All Versions
188
Latest Version
Avg Release Cycle
13 days
Latest Release
-

Changelog History
Page 1

  • v3.38.0.3 Changes

    πŸš€ Download at: http://h2o-release.s3.amazonaws.com/h2o/rel-zygmund/3/index.html

    πŸ› Bug

    • [PUBDEV-8909] - Enforced DkvClassLoader while accessing Python resources through JythonCFuncLoader.
    • [PUBDEV-8908] - Closed open file descriptors from H2OConnection.
    • πŸ›  [PUBDEV-8903] - Fixed incorrect value indicator for a partial dependence plot for its current row.
    • πŸ— [PUBDEV-8901] - Fixed GBM model with interaction_constraints only building single-depth trees.
    • πŸ›  [PUBDEV-8876] - Fixed slow estimator validation when training model with wide datasets.
    • πŸ›  [PUBDEV-8865] - Fixed GAM failure when numknots=2 for I-spline.

    Task

    • [PUBDEV-8877] - Ensured non-negative will not overwrite splines_non_negative for GAM I-spline.
    • [PUBDEV-8843] - Implemented p-value calculation for GLM with regularization.
    • πŸ‘ [PUBDEV-8836] - Verified the minimum number of knots each spline type can support for GAM.
    • πŸ‘ [PUBDEV-8835] - Implemented normal (non-monotonic) splines that can support any degrees.

    πŸ“„ Docs

    • [PUBDEV-8912] - Updated compute_p_value documentation for GLM and GAM to reflect that p-values and z-values can now be computed with regularization.
    • [PUBDEV-8911] - Documented GAM M-splines.
    • ⚑️ [PUBDEV-8910] - Updated site logo, favicon, and color scheme to reflect H2O’s brand kit.
    • ⚑️ [PUBDEV-8904] - Updated booklet links for GBM, GLM, and Deep Learning on their respective algorithm pages.
    • [PUBDEV-8897] - Fixed typo in Model Selection for build_glm_model parameter.
    • ⚑️ [PUBDEV-8892] - Updated links in the provided bibliography in the FAQ.
    • 🚚 [PUBDEV-8879] - Removed Sparkling Water booklet link from the download page.
    • [PUBDEV-8869] - Added optional Python plotting requirement matplotlib to the download page.
  • v3.38.0.2 Changes

    πŸš€ Download at: http://h2o-release.s3.amazonaws.com/h2o/rel-zygmund/2/index.html

    πŸ› Bug

    • πŸ›  [PUBDEV-8878] - Fixed H2ODeepLearningEstimator autoencoder not working without y value.
    • 🐳 [PUBDEV-8861] - Added libgomp into docker images thus enabling XGBoost multithreading.
    • ⚠ [PUBDEV-8846] - Stopped throwing warning about jobs not having proper model types when models weren’t even trained.
    • πŸ›  [PUBDEV-8833] - Fixed cross validation failure for concurrent sorting.
    • [PUBDEV-8830] - Enabled parallelism in cross validation for Isotonic Regression.

    Task

    • πŸ‘ [PUBDEV-8849] - Enabled GAM I-spline to support increasing and decreasing functions.
    • ⚑️ [PUBDEV-8836] - Updated the number of knots required for GAM I-splines to be >=2.
    • [PUBDEV-8785] - Improved ModelSelection’s mode=β€œmaxrsweep” runtime.

    πŸ“„ Docs

    • [PUBDEV-8883] - Added information on ModelSelection’s new build_glm_model parameter for mode=β€œmaxrsweep”.
    • πŸ›  [PUBDEV-8870] - Fixed incorrect header case on ModelSelection and Cox Proportional Hazards algorithm pages in the user guide.
    • [PUBDEV-8854] - Added an example to Variable Inflation Factors in the user guide.
    • πŸ›  [PUBDEV-8849] - Fixed broken links on the β€œWelcome to H2O-3” page of the user guide.
    • 🐎 [PUBDEV-8786] - Added model explainability for plotting SHAP to the β€œPerformance and Prediction” page of the user guide.
    • πŸ“š [PUBDEV-8209] - Added examples for varsplits() and feature_frequencies() to Python documentation.

