search

autogluon / autogluon

Fast and Accurate ML in 3 Lines of Code
https://auto.gluon.ai/
Apache License 2.0
7.68k stars 902 forks source link

[BUG] feature_prune_kwargs={"force_prune": True} does not work when tuning_data is on for presets="medium_quality", #4441

Open arturdaraujo opened 3 weeks ago

arturdaraujo commented 3 weeks ago 
Verbosity: 4 (Maximum Logging)
=================== System Info ===================
AutoGluon Version:  1.1.1
Python Version:     3.11.9
Operating System:   Linux
Platform Machine:   x86_64
Platform Version:   #1 SMP Fri Mar 29 23:14:13 UTC 2024
CPU Count:          8
GPU Count:          1
Memory Avail:       8.36 GB / 23.47 GB (35.6%)
Disk Space Avail:   326.38 GB / 911.84 GB (35.8%)
===================================================
Presets specified: ['medium_quality']
============ fit kwarg info ============
User Specified kwargs:
{'auto_stack': False,
 'feature_prune_kwargs': {'force_prune': True},
 'verbosity': 4}
Full kwargs:
{'_feature_generator_kwargs': None,
 '_save_bag_folds': None,
 'ag_args': None,
 'ag_args_ensemble': None,
 'ag_args_fit': None,
 'auto_stack': False,
 'calibrate': 'auto',
 'ds_args': {'clean_up_fits': True,
             'detection_time_frac': 0.25,
             'enable_ray_logging': True,
             'holdout_data': None,
             'holdout_frac': 0.1111111111111111,
             'memory_safe_fits': True,
             'n_folds': 2,
             'n_repeats': 1,
             'validation_procedure': 'holdout'},
 'excluded_model_types': None,
 'feature_generator': 'auto',
 'feature_prune_kwargs': {'force_prune': True},
 'holdout_frac': None,
 'hyperparameter_tune_kwargs': None,
 'included_model_types': None,
 'keep_only_best': False,
 'name_suffix': None,
 'num_bag_folds': None,
 'num_bag_sets': None,
 'num_stack_levels': None,
 'pseudo_data': None,
 'refit_full': False,
 'save_bag_folds': None,
 'save_space': False,
 'set_best_to_refit_full': False,
 'unlabeled_data': None,
 'use_bag_holdout': False,
 'verbosity': 4}
========================================
Warning: Training may take a very long time because `time_limit` was not specified and `train_data` is large (500000 samples, 88.0 MB).
        Consider setting `time_limit` to ensure training finishes within an expected duration or experiment with a small portion of `train_data` to identify an ideal `presets` and `hyperparameters` configuration.
Saving /mnt/d/python_directory_2/models_temp/learner.pkl
Saving /mnt/d/python_directory_2/models_temp/predictor.pkl
Beginning AutoGluon training ...
AutoGluon will save models to "/mnt/d/python_directory_2/models_temp"
Train Data Rows:    500000
Train Data Columns: 41
Tuning Data Rows:    75000
Tuning Data Columns: 41
Label Column:       target
AutoGluon infers your prediction problem is: 'binary' (because only two unique label-values observed).
        2 unique label values:  [0, 1]
        If 'binary' is not the correct problem_type, please manually specify the problem_type parameter during Predictor init (You may specify problem_type as one of: ['binary', 'multiclass', 'regression', 'quantile'])
Problem Type:       binary
Preprocessing data ...
Selected class <--> label mapping:  class 1 = 1, class 0 = 0
Using Feature Generators to preprocess the data ...
Performing general data preprocessing with merged train & validation data, so validation performance may not accurately reflect performance on new test data
Fitting AutoMLPipelineFeatureGenerator...
        Available Memory:                    8580.40 MB
        Train Data (Original)  Memory Usage: 92.13 MB (1.1% of available memory)
        Inferring data type of each feature based on column values. Set feature_metadata_in to manually specify special dtypes of the features.
        Stage 1 Generators:
                Fitting AsTypeFeatureGenerator...
                        Note: Converting 2 features to boolean dtype as they only contain 2 unique values.
                        Original Features (exact raw dtype, raw dtype):
                                ('datetime64[ns]', 'datetime') :  1 | ['time_date']
                                ('float32', 'float')           : 37 | ['STOCH_slowk_3_n_01s__SB_BinaryStats_diff_longstretch0', 'DIV_5_n_01s__DN_Mean', 'KAMA_10_n_01s__CO_FirstMin_ac', 'MACDEXT_macd_10_n_03s__SP_Summaries_welch_rect_centroid', 'MIN_15_n_05s__DN_Spread_Std', ...]
                        Types of features in original data (raw dtype, special dtypes):
                                ('datetime', []) :  1 | ['time_date']
                                ('float', [])    : 37 | ['STOCH_slowk_3_n_01s__SB_BinaryStats_diff_longstretch0', 'DIV_5_n_01s__DN_Mean', 'KAMA_10_n_01s__CO_FirstMin_ac', 'MACDEXT_macd_10_n_03s__SP_Summaries_welch_rect_centroid', 'MIN_15_n_05s__DN_Spread_Std', ...]
