Remove redundant broken encoding check which based on bytes.

Clean up warnings and errors from previous runs

29 Aug 20:18
b8a6786
Compare
Choose a tag to compare
  • Errors and warnings from previous runs are now cleaned up and do now show up in subsequent runs
  • Bump flask version (only used for tests)
  • Fixd a typo in README

1.16.3 (2019 Aug 4)

04 Aug 18:18
2092ca6
Compare
Choose a tag to compare
  • Fixes the issue with unpredictable prediction explanations columns count in the output. From now on the total number of prediction explanations columns will always be set to the value of max_prediction_explanations command line argument.

v1.16.2

29 Jul 17:08
e2b66e0
Compare
Choose a tag to compare

1.16.2 (2019 Jul 29)

Bugfixes

  • Fix 'utf-8' encoding error for Japanese data sets.

v1.16.1

28 May 10:41
13da21e
Compare
Choose a tag to compare

1.16.1 (2019 May 27)

Enhancements

  • Set the username to lowercase internally.

v1.16.0

01 Apr 09:53
239e322
Compare
Choose a tag to compare

1.16.0 (2019 Mar 28)

Enhancements

  • Added new argument --pred_threshold that allows adding the prediction threshold (which is 0.5 by default, but can be changed during deployment) as a column.
  • Added new argument --pred_decision that allows adding the prediction decision (the value predicted by the model or the class label for classification) as a column.

v1.15.3

09 Jan 20:37
50847c9
Compare
Choose a tag to compare

1.15.3 (2019 Jan 7)

Bugfixes

  • Added support for --max_prediction_explanations on DataRobot 4.3.x .

v1.15.2

19 Dec 09:56
Compare
Choose a tag to compare

1.15.2 (2018 Dec 17)

Bugfixes

  • Fixed bug where retrieving the user's API token would fail when insecure SSL is required

v1.15.1

05 Dec 10:22
Compare
Choose a tag to compare

Enhancements

v1.15.0

26 Nov 09:02
Compare
Choose a tag to compare

Enhancements

  • Added new argument -max_prediction_explanations that allows batch scoring with predictions explanations and adds explanation_N_feature and explanation_N_strength to each row in output document (where N ∈ (1, max_prediction_explanations) )