    πŸ”’ Security

    • πŸ”’ [PUBDEV-8885] - Addressed CVE-2022-42003 and CVE-2022-42889 security issues through Library upgrades.
  • v3.38.0.1 Changes

    πŸš€ Download at: http://h2o-release.s3.amazonaws.com/h2o/rel-zygmund/1/index.html

    πŸ› Bug

    • πŸ›  [PUBDEV-8811] - Fixed the sorting of h2o.make_leaderboard.
    • πŸ›  [PUBDEV-8802] - Fixed H2O dependencies overriding Jetty implementation.
    • πŸ›  [PUBDEV-8783] - Fixed Flow’s export Frame throwing an NPE because it doesn’t provide a file type.
    • πŸ›° [PUBDEV-8765] - Fixed GLM ordinal generic metrics to provide missing information in the payload.
    • πŸ›  [PUBDEV-8763] - Fixed β€œmaxrsweep” NPE in ModelSelection thrown when the replacement step stopped too early.
    • ⚑️ [PUBDEV-8758] - Fixed β€œmaxrsweep” replacement bug in ModelSelection by updating the implementation method.
    • 🐎 [PUBDEV-8742] - Fixed unnecessary transformations in the scikit-learn wrapper by using model performance API.
    • πŸš€ [PUBDEV-8733] - Fixed upload of big files in Sparkling Water deployment.
    • πŸ‘ [PUBDEV-8726] - Changed the error message that GLM does not support contributions.
    • πŸ›  [PUBDEV-8722] - Fixed QuantilesGlobal histogram type failing in GBM when all columns were categorial.
    • ⚑️ [PUBDEV-8698] - Added support for MapR 6.2 to fix the error caused by updating the cluster.
    • πŸ›  [PUBDEV-8693] - Fixed large file upload in Python.
    • πŸ–¨ [PUBDEV-8670] - Fixed inability to stop print out of model information in Python.
    • πŸ‘€ [PUBDEV-8621] - Removed -seed variable hiding in GAM.
    • ⚑️ [PUBDEV-8561] - Updated h2o.upload_mojo to also work for POJO.
    • [PUBDEV-8220] - Added unsupported operation exception when trying to use SHAP summary plot when building DRF model with binomial_double_trees.
    • πŸ”¨ [PUBDEV-7098] - Refactored the rendering logic in the Python client.
    • [PUBDEV-3524] - Added xval argument to h2o.confusionMatrix in R.

    πŸ‘Œ Improvement

    • πŸ‘ [PUBDEV-8825] - Added support for calibrating an already trained model manually.
    • πŸ‘ [PUBDEV-8801] - Added support for using Isotonic Regression for model calibration.
    • [PUBDEV-8795] - Added ability to S3A allowing it to share the built-in AWS credential providers.
    • [PUBDEV-8787] - Improved configure_s3_using_s3a allowing it to be usable in any deployment.
    • ⚑️ [PUBDEV-8727] - Updated train_segments function in R to be independent of camel casing in the algorithm name.
    • [PUBDEV-8721] - Improved runtime for QuantilesGlobal histogram by using exact split-points for low-cardinality columns.
    • [PUBDEV-8712] - Exposed the Sequential Walker for R/Python and added option to disable early stopping.
    • [PUBDEV-8692] - Cleaned up Key API by removing replicas.
    • 0️⃣ [PUBDEV-8315] - Cleaned up the default output after training a model.
    • [PUBDEV-8140] - Exposed calibrated probabilities in mojo_predict_pandas.

    πŸ†• New Feature

    • πŸ”§ [PUBDEV-8784] - Simplified the configuration of S3 for Frame exportation.
    • βœ… [PUBDEV-8725] - Added train_segments test for Isolation Forest.
    • [PUBDEV-8713] - Added ability to h2o.no_progress in R allowing it to accept expressions.
    • [PUBDEV-8685] - Implemented dispersion parameter estimation for GLM.
    • [PUBDEV-8679] - Added ability to export H2O Frame to a Parquet.
    • [PUBDEV-8589] - Added Pareto front plots to AutoML Explain.
    • [PUBDEV-8574] - Added β€œdeviance” method to dispersion for calculating p-values.
    • [PUBDEV-8572] - Implemented variable inflation factors for GLM.
    • [PUBDEV-8470] - Implemented in-training checkpoints for GBM.
    • πŸ‘ [PUBDEV-7634] - Implemented support for interactions to MOJO for CoxPH.
    • [PUBDEV-5280] - Added h2o.make_leaderboard function which scores and compares a set of models to AutoML.

    Task

    • [PUBDEV-8834] - Secured XGBoost connections in multinode environments.
    • [PUBDEV-8786] - Added missing added predictor and deleted predictor to the result frame and model summary of ModelSelection.
    • πŸ— [PUBDEV-8775] - Added support allowing you to force GLM to build a null model where the model only returns the coefficients for the intercept.
    • πŸ‘ [PUBDEV-8774] - Added support allowing GLM gamma to fix the dispersion parameter to calculate p-values.
    • [PUBDEV-8703] - Implemented β€œmaxr” speedup for Modelselection by introducting β€œmaxrsweep”.
    • [PUBDEV-8683] - Implemented dispersion factor estimation using maximum likelihood for GLM gamma family.