                        Types of features in processed data (exact raw dtype, raw dtype):
                                ('datetime64[ns]', 'datetime') :  1 | ['time_date']
                                ('float32', 'float')           : 35 | ['DIV_5_n_01s__DN_Mean', 'KAMA_10_n_01s__CO_FirstMin_ac', 'MACDEXT_macd_10_n_03s__SP_Summaries_welch_rect_centroid', 'MIN_15_n_05s__DN_Spread_Std', 'CCI_3_n_10s__SP_Summaries_welch_rect_centroid', ...]
                                ('int8', 'int')                :  2 | ['STOCH_slowk_3_n_01s__SB_BinaryStats_diff_longstretch0', 'MAX_3_n_30s__FC_LocalSimple_mean1_tauresrat']
                        Types of features in processed data (raw dtype, special dtypes):
                                ('datetime', [])  :  1 | ['time_date']
                                ('float', [])     : 35 | ['DIV_5_n_01s__DN_Mean', 'KAMA_10_n_01s__CO_FirstMin_ac', 'MACDEXT_macd_10_n_03s__SP_Summaries_welch_rect_centroid', 'MIN_15_n_05s__DN_Spread_Std', 'CCI_3_n_10s__SP_Summaries_welch_rect_centroid', ...]
                                ('int', ['bool']) :  2 | ['STOCH_slowk_3_n_01s__SB_BinaryStats_diff_longstretch0', 'MAX_3_n_30s__FC_LocalSimple_mean1_tauresrat']
                        0.2s = Fit runtime
                        38 features in original data used to generate 38 features in processed data.
        Stage 2 Generators:
                Fitting FillNaFeatureGenerator...
                        Types of features in original data (raw dtype, special dtypes):
                                ('datetime', [])  :  1 | ['time_date']
                                ('float', [])     : 35 | ['DIV_5_n_01s__DN_Mean', 'KAMA_10_n_01s__CO_FirstMin_ac', 'MACDEXT_macd_10_n_03s__SP_Summaries_welch_rect_centroid', 'MIN_15_n_05s__DN_Spread_Std', 'CCI_3_n_10s__SP_Summaries_welch_rect_centroid', ...]
                                ('int', ['bool']) :  2 | ['STOCH_slowk_3_n_01s__SB_BinaryStats_diff_longstretch0', 'MAX_3_n_30s__FC_LocalSimple_mean1_tauresrat']
                        Types of features in processed data (exact raw dtype, raw dtype):
                                ('datetime64[ns]', 'datetime') :  1 | ['time_date']
                                ('float32', 'float')           : 35 | ['DIV_5_n_01s__DN_Mean', 'KAMA_10_n_01s__CO_FirstMin_ac', 'MACDEXT_macd_10_n_03s__SP_Summaries_welch_rect_centroid', 'MIN_15_n_05s__DN_Spread_Std', 'CCI_3_n_10s__SP_Summaries_welch_rect_centroid', ...]
                                ('int8', 'int')                :  2 | ['STOCH_slowk_3_n_01s__SB_BinaryStats_diff_longstretch0', 'MAX_3_n_30s__FC_LocalSimple_mean1_tauresrat']
                        Types of features in processed data (raw dtype, special dtypes):
                                ('datetime', [])  :  1 | ['time_date']
                                ('float', [])     : 35 | ['DIV_5_n_01s__DN_Mean', 'KAMA_10_n_01s__CO_FirstMin_ac', 'MACDEXT_macd_10_n_03s__SP_Summaries_welch_rect_centroid', 'MIN_15_n_05s__DN_Spread_Std', 'CCI_3_n_10s__SP_Summaries_welch_rect_centroid', ...]
                                ('int', ['bool']) :  2 | ['STOCH_slowk_3_n_01s__SB_BinaryStats_diff_longstretch0', 'MAX_3_n_30s__FC_LocalSimple_mean1_tauresrat']
                        0.1s = Fit runtime
                        38 features in original data used to generate 38 features in processed data.
        Stage 3 Generators:
                Fitting IdentityFeatureGenerator...
                        Types of features in original data (raw dtype, special dtypes):
                                ('float', [])     : 35 | ['DIV_5_n_01s__DN_Mean', 'KAMA_10_n_01s__CO_FirstMin_ac', 'MACDEXT_macd_10_n_03s__SP_Summaries_welch_rect_centroid', 'MIN_15_n_05s__DN_Spread_Std', 'CCI_3_n_10s__SP_Summaries_welch_rect_centroid', ...]
