Predicting Columns in a Table - In Depth¶
Tip: If you are new to AutoGluon, review Predicting Columns in a Table - Quick Start to learn the basics of the AutoGluon API. To learn how to add your own custom models to the set that AutoGluon trains, tunes, and ensembles, review Adding a custom model to AutoGluon.
This tutorial describes how you can exert greater control when using
AutoGluon’s fit() or predict(). Recall that to maximize
predictive performance, you should always first try fit() with all
default arguments except eval_metric and presets, before you
experiment with other arguments covered in this in-depth tutorial like
hyperparameter_tune_kwargs, hyperparameters,
num_stack_levels, num_bag_folds, num_bag_sets, etc.
Using the same census data table as in the Predicting Columns in a Table - Quick Start
tutorial, we’ll now predict the occupation of an individual - a
multiclass classification problem. Start by importing AutoGluon’s
TabularPredictor and TabularDataset, and loading the data.
from autogluon.tabular import TabularDataset, TabularPredictor
import numpy as np
train_data = TabularDataset('https://autogluon.s3.amazonaws.com/datasets/Inc/train.csv')
subsample_size = 500 # subsample subset of data for faster demo, try setting this to much larger values
train_data = train_data.sample(n=subsample_size, random_state=0)
print(train_data.head())
label = 'occupation'
print("Summary of occupation column: \n", train_data['occupation'].describe())
new_data = TabularDataset('https://autogluon.s3.amazonaws.com/datasets/Inc/test.csv')
test_data = new_data[5000:].copy() # this should be separate data in your applications
y_test = test_data[label]
test_data_nolabel = test_data.drop(columns=[label]) # delete label column
val_data = new_data[:5000].copy()
metric = 'accuracy' # we specify eval-metric just for demo (unnecessary as it's the default)
age workclass fnlwgt education education-num 6118 51 Private 39264 Some-college 10
23204 58 Private 51662 10th 6
29590 40 Private 326310 Some-college 10
18116 37 Private 222450 HS-grad 9
33964 62 Private 109190 Bachelors 13
marital-status occupation relationship race sex 6118 Married-civ-spouse Exec-managerial Wife White Female
23204 Married-civ-spouse Other-service Wife White Female
29590 Married-civ-spouse Craft-repair Husband White Male
18116 Never-married Sales Not-in-family White Male
33964 Married-civ-spouse Exec-managerial Husband White Male
capital-gain capital-loss hours-per-week native-country class
6118 0 0 40 United-States >50K
23204 0 0 8 United-States <=50K
29590 0 0 44 United-States <=50K
18116 0 2339 40 El-Salvador <=50K
33964 15024 0 40 United-States >50K
Summary of occupation column:
count 500
unique 15
top Exec-managerial
freq 77
Name: occupation, dtype: object
Specifying hyperparameters and tuning them¶
We first demonstrate hyperparameter-tuning and how you can provide your own validation dataset that AutoGluon internally relies on to: tune hyperparameters, early-stop iterative training, and construct model ensembles. One reason you may specify validation data is when future test data will stem from a different distribution than training data (and your specified validation data is more representative of the future data that will likely be encountered).
If you don’t have a strong reason to provide your own validation
dataset, we recommend you omit the tuning_data argument. This lets
AutoGluon automatically select validation data from your provided
training set (it uses smart strategies such as stratified sampling). For
greater control, you can specify the holdout_frac argument to tell
AutoGluon what fraction of the provided training data to hold out for
validation.
Caution: Since AutoGluon tunes internal knobs based on this
validation data, performance estimates reported on this data may be
over-optimistic. For unbiased performance estimates, you should always
call predict() on a separate dataset (that was never passed to
fit()), as we did in the previous Quick-Start tutorial. We also
emphasize that most options specified in this tutorial are chosen to
minimize runtime for the purposes of demonstration and you should select
more reasonable values in order to obtain high-quality models.
fit() trains neural networks and various types of tree ensembles by
default. You can specify various hyperparameter values for each type of
model. For each hyperparameter, you can either specify a single fixed
value, or a search space of values to consider during hyperparameter
optimization. Hyperparameters which you do not specify are left at
default settings chosen automatically by AutoGluon, which may be fixed
values or search spaces.
import autogluon.core as ag
nn_options = { # specifies non-default hyperparameter values for neural network models
'num_epochs': 10, # number of training epochs (controls training time of NN models)
'learning_rate': ag.space.Real(1e-4, 1e-2, default=5e-4, log=True), # learning rate used in training (real-valued hyperparameter searched on log-scale)
'activation': ag.space.Categorical('relu', 'softrelu', 'tanh'), # activation function used in NN (categorical hyperparameter, default = first entry)
'layers': ag.space.Categorical([100], [1000], [200, 100], [300, 200, 100]), # each choice for categorical hyperparameter 'layers' corresponds to list of sizes for each NN layer to use
'dropout_prob': ag.space.Real(0.0, 0.5, default=0.1), # dropout probability (real-valued hyperparameter)
}
gbm_options = { # specifies non-default hyperparameter values for lightGBM gradient boosted trees
'num_boost_round': 100, # number of boosting rounds (controls training time of GBM models)
'num_leaves': ag.space.Int(lower=26, upper=66, default=36), # number of leaves in trees (integer hyperparameter)
}
hyperparameters = { # hyperparameters of each model type
'GBM': gbm_options,
'NN': nn_options, # NOTE: comment this line out if you get errors on Mac OSX
} # When these keys are missing from hyperparameters dict, no models of that type are trained
time_limit = 2*60 # train various models for ~2 min
num_trials = 5 # try at most 5 different hyperparameter configurations for each type of model
search_strategy = 'auto' # to tune hyperparameters using Bayesian optimization routine with a local scheduler
hyperparameter_tune_kwargs = { # HPO is not performed unless hyperparameter_tune_kwargs is specified
'num_trials': num_trials,
'scheduler' : 'local',
'searcher': search_strategy,
}
predictor = TabularPredictor(label=label, eval_metric=metric).fit(
train_data, tuning_data=val_data, time_limit=time_limit,
hyperparameters=hyperparameters, hyperparameter_tune_kwargs=hyperparameter_tune_kwargs,
)
No path specified. Models will be saved in: "AutogluonModels/ag-20210831_220529/"
Warning: hyperparameter tuning is currently experimental and may cause the process to hang.