    πŸ“„ Docs

    • πŸ“š [PUBDEV-8845] - Added documentation on Isotonic Regression.
    • [PUBDEV-8829] - Added variable inflation factors to GLM section of the user guide.
    • [PUBDEV-8827] - Added Tweedie dispersion parameter estimation to the GLM section of the user guide.
    • 🐎 [PUBDEV-8817] - Added confusion matrix calculation explanation to performance and prediction.
    • [PUBDEV-8806] - Added get_predictors_removed_per_step() and get_predictors_added_per_step() examples to ModelSelection.
    • [PUBDEV-8790] - Added use case section to the welcome page of the user guide.
    • [PUBDEV-8720] - Added MOJO import/export information to each algorithm page.
    • πŸš€ [PUBDEV-8632] - Added major release blogs to user guide and moved change log to top of the sidebar.
  • v3.36.1.5 Changes

    πŸš€ Download at: http://h2o-release.s3.amazonaws.com/h2o/rel-zumbo/5/index.html

    πŸ”’ Security

    • βž• Addressed security vulnerability CVE-2021-22569 in the h2o.jar.
  • v3.36.1.4 Changes

    πŸš€ Download at: http://h2o-release.s3.amazonaws.com/h2o/rel-zumbo/4/index.html

    πŸ› Bug

    • [PUBDEV-8773] - Disabled partial_plot for Uplift DRF temporarily.
    • πŸ‘ [PUBDEV-8766] - Added support for predicting with Autoencoder when using eigen encoding.
    • πŸ›  [PUBDEV-8757] - Fixed XGBoost failure with enabled cross validation in external cluster mode by explicitly starting external XGBoost before cross validation.

    πŸ”’ Security

    • πŸ”’ [PUBDEV-8788] - Addressed security vulnerabilities CVE-2021-22573 and CVE-2019-10172 in Steam assembly.
  • v3.36.1.3 Changes

    πŸš€ Download at: http://h2o-release.s3.amazonaws.com/h2o/rel-zumbo/3/index.html

    πŸ› Bug

    • πŸ›  [PUBDEV-8738] - Fixed CoxPH MOJO ignoring offset column.
    • πŸ›  [PUBDEV-8737] - Fixed the incorrect predictions from the CoxPH MOJO on categorical columns.
    • [PUBDEV-8736] - Fixed the View button not working after completing an AutoML job.
    • [PUBDEV-8718] - Fixed num_of_features not being used in call for varimp_heatmap().
    • πŸ›  [PUBDEV-8681] - Fixed GAM’s fold_column being treated as a normal column to score for.
    • ⚑️ [PUBDEV-8625] - Updated GBM cross validation model summary tables to reflect that some trees are removed due to a better score occurring with fewer trees.
    • πŸ›  [PUBDEV-8522] - Fixed fit_params passthrough for scikit-learn compatibility.
    • 0️⃣ [PUBDEV-7927] - Fixed validateWithCheckpoint to work with default parameter settings.
    • πŸ›  [PUBDEV-7926] - Fixed validateWithCheckpoint to work with parameters that are arrays.

    πŸ‘Œ Improvement

    • [PUBDEV-8716] - Added expert option to force-enable MOJO for CoxPH even when interactions are enabled.
    • [PUBDEV-8616] - Makes language rules generation on demand and introduced β€œEnumLimited” option for categorical encoding.

    πŸ†• New Feature

    • [PUBDEV-8750] - Added transform_frame for GLRM allowing users to obtain the new X for a new data set.
    • πŸ‘ [PUBDEV-8739] - Added support for numerical interactions in CoxPH MOJO.

    πŸ“„ Docs

    • πŸ“š [PUBDEV-8753] - Fixed the uplift_metric documentation for Uplift DRF.
    • πŸ“š [PUBDEV-8751] - Added transform_frame to GLRM documentation.
    • πŸ“š [PUBDEV-8746] - Added mode = β€œmaxrsweep” to ModelSelection documentation.
    • πŸ“š [PUBDEV-8744] - Corrected the R documentation on R2.
    • ⚑️ [PUBDEV-8719] - Updated supported MOJO list to include GAM MOJO import.