                                ('int', ['bool']) :  2 | ['STOCH_slowk_3_n_01s__SB_BinaryStats_diff_longstretch0', 'MAX_3_n_30s__FC_LocalSimple_mean1_tauresrat']
                        Types of features in processed data (exact raw dtype, raw dtype):
                                ('float32', 'float') : 35 | ['DIV_5_n_01s__DN_Mean', 'KAMA_10_n_01s__CO_FirstMin_ac', 'MACDEXT_macd_10_n_03s__SP_Summaries_welch_rect_centroid', 'MIN_15_n_05s__DN_Spread_Std', 'CCI_3_n_10s__SP_Summaries_welch_rect_centroid', ...]
                                ('int8', 'int')      :  2 | ['STOCH_slowk_3_n_01s__SB_BinaryStats_diff_longstretch0', 'MAX_3_n_30s__FC_LocalSimple_mean1_tauresrat']
                        Types of features in processed data (raw dtype, special dtypes):
                                ('float', [])     : 35 | ['DIV_5_n_01s__DN_Mean', 'KAMA_10_n_01s__CO_FirstMin_ac', 'MACDEXT_macd_10_n_03s__SP_Summaries_welch_rect_centroid', 'MIN_15_n_05s__DN_Spread_Std', 'CCI_3_n_10s__SP_Summaries_welch_rect_centroid', ...]
                                ('int', ['bool']) :  2 | ['STOCH_slowk_3_n_01s__SB_BinaryStats_diff_longstretch0', 'MAX_3_n_30s__FC_LocalSimple_mean1_tauresrat']
                        0.1s = Fit runtime
                        37 features in original data used to generate 37 features in processed data.
                Skipping CategoryFeatureGenerator: No input feature with required dtypes.
                Fitting DatetimeFeatureGenerator...
                        Types of features in original data (raw dtype, special dtypes):
                                ('datetime', []) : 1 | ['time_date']
                        Types of features in processed data (exact raw dtype, raw dtype):
                                ('int64', 'int') : 5 | ['time_date', 'time_date.year', 'time_date.month', 'time_date.day', 'time_date.dayofweek']
                        Types of features in processed data (raw dtype, special dtypes):
                                ('int', ['datetime_as_int']) : 5 | ['time_date', 'time_date.year', 'time_date.month', 'time_date.day', 'time_date.dayofweek']
                        0.1s = Fit runtime
                        1 features in original data used to generate 5 features in processed data.
                Skipping TextSpecialFeatureGenerator: No input feature with required dtypes.
                Skipping TextNgramFeatureGenerator: No input feature with required dtypes.
                Skipping IdentityFeatureGenerator: No input feature with required dtypes.
                Skipping IsNanFeatureGenerator: No input feature with required dtypes.
        Stage 4 Generators:
                Fitting DropUniqueFeatureGenerator...
                        Types of features in original data (raw dtype, special dtypes):
                                ('float', [])                : 35 | ['DIV_5_n_01s__DN_Mean', 'KAMA_10_n_01s__CO_FirstMin_ac', 'MACDEXT_macd_10_n_03s__SP_Summaries_welch_rect_centroid', 'MIN_15_n_05s__DN_Spread_Std', 'CCI_3_n_10s__SP_Summaries_welch_rect_centroid', ...]
                                ('int', ['bool'])            :  2 | ['STOCH_slowk_3_n_01s__SB_BinaryStats_diff_longstretch0', 'MAX_3_n_30s__FC_LocalSimple_mean1_tauresrat']
                                ('int', ['datetime_as_int']) :  4 | ['time_date', 'time_date.month', 'time_date.day', 'time_date.dayofweek']
                        Types of features in processed data (exact raw dtype, raw dtype):
                                ('float32', 'float') : 35 | ['DIV_5_n_01s__DN_Mean', 'KAMA_10_n_01s__CO_FirstMin_ac', 'MACDEXT_macd_10_n_03s__SP_Summaries_welch_rect_centroid', 'MIN_15_n_05s__DN_Spread_Std', 'CCI_3_n_10s__SP_Summaries_welch_rect_centroid', ...]
                                ('int64', 'int')     :  4 | ['time_date', 'time_date.month', 'time_date.day', 'time_date.dayofweek']
                                ('int8', 'int')      :  2 | ['STOCH_slowk_3_n_01s__SB_BinaryStats_diff_longstretch0', 'MAX_3_n_30s__FC_LocalSimple_mean1_tauresrat']
                        Types of features in processed data (raw dtype, special dtypes):
                                ('float', [])                : 35 | ['DIV_5_n_01s__DN_Mean', 'KAMA_10_n_01s__CO_FirstMin_ac', 'MACDEXT_macd_10_n_03s__SP_Summaries_welch_rect_centroid', 'MIN_15_n_05s__DN_Spread_Std', 'CCI_3_n_10s__SP_Summaries_welch_rect_centroid', ...]
                                ('int', ['bool'])            :  2 | ['STOCH_slowk_3_n_01s__SB_BinaryStats_diff_longstretch0', 'MAX_3_n_30s__FC_LocalSimple_mean1_tauresrat']
                                ('int', ['datetime_as_int']) :  4 | ['time_date', 'time_date.month', 'time_date.day', 'time_date.dayofweek']
                        0.2s = Fit runtime
                        41 features in original data used to generate 41 features in processed data.