Beginning AutoGluon training ... Time limit = 120s
AutoGluon will save models to "AutogluonModels/ag-20210831_220529/"
AutoGluon Version: 0.3.1b20210831
Train Data Rows: 500
Train Data Columns: 14
Tuning Data Rows: 5000
Tuning Data Columns: 14
Preprocessing data ...
AutoGluon infers your prediction problem is: 'multiclass' (because dtype of label-column == object).
First 10 (of 15) unique label values: [' Exec-managerial', ' Other-service', ' Craft-repair', ' Sales', ' Prof-specialty', ' Protective-serv', ' ?', ' Adm-clerical', ' Machine-op-inspct', ' Tech-support']
If 'multiclass' is not the correct problem_type, please manually specify the problem_type argument in fit() (You may specify problem_type as one of: ['binary', 'multiclass', 'regression'])
Warning: Some classes in the training set have fewer than 10 examples. AutoGluon will only keep 12 out of 15 classes for training and will not try to predict the rare classes. To keep more classes, increase the number of datapoints from these rare classes in the training data or reduce label_count_threshold.
Fraction of data from classes with at least 10 examples that will be kept for training models: 0.978
Train Data Class Count: 12
Using Feature Generators to preprocess the data ...
Fitting AutoMLPipelineFeatureGenerator...
Available Memory: 22344.12 MB
Train Data (Original) Memory Usage: 3.11 MB (0.0% 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.
Stage 2 Generators:
Fitting FillNaFeatureGenerator...
Stage 3 Generators:
Fitting IdentityFeatureGenerator...
Fitting CategoryFeatureGenerator...
Fitting CategoryMemoryMinimizeFeatureGenerator...
Stage 4 Generators:
Fitting DropUniqueFeatureGenerator...
Types of features in original data (raw dtype, special dtypes):
('int', []) : 6 | ['age', 'fnlwgt', 'education-num', 'capital-gain', 'capital-loss', ...]
('object', []) : 8 | ['workclass', 'education', 'marital-status', 'relationship', 'race', ...]
Types of features in processed data (raw dtype, special dtypes):
('category', []) : 6 | ['workclass', 'education', 'marital-status', 'relationship', 'race', ...]
('int', []) : 6 | ['age', 'fnlwgt', 'education-num', 'capital-gain', 'capital-loss', ...]
('int', ['bool']) : 2 | ['sex', 'class']
0.1s = Fit runtime
14 features in original data used to generate 14 features in processed data.
Train Data (Processed) Memory Usage: 0.3 MB (0.0% of available memory)
Data preprocessing and feature engineering runtime = 0.11s ...
AutoGluon will gauge predictive performance using evaluation metric: 'accuracy'
To change this, specify the eval_metric argument of fit()
Fitting 2 L1 models ...
Hyperparameter tuning model: LightGBM ...
0%| | 0/5 [00:00<?, ?it/s]
Fitted model: LightGBM/T0 ...
0.3033 = Validation score (accuracy)
0.65s = Training runtime
0.04s = Validation runtime
Fitted model: LightGBM/T1 ...
0.3086 = Validation score (accuracy)
0.57s = Training runtime
0.07s = Validation runtime
Fitted model: LightGBM/T2 ...
0.3024 = Validation score (accuracy)
0.68s = Training runtime
0.07s = Validation runtime
Fitted model: LightGBM/T3 ...
0.2707 = Validation score (accuracy)
1.38s = Training runtime
0.07s = Validation runtime
Fitted model: LightGBM/T4 ...
0.3121 = Validation score (accuracy)
0.51s = Training runtime
0.02s = Validation runtime
Hyperparameter tuning model: NeuralNetMXNet ...
0%| | 0/5 [00:00<?, ?it/s]
Fitted model: NeuralNetMXNet/T0 ...
0.17 = Validation score (accuracy)
5.67s = Training runtime
0.41s = Validation runtime
Fitted model: NeuralNetMXNet/T1 ...
0.1735 = Validation score (accuracy)
5.02s = Training runtime
0.41s = Validation runtime
Fitted model: NeuralNetMXNet/T2 ...
0.1306 = Validation score (accuracy)
5.12s = Training runtime
0.41s = Validation runtime
Fitted model: NeuralNetMXNet/T3 ...
0.1726 = Validation score (accuracy)
4.97s = Training runtime
0.41s = Validation runtime
Fitted model: NeuralNetMXNet/T4 ...
