Skip to content
Permalink
Branch: master
Commits on Mar 25, 2018
  1. Merge pull request #395 from seekingpeace/patch-1

    ClimbsRocks committed Mar 25, 2018
    Update README.md
Commits on Mar 6, 2018
  1. Merge pull request #389 from miyamamoto/patch-1

    ClimbsRocks committed Mar 6, 2018
    Update README.md
Commits on Feb 22, 2018
  1. v2.9.10 for training_features, and more flexibility in column_descrip…

    ClimbsRocks committed Feb 22, 2018
    …tions
  2. Usability (#388)

    ClimbsRocks committed Feb 22, 2018
    * adds .training_features to saved pipeline, more flexibility for column_descriptions input
    
    * bugfix in .training_features- thanks test suite!
Commits on Feb 9, 2018
  1. v2.9.9 for bugfix in analytics logging, and fixes to grid search edge…

    ClimbsRocks committed Feb 9, 2018
    … cases
  2. adds logging of version number to help with debugging

    ClimbsRocks committed Feb 9, 2018
  3. Bugfixes (#384)

    ClimbsRocks committed Feb 9, 2018
    * adds copy_reg back in
    
    * removes duplicate feature importance logging, sets refit=False, stops overwriting n_estimators with a much larger number
    
    * wraps print_results in try/except
    
    * removes max_bin from grid_search params. deprecation warning will still show when instantiating new instances with the sklearn interface
Commits on Feb 8, 2018
  1. Merge pull request #380 from ClimbsRocks/dev5

    ClimbsRocks committed Feb 8, 2018
    Dev5
  2. Merge branch 'master' into dev5

    ClimbsRocks committed Feb 8, 2018
  3. Travis2 (#381)

    ClimbsRocks committed Feb 8, 2018
    * trying different package versions for tensorflow and keras due to unannounced breaking changed to APIs
    
    * removes only-binary from install to see if that helps with install issues
    
    * two tests surpassed our best expected performance
    
    * removes unnecessary spacing
Commits on Jan 21, 2018
  1. v2.9.8 for copying prediction data

    ClimbsRocks committed Jan 21, 2018
  2. tries sudo:false to speed up builds

    ClimbsRocks committed Jan 21, 2018
  3. copies data for predictions always, particularly because auto_ml rese…

    ClimbsRocks committed Jan 21, 2018
    …ts indexes in DataFrames
  4. v2.9.7

    ClimbsRocks committed Jan 21, 2018
  5. Merge branch 'dev4'

    ClimbsRocks committed Jan 21, 2018
  6. only creates the pool if we are actually going to use it

    ClimbsRocks committed Jan 21, 2018
  7. makes load_ml_model more easily importable

    ClimbsRocks committed Jan 21, 2018
  8. v2.9.6 for .feature_importances_ and .versions on saved pipelines, lo…

    ClimbsRocks committed Jan 21, 2018
    …ad_ml_model importable from auto_ml, not auto_ml.utils_models, and improved garbage collection, particularly in jupyter notebooks
  9. Merge branch 'master' into dev4

    ClimbsRocks committed Jan 21, 2018
  10. adds some conditional logic to only add .feature_importances_ when we…

    ClimbsRocks committed Jan 21, 2018
    … can
  11. speeds up test suite

    ClimbsRocks committed Jan 21, 2018
  12. adds garbage collection, since jupyter doesn't always play nicely wit…

    ClimbsRocks committed Jan 21, 2018
    …h garbage collection
  13. adds feature_importances_ attribute to saved pipeline

    ClimbsRocks committed Jan 21, 2018
Commits on Jan 20, 2018
  1. temporarily adding in tests for py3.4 and py3.5 to verify this works …

    ClimbsRocks committed Jan 20, 2018
    …in those environments
Commits on Jan 18, 2018
  1. Merge pull request #375 from dakl/patch-1

    ClimbsRocks committed Jan 18, 2018
    fix formatting that hurts my eyes
  2. v2.9.5 for less parallelism, test suite fix to ignore breaking keras …

    ClimbsRocks committed Jan 18, 2018
    …change in 2.1.3, and handling of np numerical types in labelencoder
  3. uses a 2x faster earlier version of keras. sigh

    ClimbsRocks committed Jan 18, 2018
  4. more compatibility checks with input types of numpy

    ClimbsRocks committed Jan 18, 2018
  5. encodes np numbers as strings for label encoder

    ClimbsRocks committed Jan 18, 2018
Commits on Dec 15, 2017
  1. only creates the pool if we are actually going to use it

    ClimbsRocks committed Dec 15, 2017
Commits on Dec 8, 2017
  1. v2.9.4 for switching prediction intervals back to sklearn, prediction…

    ClimbsRocks committed Dec 8, 2017
    …_interval_params, and a big, matrix-based refactor of the test suite
Older
You can’t perform that action at this time.