        Stage 5 Generators:
                Fitting DropDuplicatesFeatureGenerator...
                        Types of features in original data (raw dtype, special dtypes):
                                ('float', [])                : 35 | ['DIV_5_n_01s__DN_Mean', 'KAMA_10_n_01s__CO_FirstMin_ac', 'MACDEXT_macd_10_n_03s__SP_Summaries_welch_rect_centroid', 'MIN_15_n_05s__DN_Spread_Std', 'CCI_3_n_10s__SP_Summaries_welch_rect_centroid', ...]
                                ('int', ['bool'])            :  2 | ['STOCH_slowk_3_n_01s__SB_BinaryStats_diff_longstretch0', 'MAX_3_n_30s__FC_LocalSimple_mean1_tauresrat']
                                ('int', ['datetime_as_int']) :  4 | ['time_date', 'time_date.month', 'time_date.day', 'time_date.dayofweek']
                        Types of features in processed data (exact raw dtype, raw dtype):
                                ('float32', 'float') : 35 | ['DIV_5_n_01s__DN_Mean', 'KAMA_10_n_01s__CO_FirstMin_ac', 'MACDEXT_macd_10_n_03s__SP_Summaries_welch_rect_centroid', 'MIN_15_n_05s__DN_Spread_Std', 'CCI_3_n_10s__SP_Summaries_welch_rect_centroid', ...]
                                ('int64', 'int')     :  4 | ['time_date', 'time_date.month', 'time_date.day', 'time_date.dayofweek']
                                ('int8', 'int')      :  2 | ['STOCH_slowk_3_n_01s__SB_BinaryStats_diff_longstretch0', 'MAX_3_n_30s__FC_LocalSimple_mean1_tauresrat']
                        Types of features in processed data (raw dtype, special dtypes):
                                ('float', [])                : 35 | ['DIV_5_n_01s__DN_Mean', 'KAMA_10_n_01s__CO_FirstMin_ac', 'MACDEXT_macd_10_n_03s__SP_Summaries_welch_rect_centroid', 'MIN_15_n_05s__DN_Spread_Std', 'CCI_3_n_10s__SP_Summaries_welch_rect_centroid', ...]
                                ('int', ['bool'])            :  2 | ['STOCH_slowk_3_n_01s__SB_BinaryStats_diff_longstretch0', 'MAX_3_n_30s__FC_LocalSimple_mean1_tauresrat']
                                ('int', ['datetime_as_int']) :  4 | ['time_date', 'time_date.month', 'time_date.day', 'time_date.dayofweek']
                        0.1s = Fit runtime
                        41 features in original data used to generate 41 features in processed data.
        Useless Original Features (Count: 3): ['STDDEV_30_n_03s__SC_FluctAnal_2_rsrangefit_50_1_logi_prop_r1', 'STOCHRSI_fastk_3_n_03m__CO_HistogramAMI_even_2_5', 'TEMA_3_n_15m__SC_FluctAnal_2_dfa_50_1_2_logi_prop_r1']
                These features carry no predictive signal and should be manually investigated.
                This is typically a feature which has the same value for all rows.
                These features do not need to be present at inference time.
        Types of features in original data (exact raw dtype, raw dtype):
                ('datetime64[ns]', 'datetime') :  1 | ['time_date']
                ('float32', 'float')           : 37 | ['STOCH_slowk_3_n_01s__SB_BinaryStats_diff_longstretch0', 'DIV_5_n_01s__DN_Mean', 'KAMA_10_n_01s__CO_FirstMin_ac', 'MACDEXT_macd_10_n_03s__SP_Summaries_welch_rect_centroid', 'MIN_15_n_05s__DN_Spread_Std', ...]
        Types of features in original data (raw dtype, special dtypes):
                ('datetime', []) :  1 | ['time_date']
                ('float', [])    : 37 | ['STOCH_slowk_3_n_01s__SB_BinaryStats_diff_longstretch0', 'DIV_5_n_01s__DN_Mean', 'KAMA_10_n_01s__CO_FirstMin_ac', 'MACDEXT_macd_10_n_03s__SP_Summaries_welch_rect_centroid', 'MIN_15_n_05s__DN_Spread_Std', ...]
        Types of features in processed data (exact raw dtype, raw dtype):
                ('float32', 'float') : 35 | ['DIV_5_n_01s__DN_Mean', 'KAMA_10_n_01s__CO_FirstMin_ac', 'MACDEXT_macd_10_n_03s__SP_Summaries_welch_rect_centroid', 'MIN_15_n_05s__DN_Spread_Std', 'CCI_3_n_10s__SP_Summaries_welch_rect_centroid', ...]