0.2649 = Validation score (accuracy)
5.61s = Training runtime
0.45s = Validation runtime
Fitting model: WeightedEnsemble_L2 ... Training model for up to 119.89s of the 81.59s of remaining time.
0.3186 = Validation score (accuracy)
1.26s = Training runtime
0.0s = Validation runtime
AutoGluon training complete, total runtime = 39.69s ...
TabularPredictor saved. To load, use: predictor = TabularPredictor.load("AutogluonModels/ag-20210831_220529/")
We again demonstrate how to use the trained models to predict on the test data.
y_pred = predictor.predict(test_data_nolabel)
print("Predictions: ", list(y_pred)[:5])
perf = predictor.evaluate(test_data, auxiliary_metrics=False)
Predictions: [' Exec-managerial', ' Craft-repair', ' Craft-repair', ' Adm-clerical', ' Sales']
Evaluation: accuracy on test data: 0.2914657160830363
Evaluations on test data:
{
"accuracy": 0.2914657160830363
}
Use the following to view a summary of what happened during fit. Now this command will show details of the hyperparameter-tuning process for each type of model:
results = predictor.fit_summary(show_plot=True)
* Summary of fit() *
Estimated performance of each model:
model score_val pred_time_val fit_time pred_time_val_marginal fit_time_marginal stack_level can_infer fit_order
0 WeightedEnsemble_L2 0.318639 0.872146 13.062934 0.000993 1.264167 2 True 11
1 LightGBM/T4 0.312077 0.016086 0.514884 0.016086 0.514884 1 True 5
2 LightGBM/T1 0.308591 0.067504 0.573090 0.067504 0.573090 1 True 2
3 LightGBM/T0 0.303260 0.039296 0.651870 0.039296 0.651870 1 True 1
4 LightGBM/T2 0.302440 0.068444 0.683065 0.068444 0.683065 1 True 3
5 LightGBM/T3 0.270658 0.068377 1.384378 0.068377 1.384378 1 True 4
6 NeuralNetMXNet/T4 0.264917 0.446777 5.610023 0.446777 5.610023 1 True 10
7 NeuralNetMXNet/T1 0.173467 0.413294 5.022161 0.413294 5.022161 1 True 7
8 NeuralNetMXNet/T3 0.172647 0.407352 4.972622 0.407352 4.972622 1 True 9
9 NeuralNetMXNet/T0 0.169982 0.408290 5.673861 0.408290 5.673861 1 True 6
10 NeuralNetMXNet/T2 0.130613 0.407313 5.119499 0.407313 5.119499 1 True 8
Number of models trained: 11
Types of models trained:
{'LGBModel', 'WeightedEnsembleModel', 'TabularNeuralNetModel'}
Bagging used: False
Multi-layer stack-ensembling used: False
Feature Metadata (Processed):
(raw dtype, special dtypes):
('category', []) : 6 | ['workclass', 'education', 'marital-status', 'relationship', 'race', ...]
('int', []) : 6 | ['age', 'fnlwgt', 'education-num', 'capital-gain', 'capital-loss', ...]
('int', ['bool']) : 2 | ['sex', 'class']
* End of fit() summary *
/var/lib/jenkins/workspace/workspace/autogluon-tutorial-tabular-v3/core/src/autogluon/core/utils/plots.py:138: UserWarning: AutoGluon summary plots cannot be created because bokeh is not installed. To see plots, please do: "pip install bokeh==2.0.1"
warnings.warn('AutoGluon summary plots cannot be created because bokeh is not installed. To see plots, please do: "pip install bokeh==2.0.1"')
In the above example, the predictive performance may be poor because we
specified very little training to ensure quick runtimes. You can call
fit() multiple times while modifying the above settings to better
understand how these choices affect performance outcomes. For example:
you can comment out the train_data.head command or increase
subsample_size to train using a larger dataset, increase the
num_epochs and num_boost_round hyperparameters, and increase the
time_limit (which you should do for all code in these tutorials). To
see more detailed output during the execution of fit(), you can also
pass in the argument: verbosity = 3.
Model ensembling with stacking/bagging¶
Beyond hyperparameter-tuning with a correctly-specified evaluation
metric, two other methods to boost predictive performance are bagging
and stack-ensembling. You’ll often
see performance improve if you specify num_bag_folds = 5-10,
num_stack_levels = 1-3 in the call to fit(), but this will
increase training times and memory/disk usage.
predictor = TabularPredictor(label=label, eval_metric=metric).fit(train_data,
num_bag_folds=5, num_bag_sets=1, num_stack_levels=1,
hyperparameters = {'NN': {'num_epochs': 2}, 'GBM': {'num_boost_round': 20}}, # last argument is just for quick demo here, omit it in real applications
)
No path specified. Models will be saved in: "AutogluonModels/ag-20210831_220611/"
Beginning AutoGluon training ...
AutoGluon will save models to "AutogluonModels/ag-20210831_220611/"
AutoGluon Version: 0.3.1b20210831
Train Data Rows: 500
Train Data Columns: 14
Preprocessing data ...
AutoGluon infers your prediction problem is: 'multiclass' (because dtype of label-column == object).
First 10 (of 15) unique label values: [' Exec-managerial', ' Other-service', ' Craft-repair', ' Sales', ' Prof-specialty', ' Protective-serv', ' ?', ' Adm-clerical', ' Machine-op-inspct', ' Tech-support']
If 'multiclass' is not the correct problem_type, please manually specify the problem_type argument in fit() (You may specify problem_type as one of: ['binary', 'multiclass', 'regression'])
Warning: Some classes in the training set have fewer than 10 examples. AutoGluon will only keep 12 out of 15 classes for training and will not try to predict the rare classes. To keep more classes, increase the number of datapoints from these rare classes in the training data or reduce label_count_threshold.
