0.7.15

#Enhancements

• #274 limit number of iteration in CatBoost

0.7.13

Enhancements

• #92 add time checks
• #270 disable stacking for validation type split and repeats > 1
• #271 disable ldb model sort

0.7.12

Enhancements

• #223 Support for repeated validation
• #266 Adjust validation for small datasets

Bug fixes

• #265 fix validation warning
• #264 fix EDA tests
• #261 better error message for missing golden features

Dependencies

• #260 update fastparquet to 0.4.1

0.7.11

Bug fixes

• #258 Fix cant load automl when adjusted validation is used

0.7.10

Enhancements

• #250 New strategies for categorical encoding
• #257 Control algorithm order in not-so-random step

Bug fixes

• #255 Fix overwrite in adjusted models

0.7.9

Enhancements

• #249 Adjust validation type in Compete mode

0.7.8

Enhancements

• #249 Adjust validation type based on data
• #251 add more eval_metrics in regression
• #252 add traceback to error reports

Bug fixes

• #253 Fix error when text data has missing values in test fold

0.7.7

Enhancements

• #73 Optimize AUC

Bug fixes

• #136 RMSE in Extra Trees and Random Forest
• #243 Switch off Xgboost and CatBoost for multiclass with many classes (in extreme switch of Extra Trees and Random Forest)
• #245 Fix ordering of prediction columns

0.7.6

Enhancements

• #240 Change algorithm execution order for default algorithms

Bug fixes:

• #236 Wrong labels for target predictions in the case of -1, 1 target
• #238 Object of type float32 is not JSON serializable
• #239 Value Error: Input contains NaN in numpy training array

0.7.5

Bug fixes

• (#216) Raise exception when all models with error
• (#234) Fix target with first empty value