                ('int64', 'int')     :  4 | ['time_date', 'time_date.month', 'time_date.day', 'time_date.dayofweek']
                ('int8', 'int')      :  2 | ['STOCH_slowk_3_n_01s__SB_BinaryStats_diff_longstretch0', 'MAX_3_n_30s__FC_LocalSimple_mean1_tauresrat']
        Types of features in processed data (raw dtype, special dtypes):
                ('float', [])                : 35 | ['DIV_5_n_01s__DN_Mean', 'KAMA_10_n_01s__CO_FirstMin_ac', 'MACDEXT_macd_10_n_03s__SP_Summaries_welch_rect_centroid', 'MIN_15_n_05s__DN_Spread_Std', 'CCI_3_n_10s__SP_Summaries_welch_rect_centroid', ...]
                ('int', ['bool'])            :  2 | ['STOCH_slowk_3_n_01s__SB_BinaryStats_diff_longstretch0', 'MAX_3_n_30s__FC_LocalSimple_mean1_tauresrat']
                ('int', ['datetime_as_int']) :  4 | ['time_date', 'time_date.month', 'time_date.day', 'time_date.dayofweek']
        1.0s = Fit runtime
        38 features in original data used to generate 41 features in processed data.
        Train Data (Processed) Memory Usage: 95.42 MB (1.1% of available memory)
Data preprocessing and feature engineering runtime = 1.11s ...
AutoGluon will gauge predictive performance using evaluation metric: 'roc_auc'
        This metric expects predicted probabilities rather than predicted class labels, so you'll need to use predict_proba() instead of predict()
        To change this, specify the eval_metric parameter of Predictor()
Saving /mnt/d/python_directory_2/models_temp/learner.pkl
User-specified model hyperparameters to be fit:
{
        'GBM': [{'extra_trees': True, 'ag_args': {'name_suffix': 'XT'}, 'learning_rate': 0.45}, {'learning_rate': 0.45}],
        'CAT': {'iterations': 10000, 'learning_rate': 0.75, 'allow_writing_files': False, 'eval_metric': 'Logloss', 'thread_count': 8},
}
Saving /mnt/d/python_directory_2/models_temp/utils/data/X.pkl
Saving /mnt/d/python_directory_2/models_temp/utils/data/y.pkl
Saving /mnt/d/python_directory_2/models_temp/utils/data/X_val.pkl
Saving /mnt/d/python_directory_2/models_temp/utils/data/y_val.pkl
Model configs that will be trained (in order):
        LightGBMXT:     {'extra_trees': True, 'ag_args': {'name_suffix': 'XT', 'model_type': <class 'autogluon.tabular.models.lgb.lgb_model.LGBModel'>, 'priority': 90}, 'learning_rate': 0.45}
        LightGBM:       {'learning_rate': 0.45, 'ag_args': {'model_type': <class 'autogluon.tabular.models.lgb.lgb_model.LGBModel'>, 'priority': 90}}
        CatBoost:       {'iterations': 10000, 'learning_rate': 0.75, 'allow_writing_files': False, 'eval_metric': 'Logloss', 'thread_count': 8, 'ag_args': {'model_type': <class 'autogluon.tabular.models.catboost.catboost_model.CatBoostModel'>, 'priority': 70}}
Fitting 3 L1 models ...
Fitting model: LightGBMXT ...
        Dropped 0 of 41 features.
        Fitting LightGBMXT with 'num_gpus': 0, 'num_cpus': 8
        Fitting 10000 rounds... Hyperparameters: {'learning_rate': 0.45, 'extra_trees': True}
[1]     valid_set's binary_logloss: 0.690857
[2]     valid_set's binary_logloss: 0.68494
[3]     valid_set's binary_logloss: 0.684353
[4]     valid_set's binary_logloss: 0.686686
[5]     valid_set's binary_logloss: 0.691233
[6]     valid_set's binary_logloss: 0.691459
[7]     valid_set's binary_logloss: 0.701168
[8]     valid_set's binary_logloss: 0.712405
[9]     valid_set's binary_logloss: 0.716977
[10]    valid_set's binary_logloss: 0.717592
[11]    valid_set's binary_logloss: 0.715944
[12]    valid_set's binary_logloss: 0.700291
[13]    valid_set's binary_logloss: 0.702016
[14]    valid_set's binary_logloss: 0.706436
[15]    valid_set's binary_logloss: 0.703444
[16]    valid_set's binary_logloss: 0.702883
[17]    valid_set's binary_logloss: 0.702753
[18]    valid_set's binary_logloss: 0.70785
[19]    valid_set's binary_logloss: 0.70979
[20]    valid_set's binary_logloss: 0.711334
[21]    valid_set's binary_logloss: 0.710002
[22]    valid_set's binary_logloss: 0.709985
[23]    valid_set's binary_logloss: 0.710674
[24]    valid_set's binary_logloss: 0.711116
Saving /mnt/d/python_directory_2/models_temp/models/LightGBMXT/model.pkl
Saving /mnt/d/python_directory_2/models_temp/utils/attr/LightGBMXT/y_pred_proba_val.pkl
        0.541    = Validation score   (roc_auc)
        1.13s    = Training   runtime
        0.02s    = Validation runtime
        3801989.4        = Inference  throughput (rows/s | 75000 batch size)
Saving /mnt/d/python_directory_2/models_temp/models/trainer.pkl
Fitting model: LightGBM ...