Fraction of data from classes with at least 10 examples that will be kept for training models: 0.978
Train Data Class Count: 12
Using Feature Generators to preprocess the data ...
Fitting AutoMLPipelineFeatureGenerator...
Available Memory: 22087.94 MB
Train Data (Original) Memory Usage: 0.28 MB (0.0% 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.
Stage 2 Generators:
Fitting FillNaFeatureGenerator...
Stage 3 Generators:
Fitting IdentityFeatureGenerator...
Fitting CategoryFeatureGenerator...
Fitting CategoryMemoryMinimizeFeatureGenerator...
Stage 4 Generators:
Fitting DropUniqueFeatureGenerator...
Types of features in original data (raw dtype, special dtypes):
('int', []) : 6 | ['age', 'fnlwgt', 'education-num', 'capital-gain', 'capital-loss', ...]
('object', []) : 8 | ['workclass', 'education', 'marital-status', 'relationship', 'race', ...]
Types of features in processed data (raw dtype, special dtypes):
('category', []) : 6 | ['workclass', 'education', 'marital-status', 'relationship', 'race', ...]
('int', []) : 6 | ['age', 'fnlwgt', 'education-num', 'capital-gain', 'capital-loss', ...]
('int', ['bool']) : 2 | ['sex', 'class']
0.1s = Fit runtime
14 features in original data used to generate 14 features in processed data.
Train Data (Processed) Memory Usage: 0.03 MB (0.0% of available memory)
Data preprocessing and feature engineering runtime = 0.08s ...
AutoGluon will gauge predictive performance using evaluation metric: 'accuracy'
To change this, specify the eval_metric argument of fit()
AutoGluon will fit 2 stack levels (L1 to L2) ...
Fitting 2 L1 models ...
Fitting model: LightGBM_BAG_L1 ...
0.3067 = Validation score (accuracy)
0.67s = Training runtime
0.03s = Validation runtime
Fitting model: NeuralNetMXNet_BAG_L1 ...
0.1002 = Validation score (accuracy)
1.56s = Training runtime
0.1s = Validation runtime
Fitting model: WeightedEnsemble_L2 ...
0.3067 = Validation score (accuracy)
0.11s = Training runtime
0.0s = Validation runtime
Fitting 2 L2 models ...
Fitting model: LightGBM_BAG_L2 ...
0.2965 = Validation score (accuracy)
0.89s = Training runtime
0.03s = Validation runtime
Fitting model: NeuralNetMXNet_BAG_L2 ...
0.0777 = Validation score (accuracy)
1.99s = Training runtime
0.14s = Validation runtime
Fitting model: WeightedEnsemble_L3 ...
0.2986 = Validation score (accuracy)
0.11s = Training runtime
0.0s = Validation runtime
AutoGluon training complete, total runtime = 5.89s ...
TabularPredictor saved. To load, use: predictor = TabularPredictor.load("AutogluonModels/ag-20210831_220611/")
You should not provide tuning_data when stacking/bagging, and
instead provide all your available data as train_data (which
AutoGluon will split in more intellgent ways). num_bag_sets controls
how many times the k-fold bagging process is repeated to further reduce
variance (increasing this may further boost accuracy but will
substantially increase training times, inference latency, and
memory/disk usage). Rather than manually searching for good
bagging/stacking values yourself, AutoGluon will automatically select
good values for you if you specify auto_stack instead:
save_path = 'agModels-predictOccupation' # folder where to store trained models
predictor = TabularPredictor(label=label, eval_metric=metric, path=save_path).fit(
train_data, auto_stack=True,
time_limit=30, hyperparameters={'NN': {'num_epochs': 2}, 'GBM': {'num_boost_round': 20}} # last 2 arguments are for quick demo, omit them in real applications
)
Beginning AutoGluon training ... Time limit = 30s
AutoGluon will save models to "agModels-predictOccupation/"
AutoGluon Version: 0.3.1b20210831
Train Data Rows: 500
Train Data Columns: 14
Preprocessing data ...
AutoGluon infers your prediction problem is: 'multiclass' (because dtype of label-column == object).
First 10 (of 15) unique label values: [' Exec-managerial', ' Other-service', ' Craft-repair', ' Sales', ' Prof-specialty', ' Protective-serv', ' ?', ' Adm-clerical', ' Machine-op-inspct', ' Tech-support']
If 'multiclass' is not the correct problem_type, please manually specify the problem_type argument in fit() (You may specify problem_type as one of: ['binary', 'multiclass', 'regression'])
Warning: Some classes in the training set have fewer than 10 examples. AutoGluon will only keep 12 out of 15 classes for training and will not try to predict the rare classes. To keep more classes, increase the number of datapoints from these rare classes in the training data or reduce label_count_threshold.
Fraction of data from classes with at least 10 examples that will be kept for training models: 0.978
Train Data Class Count: 12
Using Feature Generators to preprocess the data ...
Fitting AutoMLPipelineFeatureGenerator...
Available Memory: 22093.03 MB
Train Data (Original) Memory Usage: 0.28 MB (0.0% 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.
Stage 2 Generators:
Fitting FillNaFeatureGenerator...