    πŸ”’ Security

    • πŸ”’ [PUBDEV-8735] - Fixed security issue in genmodel (CVE-2022-25647).
  • v3.36.1.2 Changes

    πŸš€ Download at: http://h2o-release.s3.amazonaws.com/h2o/rel-zumbo/2/index.html

    πŸ› Bug

    • πŸ”¨ [PUBDEV-8702] - Refactored Uplift DRF methods.
    • 🚚 [PUBDEV-8699] - Removed duplicate runs in MaxR.
    • πŸ›  [PUBDEV-8682] - Fixed the ambiguity check in Explain’s consolidate varimp.
    • [PUBDEV-8673] - Improved efficiency of pd_plot and ice_plot and made rug optional.
    • πŸ›  [PUBDEV-8672] - Fixed H2O failing with null pointer exception when providing an improper -network to h2o.jar.
    • πŸ›  [PUBDEV-8671] - Fixed external XGBoost on K8s.
    • πŸ›  [PUBDEV-8640] - Fixed failing concurrent-running GLM training processes.
    • πŸ›  [PUBDEV-8375] - Fixed missing time values SHAP summary plot error.
    • πŸ›  [PUBDEV-7913] - Fixed Partial Dependence Plot’s date/time handling from explainability modules.

    πŸ†• New Feature

    πŸ“„ Docs

    • [PUBDEV-8700] - Added model_summary examples for GLM.
    • ⚑️ [PUBDEV-8687] - Updated incorrect formula in GLM booklet.
    • πŸ“š [PUBDEV-8442] - Updated Python Module documentation readability.
  • v3.36.1.1 Changes

    πŸš€ Download at: http://h2o-release.s3.amazonaws.com/h2o/rel-zumbo/1/index.html

    πŸ› Bug

    • πŸ›  [PUBDEV-8653] - Fixed Residual Analysis plot flipping the residual calculation.
    • πŸ‘» [PUBDEV-8643] - Added more detailed exception when disconnected due to error caused by Rcurl.
    • πŸ“¦ [PUBDEV-8633] - Made R client attempt to connect to curl package instead of Rcurl package first.
    • [PUBDEV-8622] - Ensures GLM models fail instead of throwing warnings when beta_contraints and `non_negative are used with multinomial or ordinal families.
    • πŸ›  [PUBDEV-8620] - Fixed how cv_computeAndSetOptimalParameters deals with multiple alpha and lambda values across different folds.
    • [PUBDEV-8619] - Increased MaxR running speed.
    • πŸ›  [PUBDEV-8602] - Fixed getGLMRegularizationPath erroring out when standardize = False.
    • πŸ›  [PUBDEV-8468] - Fixed Keystore not generating on Java 16+.
    • [PUBDEV-8014] - Added a num_of_features argument to h2o.varimp_heatmap to limit the number of displayed variables.
    • [PUBDEV-5258] - Fixed cross_validation_metrics_summary not being accessible for Stacked Ensemble.

    πŸ‘Œ Improvement

    • [PUBDEV-8661] - Improved AUUC result information in Uplift DRF by adding information on number of bins.
    • πŸ“¦ [PUBDEV-8649] - Replaced class() with inherits() in R package.
    • πŸ“¦ [PUBDEV-8644] - Fixed invalid URLs in R Package.
    • [PUBDEV-8629] - Added normalized AUUC to Uplift DRF.
    • [PUBDEV-8617] - Sped-up AutoML by avoiding sleep-waiting.
    • 🚚 [PUBDEV-8614] - Removed Stacked Ensembles with XGB metalearner to increase speed.
    • [PUBDEV-8613] - Ensures AutoML reproducibility when max_models is used.
    • ⚑️ [PUBDEV-8528] - Updated AutoML default leaderboard regression sorting to RMSE.

    πŸ†• New Feature

    • [PUBDEV-8652] - Bundled several basic datasets with H2O for use in examples.
    • πŸš€ [PUBDEV-8630] - Added h2o.jar assembly for secure Steam deployments and excluded PAM authentication from minimal/Steam builds.
    • [PUBDEV-8624] - Added ability to ingest data from secured Hive using h2odriver.jar in standalone.
    • [PUBDEV-8592] - Bundled KrbStandalone extension in h2odriver.jar.
    • 0️⃣ [PUBDEV-8580] - Implemented new method for defining histogram split-points in GBM/DRF designed to address outlier issues with default UniformAdaptive method.
    • [PUBDEV-8573] - Added ability to reorder frame levels based on their frequencies and to relieve only topN levels for GLM.
    • [PUBDEV-8571] - Added a function to calculate predicted versus actual response in GLM.
    • [PUBDEV-8401] - Implemented MOJO for Extended Isolation Forest.
    • [PUBDEV-8398] - Added monotone splines to GAM.
    • [PUBDEV-8388] - Added a plot function for gains/lift to R and Python.
    • ⚑️ [PUBDEV-8373] - Added ability to acquire metric builder updates for Sparkling Water calculation without H2O runtime.
    • πŸ‘ [PUBDEV-7980] - Added support for interaction_constraints to GBM.
    • [PUBDEV-7858] - Exposed distribution parameter in AutoML