        Dropped 0 of 41 features.
        Fitting LightGBM with 'num_gpus': 0, 'num_cpus': 8
        Fitting 10000 rounds... Hyperparameters: {'learning_rate': 0.45}
[1]     valid_set's binary_logloss: 0.726226
[2]     valid_set's binary_logloss: 0.741846
[3]     valid_set's binary_logloss: 0.737792
[4]     valid_set's binary_logloss: 0.791165
[5]     valid_set's binary_logloss: 0.838001
[6]     valid_set's binary_logloss: 0.838667
[7]     valid_set's binary_logloss: 0.839008
[8]     valid_set's binary_logloss: 0.861499
[9]     valid_set's binary_logloss: 0.85956
[10]    valid_set's binary_logloss: 0.847175
[11]    valid_set's binary_logloss: 0.84703
[12]    valid_set's binary_logloss: 0.846104
[13]    valid_set's binary_logloss: 0.832776
[14]    valid_set's binary_logloss: 0.833003
[15]    valid_set's binary_logloss: 0.831728
[16]    valid_set's binary_logloss: 0.834595
[17]    valid_set's binary_logloss: 0.833374
[18]    valid_set's binary_logloss: 0.833442
[19]    valid_set's binary_logloss: 0.833311
[20]    valid_set's binary_logloss: 0.824875
[21]    valid_set's binary_logloss: 0.822567
Saving /mnt/d/python_directory_2/models_temp/models/LightGBM/model.pkl
Saving /mnt/d/python_directory_2/models_temp/utils/attr/LightGBM/y_pred_proba_val.pkl
        0.5      = Validation score   (roc_auc)
        1.01s    = Training   runtime
        0.02s    = Validation runtime
        3795109.1        = Inference  throughput (rows/s | 75000 batch size)
Saving /mnt/d/python_directory_2/models_temp/models/trainer.pkl
Fitting model: CatBoost ...
        Dropped 0 of 41 features.
        Fitting CatBoost with 'num_gpus': 0, 'num_cpus': 8
        Catboost model hyperparameters: {'iterations': 10000, 'learning_rate': 0.75, 'random_seed': 0, 'allow_writing_files': False, 'eval_metric': 'Logloss', 'thread_count': 8}
0:      learn: 0.6533180        test: 0.7174133 best: 0.7174133 (0)     total: 44.1ms   remaining: 7m 20s
1:      learn: 0.6298120        test: 0.7229527 best: 0.7174133 (0)     total: 82.8ms   remaining: 6m 53s
2:      learn: 0.6126701        test: 0.7287912 best: 0.7174133 (0)     total: 122ms    remaining: 6m 46s
3:      learn: 0.6073046        test: 0.7166248 best: 0.7166248 (3)     total: 158ms    remaining: 6m 35s
4:      learn: 0.5929676        test: 0.7188954 best: 0.7166248 (3)     total: 194ms    remaining: 6m 27s
5:      learn: 0.5770597        test: 0.7453000 best: 0.7166248 (3)     total: 227ms    remaining: 6m 18s
6:      learn: 0.5684997        test: 0.7506057 best: 0.7166248 (3)     total: 271ms    remaining: 6m 26s
7:      learn: 0.5586667        test: 0.6851480 best: 0.6851480 (7)     total: 313ms    remaining: 6m 31s
8:      learn: 0.5533758        test: 0.6924916 best: 0.6851480 (7)     total: 356ms    remaining: 6m 35s
9:      learn: 0.5475293        test: 0.6908779 best: 0.6851480 (7)     total: 398ms    remaining: 6m 37s
10:     learn: 0.5413246        test: 0.6897852 best: 0.6851480 (7)     total: 431ms    remaining: 6m 31s
11:     learn: 0.5372513        test: 0.6887706 best: 0.6851480 (7)     total: 466ms    remaining: 6m 27s
12:     learn: 0.5334703        test: 0.6926530 best: 0.6851480 (7)     total: 500ms    remaining: 6m 24s
13:     learn: 0.5298752        test: 0.6958383 best: 0.6851480 (7)     total: 541ms    remaining: 6m 25s
14:     learn: 0.5268030        test: 0.6941996 best: 0.6851480 (7)     total: 581ms    remaining: 6m 26s
15:     learn: 0.5180547        test: 0.6958168 best: 0.6851480 (7)     total: 631ms    remaining: 6m 33s
16:     learn: 0.5125751        test: 0.6883260 best: 0.6851480 (7)     total: 670ms    remaining: 6m 33s