Stage 3 Generators:
Fitting IdentityFeatureGenerator...
Fitting CategoryFeatureGenerator...
Fitting CategoryMemoryMinimizeFeatureGenerator...
Stage 4 Generators:
Fitting DropUniqueFeatureGenerator...
Types of features in original data (raw dtype, special dtypes):
('int', []) : 6 | ['age', 'fnlwgt', 'education-num', 'capital-gain', 'capital-loss', ...]
('object', []) : 8 | ['workclass', 'education', 'marital-status', 'relationship', 'race', ...]
Types of features in processed data (raw dtype, special dtypes):
('category', []) : 6 | ['workclass', 'education', 'marital-status', 'relationship', 'race', ...]
('int', []) : 6 | ['age', 'fnlwgt', 'education-num', 'capital-gain', 'capital-loss', ...]
('int', ['bool']) : 2 | ['sex', 'class']
0.1s = Fit runtime
14 features in original data used to generate 14 features in processed data.
Train Data (Processed) Memory Usage: 0.03 MB (0.0% of available memory)
Data preprocessing and feature engineering runtime = 0.09s ...
AutoGluon will gauge predictive performance using evaluation metric: 'accuracy'
To change this, specify the eval_metric argument of fit()
Fitting 2 L1 models ...
Fitting model: LightGBM_BAG_L1 ... Training model for up to 29.91s of the 29.91s of remaining time.
0.3067 = Validation score (accuracy)
0.68s = Training runtime
0.03s = Validation runtime
Fitting model: NeuralNetMXNet_BAG_L1 ... Training model for up to 29.18s of the 29.18s of remaining time.
0.0961 = Validation score (accuracy)
1.78s = Training runtime
0.1s = Validation runtime
Repeating k-fold bagging: 2/20
Fitting model: LightGBM_BAG_L1 ... Training model for up to 27.26s of the 27.26s of remaining time.
0.3149 = Validation score (accuracy)
1.36s = Training runtime
0.06s = Validation runtime
Fitting model: NeuralNetMXNet_BAG_L1 ... Training model for up to 26.53s of the 26.53s of remaining time.
0.0818 = Validation score (accuracy)
3.47s = Training runtime
0.2s = Validation runtime
Repeating k-fold bagging: 3/20
Fitting model: LightGBM_BAG_L1 ... Training model for up to 24.71s of the 24.71s of remaining time.
0.3292 = Validation score (accuracy)
2.03s = Training runtime
0.09s = Validation runtime
Fitting model: NeuralNetMXNet_BAG_L1 ... Training model for up to 23.98s of the 23.98s of remaining time.
0.1104 = Validation score (accuracy)
5.25s = Training runtime
0.31s = Validation runtime
Repeating k-fold bagging: 4/20
Fitting model: LightGBM_BAG_L1 ... Training model for up to 22.05s of the 22.05s of remaining time.
0.3129 = Validation score (accuracy)
2.72s = Training runtime
0.12s = Validation runtime
Fitting model: NeuralNetMXNet_BAG_L1 ... Training model for up to 21.31s of the 21.31s of remaining time.
0.0961 = Validation score (accuracy)
6.85s = Training runtime
0.41s = Validation runtime
Repeating k-fold bagging: 5/20
Fitting model: LightGBM_BAG_L1 ... Training model for up to 19.56s of the 19.56s of remaining time.
0.3108 = Validation score (accuracy)
3.41s = Training runtime
0.14s = Validation runtime
Fitting model: NeuralNetMXNet_BAG_L1 ... Training model for up to 18.83s of the 18.83s of remaining time.
0.1063 = Validation score (accuracy)
8.54s = Training runtime
0.52s = Validation runtime
Repeating k-fold bagging: 6/20
Fitting model: LightGBM_BAG_L1 ... Training model for up to 17.0s of the 17.0s of remaining time.
0.3088 = Validation score (accuracy)
4.09s = Training runtime
0.17s = Validation runtime
Fitting model: NeuralNetMXNet_BAG_L1 ... Training model for up to 16.27s of the 16.27s of remaining time.
0.0941 = Validation score (accuracy)
10.25s = Training runtime
0.62s = Validation runtime
Repeating k-fold bagging: 7/20
Fitting model: LightGBM_BAG_L1 ... Training model for up to 14.42s of the 14.42s of remaining time.
0.3088 = Validation score (accuracy)
4.78s = Training runtime
0.2s = Validation runtime
Fitting model: NeuralNetMXNet_BAG_L1 ... Training model for up to 13.68s of the 13.68s of remaining time.
0.0961 = Validation score (accuracy)
11.84s = Training runtime
0.73s = Validation runtime
Repeating k-fold bagging: 8/20
Fitting model: LightGBM_BAG_L1 ... Training model for up to 11.94s of the 11.94s of remaining time.
0.3067 = Validation score (accuracy)
5.45s = Training runtime
0.23s = Validation runtime
Fitting model: NeuralNetMXNet_BAG_L1 ... Training model for up to 11.21s of the 11.2s of remaining time.
0.1391 = Validation score (accuracy)
13.44s = Training runtime
0.83s = Validation runtime
Repeating k-fold bagging: 9/20
Fitting model: LightGBM_BAG_L1 ... Training model for up to 9.46s of the 9.46s of remaining time.
0.3047 = Validation score (accuracy)
6.14s = Training runtime
0.26s = Validation runtime
Fitting model: NeuralNetMXNet_BAG_L1 ... Training model for up to 8.73s of the 8.73s of remaining time.