    Task

    • [PUBDEV-8587] - Decoupled Infogram and XGBoost removing Infograms reliance on XGBoost to work.
    • [PUBDEV-8585] - Verified GLM binomial IRLSM implementation and p-value calculation.
    • [PUBDEV-8576] - Added private ModelBuilder parameter to AutoML to enforce the time budget on the final model after cross-validation.

    Sub-Task

    • [PUBDEV-8601] - Made Ice plot functionalities also available on pd_plot.
    • [PUBDEV-8502] - Added option to normalize y-axis values.
    • [PUBDEV-8499] - Added option to display logodds for binary models for Ice plots.
    • [PUBDEV-8498] - Added ability to save final graphing data to a frame for Ice plots.
    • [PUBDEV-8497] - Added option to specify a grouping variable for Ice plots.
    • [PUBDEV-8496] - Shows original observation values as points on the line for Ice plots.
    • [PUBDEV-8495] - Added option to toggle PDP vs Ice lines on or off.

    πŸ“„ Docs

    • πŸ“š [PUBDEV-8654] - Added documentation on the monotone spline for GAM.
    • [PUBDEV-8665] - Added links to the Additional Resources page to the sites where users can ask questions.
    • ⚑️ [PUBDEV-8658] - Updated the examples for the Residual Analysis Plot.
    • πŸš€ [PUBDEV-8650] - Updated the K8s deployment tutorial.
    • πŸ“š [PUBDEV-8639] - Improved Uplift DRF User Guide documentation.
    • πŸ“„ [PUBDEV-8631] - Shifted the links from the H2O-3 docs page to the User Guide β€œAdditional Resources” page.
    • πŸ›  [PUBDEV-8604] - Fixed MOJO importable/exportable table in User Guide.
    • πŸ— [PUBDEV-8603] - Added a note that MOJOs won’t build if interactions are specified.
    • [PUBDEV-8595] - Added information on how H2O handles date columns.
    • ✏️ [PUBDEV-8590] - Fixed code typos on Admissible ML page in User Guide.
    • [PUBDEV-8591] - Added information on the -hdfs_config tag.
  • v3.36.0.4 Changes

    πŸš€ Download at: http://h2o-release.s3.amazonaws.com/h2o/rel-zorn/4/index.html

    πŸ› Bug

    • πŸ“¦ [PUBDEV-8634] - Fixed logic operations error in R package.
    • [PUBDEV-8626] - Clarified that enum and eigen categorical_encoding values do not work for XGBoost.

    πŸ‘Œ Improvement

    πŸ“„ Docs

    • 🐎 [PUBDEV-8505] - Added information on the make_metrics command to the Performance and Prediction section of the User Guide.
  • v3.36.0.3 Changes

    πŸš€ Download at: http://h2o-release.s3.amazonaws.com/h2o/rel-zorn/3/index.html

    πŸ› Bug

    • [PUBDEV-8567] - Fixed S3 file downloads not working by adding aws_java_sdk_sts as a dependency of H2O persist S3.
    • [PUBDEV-8563] - Added note to GBM, DRF, IF, and EIF that build_tree_one_node=True does not work with current release.
    • 0️⃣ [PUBDEV-8557] - Extended AWS default credential chain instead of replacing it.
    • πŸ›  [PUBDEV-8553] - Fixed import failures for URLs longer than 152 characters.
    • [PUBDEV-8552] - Fix AutoML ignoring verbosity setting.
    • πŸ›  [PUBDEV-8524] - Fixed Huber distribution bug for deviance.

    πŸ‘Œ Improvement

    • 🚚 [PUBDEV-8475] - Removed β€œH2O API Extensions” from h2o.init() output.

    πŸ“„ Docs

    • πŸ“š [PUBDEV-8569] - Corrected typos and inconsistencies in Admissible ML documentation.
    • πŸ“š [PUBDEV-8546] - Updated copyright year in documentation.
    • [PUBDEV-8535] - Clarified feasible intervals for tweedie power.
    • [PUBDEV-8466] - Clarified Java requirements when running H2O on Hadoop.