17:     learn: 0.5112367        test: 0.6874500 best: 0.6851480 (7)     total: 719ms    remaining: 6m 38s
18:     learn: 0.5060452        test: 0.6895496 best: 0.6851480 (7)     total: 765ms    remaining: 6m 42s
19:     learn: 0.5014992        test: 0.6867924 best: 0.6851480 (7)     total: 808ms    remaining: 6m 43s
20:     learn: 0.4988651        test: 0.6899766 best: 0.6851480 (7)     total: 848ms    remaining: 6m 43s
21:     learn: 0.4955378        test: 0.6872165 best: 0.6851480 (7)     total: 894ms    remaining: 6m 45s
22:     learn: 0.4904387        test: 0.6821777 best: 0.6821777 (22)    total: 944ms    remaining: 6m 49s
23:     learn: 0.4883393        test: 0.6846079 best: 0.6821777 (22)    total: 985ms    remaining: 6m 49s
24:     learn: 0.4851253        test: 0.6906374 best: 0.6821777 (22)    total: 1.02s    remaining: 6m 47s
25:     learn: 0.4807459        test: 0.6861663 best: 0.6821777 (22)    total: 1.07s    remaining: 6m 49s
26:     learn: 0.4782397        test: 0.6860673 best: 0.6821777 (22)    total: 1.12s    remaining: 6m 52s
27:     learn: 0.4775118        test: 0.6907217 best: 0.6821777 (22)    total: 1.16s    remaining: 6m 54s
28:     learn: 0.4745040        test: 0.7047771 best: 0.6821777 (22)    total: 1.2s     remaining: 6m 53s
29:     learn: 0.4728311        test: 0.7234329 best: 0.6821777 (22)    total: 1.24s    remaining: 6m 53s
30:     learn: 0.4691995        test: 0.7221445 best: 0.6821777 (22)    total: 1.29s    remaining: 6m 54s
31:     learn: 0.4675475        test: 0.7224027 best: 0.6821777 (22)    total: 1.33s    remaining: 6m 53s
32:     learn: 0.4658990        test: 0.7221766 best: 0.6821777 (22)    total: 1.37s    remaining: 6m 54s
33:     learn: 0.4650485        test: 0.7219495 best: 0.6821777 (22)    total: 1.4s     remaining: 6m 51s
34:     learn: 0.4638509        test: 0.7243998 best: 0.6821777 (22)    total: 1.45s    remaining: 6m 52s
35:     learn: 0.4621719        test: 0.7230784 best: 0.6821777 (22)    total: 1.49s    remaining: 6m 52s
36:     learn: 0.4615038        test: 0.7235578 best: 0.6821777 (22)    total: 1.53s    remaining: 6m 52s
37:     learn: 0.4609860        test: 0.7225358 best: 0.6821777 (22)    total: 1.57s    remaining: 6m 51s
38:     learn: 0.4590647        test: 0.7220221 best: 0.6821777 (22)    total: 1.61s    remaining: 6m 52s
39:     learn: 0.4584479        test: 0.7252117 best: 0.6821777 (22)    total: 1.66s    remaining: 6m 52s
40:     learn: 0.4566977        test: 0.7285681 best: 0.6821777 (22)    total: 1.71s    remaining: 6m 54s
41:     learn: 0.4550106        test: 0.7298182 best: 0.6821777 (22)    total: 1.76s    remaining: 6m 58s
42:     learn: 0.4537528        test: 0.7286350 best: 0.6821777 (22)    total: 1.82s    remaining: 7m 2s
43:     learn: 0.4530702        test: 0.7217189 best: 0.6821777 (22)    total: 1.91s    remaining: 7m 13s
44:     learn: 0.4515620        test: 0.7219945 best: 0.6821777 (22)    total: 1.99s    remaining: 7m 20s
45:     learn: 0.4501892        test: 0.7196554 best: 0.6821777 (22)    total: 2.07s    remaining: 7m 27s
46:     learn: 0.4476642        test: 0.7128948 best: 0.6821777 (22)    total: 2.12s    remaining: 7m 28s
47:     learn: 0.4466513        test: 0.7867258 best: 0.6821777 (22)    total: 2.17s    remaining: 7m 29s
48:     learn: 0.4461165        test: 0.7878465 best: 0.6821777 (22)    total: 2.22s    remaining: 7m 30s
49:     learn: 0.4450580        test: 0.7877115 best: 0.6821777 (22)    total: 2.27s    remaining: 7m 32s

bestTest = 0.6821776997
bestIteration = 22

Shrink model to first 23 iterations.