0.1288 = Validation score (accuracy)
15.13s = Training runtime
0.93s = Validation runtime
Repeating k-fold bagging: 10/20
Fitting model: LightGBM_BAG_L1 ... Training model for up to 6.9s of the 6.9s of remaining time.
0.3006 = Validation score (accuracy)
6.81s = Training runtime
0.29s = Validation runtime
Fitting model: NeuralNetMXNet_BAG_L1 ... Training model for up to 6.17s of the 6.17s of remaining time.
0.1329 = Validation score (accuracy)
16.83s = Training runtime
1.04s = Validation runtime
Repeating k-fold bagging: 11/20
Fitting model: LightGBM_BAG_L1 ... Training model for up to 4.33s of the 4.33s of remaining time.
0.2965 = Validation score (accuracy)
7.5s = Training runtime
0.32s = Validation runtime
Fitting model: NeuralNetMXNet_BAG_L1 ... Training model for up to 3.59s of the 3.58s of remaining time.
0.1391 = Validation score (accuracy)
18.73s = Training runtime
1.14s = Validation runtime
Completed 11/20 k-fold bagging repeats ...
Fitting model: WeightedEnsemble_L2 ... Training model for up to 29.91s of the 1.53s of remaining time.
0.3129 = Validation score (accuracy)
0.11s = Training runtime
0.0s = Validation runtime
AutoGluon training complete, total runtime = 28.59s ...
TabularPredictor saved. To load, use: predictor = TabularPredictor.load("agModels-predictOccupation/")
Often stacking/bagging will produce superior accuracy than
hyperparameter-tuning, but you may try combining both techniques (note:
specifying presets='best_quality' in fit() simply sets
auto_stack=True).
Prediction options (inference)¶
Even if you’ve started a new Python session since last calling
fit(), you can still load a previously trained predictor from disk:
predictor = TabularPredictor.load(save_path) # `predictor.path` is another way to get the relative path needed to later load predictor.
Above save_path is the same folder previously passed to
TabularPredictor, in which all the trained models have been saved.
You can train easily models on one machine and deploy them on another.
Simply copy the save_path folder to the new machine and specify its
new path in TabularPredictor.load().
To find out the required feature columns to make predictions, call
predictor.features():
predictor.features()
['age',
'workclass',
'fnlwgt',
'education',
'education-num',
'marital-status',
'relationship',
'race',
'sex',
'capital-gain',
'capital-loss',
'hours-per-week',
'native-country',
'class']
We can make a prediction on an individual example rather than a full dataset:
datapoint = test_data_nolabel.iloc[[0]] # Note: .iloc[0] won't work because it returns pandas Series instead of DataFrame
print(datapoint)
predictor.predict(datapoint)
age workclass fnlwgt education education-num marital-status 5000 49 Private 259087 Some-college 10 Divorced
relationship race sex capital-gain capital-loss 5000 Not-in-family White Female 0 0
hours-per-week native-country class
5000 40 United-States <=50K
5000 Exec-managerial
Name: occupation, dtype: object
To output predicted class probabilities instead of predicted classes, you can use:
predictor.predict_proba(datapoint) # returns a DataFrame that shows which probability corresponds to which class
| ? | Adm-clerical | Armed-Forces | Craft-repair | Exec-managerial | Farming-fishing | Handlers-cleaners | Machine-op-inspct | Other-service | Priv-house-serv | Prof-specialty | Protective-serv | Sales | Tech-support | Transport-moving | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 5000 | 0.072905 | 0.100064 | 0.0 | 0.093995 | 0.113477 | 0.068584 | 0.073102 | 0.075493 | 0.078808 | 0.0 | 0.085901 | 0.0 | 0.089477 | 0.067967 | 0.080228 |
By default, predict() and predict_proba() will utilize the model
that AutoGluon thinks is most accurate, which is usually an ensemble of
many individual models. Here’s how to see which model this is:
predictor.get_model_best()
'WeightedEnsemble_L2'
We can instead specify a particular model to use for predictions (e.g. to reduce inference latency). Note that a ‘model’ in AutoGluon may refer to for example a single Neural Network, a bagged ensemble of many Neural Network copies trained on different training/validation splits, a weighted ensemble that aggregates the predictions of many other models, or a stacker model that operates on predictions output by other models. This is akin to viewing a Random Forest as one ‘model’ when it is in fact an ensemble of many decision trees.
Before deciding which model to use, let’s evaluate all of the models AutoGluon has previously trained on our test data:
predictor.leaderboard(test_data, silent=True)
| model | score_test | score_val | pred_time_test | pred_time_val | fit_time | pred_time_test_marginal | pred_time_val_marginal | fit_time_marginal | stack_level | can_infer | fit_order | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 0 | LightGBM_BAG_L1 | 0.284546 | 0.296524 | 0.917929 | 0.317307 | 7.497480 | 0.917929 | 0.317307 | 7.497480 | 1 | True | 1 |
| 1 | WeightedEnsemble_L2 | 0.278465 | 0.312883 | 23.556926 | 1.458627 | 26.340142 | 0.002352 | 0.000440 | 0.108962 | 2 | True | 3 |
| 2 | NeuralNetMXNet_BAG_L1 | 0.147201 | 0.139059 | 22.636646 | 1.140880 | 18.733700 | 22.636646 | 1.140880 | 18.733700 | 1 | True | 2 |
The leaderboard shows each model’s predictive performance on the test
data (score_test) and validation data (score_val), as well as
the time required to: produce predictions for the test data
(pred_time_val), produce predictions on the validation data
(pred_time_val), and train only this model (fit_time). Below, we
show that a leaderboard can be produced without new data (just uses the
data previously reserved for validation inside fit) and can display
extra information about each model:
predictor.leaderboard(extra_info=True, silent=True)
| model | score_val | pred_time_val | fit_time | pred_time_val_marginal | fit_time_marginal | stack_level | can_infer | fit_order | num_features | ... | child_model_type | hyperparameters | hyperparameters_fit | ag_args_fit | features | child_hyperparameters | child_hyperparameters_fit | child_ag_args_fit | ancestors | descendants | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 0 | WeightedEnsemble_L2 | 0.312883 | 1.458627 | 26.340142 | 0.000440 | 0.108962 | 2 | True | 3 | 24 | ... | GreedyWeightedEnsembleModel | {'use_orig_features': False, 'max_base_models'... | {} | {'max_memory_usage_ratio': 1.0, 'max_time_limi... | [LightGBM_BAG_L1_5, NeuralNetMXNet_BAG_L1_5, N... | {'ensemble_size': 100} | {'ensemble_size': 5} | {'max_memory_usage_ratio': 1.0, 'max_time_limi... | [NeuralNetMXNet_BAG_L1, LightGBM_BAG_L1] | [] |
| 1 | LightGBM_BAG_L1 | 0.296524 | 0.317307 | 7.497480 | 0.317307 | 7.497480 | 1 | True | 1 | 14 | ... | LGBModel | {'use_orig_features': True, 'max_base_models':... | {} | {'max_memory_usage_ratio': 1.0, 'max_time_limi... | [relationship, fnlwgt, class, age, sex, workcl... | {'num_boost_round': 20, 'num_threads': -1, 'le... | {'num_boost_round': 12} | {'max_memory_usage_ratio': 1.0, 'max_time_limi... | [] | [WeightedEnsemble_L2] |
| 2 | NeuralNetMXNet_BAG_L1 | 0.139059 | 1.140880 | 18.733700 | 1.140880 | 18.733700 | 1 | True | 2 | 14 | ... | TabularNeuralNetModel | {'use_orig_features': True, 'max_base_models':... | {} | {'max_memory_usage_ratio': 1.0, 'max_time_limi... | [relationship, fnlwgt, class, age, sex, workcl... | {'num_epochs': 2, 'epochs_wo_improve': 20, 'se... | {'num_epochs': 2} | {'max_memory_usage_ratio': 1.0, 'max_time_limi... | [] | [WeightedEnsemble_L2] |
3 rows × 29 columns
The expanded leaderboard shows properties like how many features are
used by each model (num_features), which other models are ancestors
whose predictions are required inputs for each model (ancestors),
and how much memory each model and all its ancestors would occupy if
simultaneously persisted (memory_size_w_ancestors). See the
leaderboard
documentation
for full details.
To show scores for other metrics, you can specify the extra_metrics
argument when passing in test_data:
predictor.leaderboard(test_data, extra_metrics=['accuracy', 'balanced_accuracy', 'log_loss'], silent=True)
| model | score_test | accuracy | balanced_accuracy | log_loss | score_val | pred_time_test | pred_time_val | fit_time | pred_time_test_marginal | pred_time_val_marginal | fit_time_marginal | stack_level | can_infer | fit_order | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 0 | LightGBM_BAG_L1 | 0.284546 | 0.284546 | 0.178211 | -11.748757 | 0.296524 | 0.916015 | 0.317307 | 7.497480 | 0.916015 | 0.317307 | 7.497480 | 1 | True | 1 |
| 1 | WeightedEnsemble_L2 | 0.278465 | 0.278465 | 0.172589 | -11.740380 | 0.312883 | 23.357587 | 1.458627 | 26.340142 | 0.002362 | 0.000440 | 0.108962 | 2 | True | 3 |
| 2 | NeuralNetMXNet_BAG_L1 | 0.147201 | 0.147201 | 0.077969 | -11.765823 | 0.139059 | 22.439210 | 1.140880 | 18.733700 | 22.439210 | 1.140880 | 18.733700 | 1 | True | 2 |
Notice that log_loss scores are negative. This is because metrics in
AutoGluon are always shown in higher_is_better form. This means that
metrics such as log_loss and root_mean_squared_error will have
their signs FLIPPED, and values will be negative. This is necessary to
avoid the user needing to know the metric to understand if higher is
better when looking at leaderboard.
One additional caviat: It is possible that log_loss values can be
-inf when computed via extra_metrics. This is because the models
were not optimized with log_loss in mind during training and may
have prediction probabilities giving a class 0 (particularly common
with K-Nearest-Neighbors models). Because log_loss gives infinite
error when the correct class was given 0 probability, this results
in a score of -inf. It is therefore recommended that log_loss
should not be used as a secondary metric to determine model quality.
Either use log_loss as the eval_metric or avoid it altogether.