Saving /mnt/d/python_directory_2/models_temp/models/CatBoost/model.pkl
Saving /mnt/d/python_directory_2/models_temp/utils/attr/CatBoost/y_pred_proba_val.pkl
        0.6525   = Validation score   (roc_auc)
        2.64s    = Training   runtime
        0.01s    = Validation runtime
        7341255.5        = Inference  throughput (rows/s | 75000 batch size)
Saving /mnt/d/python_directory_2/models_temp/models/trainer.pkl
Loading: /mnt/d/python_directory_2/models_temp/models/LightGBMXT/model.pkl
Performing feature pruning with model: FeatureSelector_LightGBMXT, total time limit: 300s, stop threshold: 10, prune ratio: 0.05, prune threshold: noise.
        Number of training samples 500000 is greater than 50000. Using 50000 samples as training data.
Traceback (most recent call last):
  File "<stdin>", line 3, in <module>
  File "/home/artur/miniforge3/envs/py311_ag/lib/python3.11/site-packages/autogluon/core/utils/decorators.py", line 31, in _call
    return f(*gargs, **gkwargs)
           ^^^^^^^^^^^^^^^^^^^^
  File "/home/artur/miniforge3/envs/py311_ag/lib/python3.11/site-packages/autogluon/tabular/predictor/predictor.py", line 1167, in fit
    self._fit(ag_fit_kwargs=ag_fit_kwargs, ag_post_fit_kwargs=ag_post_fit_kwargs)
  File "/home/artur/miniforge3/envs/py311_ag/lib/python3.11/site-packages/autogluon/tabular/predictor/predictor.py", line 1173, in _fit
    self._learner.fit(**ag_fit_kwargs)
  File "/home/artur/miniforge3/envs/py311_ag/lib/python3.11/site-packages/autogluon/tabular/learner/abstract_learner.py", line 159, in fit
    return self._fit(X=X, X_val=X_val, **kwargs)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/artur/miniforge3/envs/py311_ag/lib/python3.11/site-packages/autogluon/tabular/learner/default_learner.py", line 122, in _fit
    trainer.fit(
  File "/home/artur/miniforge3/envs/py311_ag/lib/python3.11/site-packages/autogluon/tabular/trainer/auto_trainer.py", line 125, in fit
    self._train_multi_and_ensemble(
  File "/home/artur/miniforge3/envs/py311_ag/lib/python3.11/site-packages/autogluon/core/trainer/abstract_trainer.py", line 2589, in _train_multi_and_ensemble
    model_names_fit = self.train_multi_levels(
                      ^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/artur/miniforge3/envs/py311_ag/lib/python3.11/site-packages/autogluon/core/trainer/abstract_trainer.py", line 452, in train_multi_levels
    base_model_names, aux_models = self.stack_new_level(
                                   ^^^^^^^^^^^^^^^^^^^^^
  File "/home/artur/miniforge3/envs/py311_ag/lib/python3.11/site-packages/autogluon/core/trainer/abstract_trainer.py", line 600, in stack_new_level
    core_models = self.stack_new_level_core(
                  ^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/artur/miniforge3/envs/py311_ag/lib/python3.11/site-packages/autogluon/core/trainer/abstract_trainer.py", line 730, in stack_new_level_core
    return self._train_multi(
           ^^^^^^^^^^^^^^^^^^
  File "/home/artur/miniforge3/envs/py311_ag/lib/python3.11/site-packages/autogluon/core/trainer/abstract_trainer.py", line 2539, in _train_multi
    model_names_trained = self._train_multi_initial(
                          ^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/artur/miniforge3/envs/py311_ag/lib/python3.11/site-packages/autogluon/core/trainer/abstract_trainer.py", line 2422, in _train_multi_initial
    candidate_features = self._proxy_model_feature_prune(
                         ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/artur/miniforge3/envs/py311_ag/lib/python3.11/site-packages/autogluon/core/trainer/abstract_trainer.py", line 2669, in _proxy_model_feature_prune
    candidate_features = selector.select_features(**feature_prune_kwargs, **model_fit_kwargs)
                         ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/artur/miniforge3/envs/py311_ag/lib/python3.11/site-packages/autogluon/core/utils/feature_selection.py", line 225, in select_features
    X, y, X_val, y_val, X_fi, y_fi, prune_threshold, noise_columns, feature_metadata = self.setup(
                                                                                       ^^^^^^^^^^^
  File "/home/artur/miniforge3/envs/py311_ag/lib/python3.11/site-packages/autogluon/core/utils/feature_selection.py", line 549, in setup
    X_train, _, y_train, _ = generate_train_test_split(X=X, y=y, problem_type=self.problem_type, random_state=random_state, test_size=drop_ratio)
                             ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/artur/miniforge3/envs/py311_ag/lib/python3.11/site-packages/autogluon/core/utils/utils.py", line 513, in generate_train_test_split
    random.seed(random_state)
  File "/home/artur/miniforge3/envs/py311_ag/lib/python3.11/random.py", line 160, in seed
    raise TypeError('The only supported seed types are: None,\n'
TypeError: The only supported seed types are: None,
int, float, str, bytes, and bytearray.
arturdaraujo commented 3 weeks ago

Update: The problem is coming from the training data, I'm still not sure what is causing.