Here’s how to specify a particular model to use for prediction instead of AutoGluon’s default model-choice:
i = 0 # index of model to use
model_to_use = predictor.get_model_names()[i]
model_pred = predictor.predict(datapoint, model=model_to_use)
print("Prediction from %s model: %s" % (model_to_use, model_pred.iloc[0]))
Prediction from LightGBM_BAG_L1 model: Exec-managerial
We can easily access various information about the trained predictor or a particular model:
all_models = predictor.get_model_names()
model_to_use = all_models[i]
specific_model = predictor._trainer.load_model(model_to_use)
# Objects defined below are dicts of various information (not printed here as they are quite large):
model_info = specific_model.get_info()
predictor_information = predictor.info()
The predictor also remembers what metric predictions should be
evaluated with, which can be done with ground truth labels as follows:
y_pred_proba = predictor.predict_proba(test_data_nolabel)
perf = predictor.evaluate_predictions(y_true=y_test, y_pred=y_pred_proba)
Evaluation: accuracy on test data: 0.2784650870203397
Evaluations on test data:
{
"accuracy": 0.2784650870203397,
"balanced_accuracy": 0.17258912992867856,
"mcc": 0.18580044015660524
}
Since the label columns remains in the test_data DataFrame, we can
instead use the shorthand:
perf = predictor.evaluate(test_data)
Evaluation: accuracy on test data: 0.2784650870203397
Evaluations on test data:
{
"accuracy": 0.2784650870203397,
"balanced_accuracy": 0.17258912992867856,
"mcc": 0.18580044015660524
}
Interpretability (feature importance)¶
To better understand our trained predictor, we can estimate the overall importance of each feature:
predictor.feature_importance(test_data)
Computing feature importance via permutation shuffling for 14 features using 1000 rows with 3 shuffle sets...
272.37s = Expected runtime (90.79s per shuffle set)
218.16s = Actual runtime (Completed 3 of 3 shuffle sets)
| importance | stddev | p_value | n | p99_high | p99_low | |
|---|---|---|---|---|---|---|
| education-num | 0.066333 | 0.010214 | 0.003906 | 3 | 0.124863 | 0.007804 |
| workclass | 0.040667 | 0.004933 | 0.002434 | 3 | 0.068933 | 0.012401 |
| sex | 0.032000 | 0.013892 | 0.028732 | 3 | 0.111605 | -0.047605 |
| hours-per-week | 0.023667 | 0.007234 | 0.014881 | 3 | 0.065119 | -0.017786 |
| age | 0.014333 | 0.018148 | 0.152366 | 3 | 0.118321 | -0.089654 |
| class | 0.002667 | 0.008083 | 0.312683 | 3 | 0.048983 | -0.043649 |
| fnlwgt | 0.001333 | 0.013650 | 0.440609 | 3 | 0.079552 | -0.076885 |
| marital-status | 0.000333 | 0.001528 | 0.370901 | 3 | 0.009086 | -0.008420 |
| race | 0.000333 | 0.000577 | 0.211325 | 3 | 0.003642 | -0.002975 |
| education | 0.000000 | 0.003000 | 0.500000 | 3 | 0.017190 | -0.017190 |
| native-country | 0.000000 | 0.000000 | 0.500000 | 3 | 0.000000 | 0.000000 |
| capital-loss | -0.000333 | 0.000577 | 0.788675 | 3 | 0.002975 | -0.003642 |
| capital-gain | -0.001000 | 0.001000 | 0.887298 | 3 | 0.004730 | -0.006730 |
| relationship | -0.001333 | 0.004041 | 0.687317 | 3 | 0.021825 | -0.024491 |
Computed via
permutation-shuffling, these
feature importance scores quantify the drop in predictive performance
(of the already trained predictor) when one column’s values are randomly
shuffled across rows. The top features in this list contribute most to
AutoGluon’s accuracy (for predicting when/if a patient will be
readmitted to the hospital). Features with non-positive importance score
hardly contribute to the predictor’s accuracy, or may even be actively
harmful to include in the data (consider removing these features from
your data and calling fit again). These scores facilitate
interpretability of the predictor’s global behavior (which features it
relies on for all predictions) rather than local
explanations
that only rationalize one particular prediction.
Accelerating inference¶
We describe multiple ways to reduce the time it takes for AutoGluon to produce predictions.
Keeping models in memory¶
By default, AutoGluon loads models into memory one at a time and only when they are needed for prediction. This strategy is robust for large stacked/bagged ensembles, but leads to slower prediction times. If you plan to repeatedly make predictions (e.g. on new datapoints one at a time rather than one large test dataset), you can first specify that all models required for inference should be loaded into memory as follows:
predictor.persist_models()
num_test = 20
preds = np.array(['']*num_test, dtype='object')
for i in range(num_test):
datapoint = test_data_nolabel.iloc[[i]]
pred_numpy = predictor.predict(datapoint, as_pandas=False)
preds[i] = pred_numpy[0]
perf = predictor.evaluate_predictions(y_test[:num_test], preds, auxiliary_metrics=True)
print("Predictions: ", preds)
predictor.unpersist_models() # free memory by clearing models, future predict() calls will load models from disk
Persisting 3 models in memory. Models will require 0.27% of memory.
Evaluation: accuracy on test data: 0.25
Evaluations on test data:
{
"accuracy": 0.25,
"balanced_accuracy": 0.3208333333333336,
"mcc": 0.12935842095105549
}
Unpersisted 3 models: ['WeightedEnsemble_L2', 'NeuralNetMXNet_BAG_L1', 'LightGBM_BAG_L1']
Predictions: [' Exec-managerial' ' Exec-managerial' ' Craft-repair' ' Adm-clerical'
' Sales' ' Exec-managerial' ' Exec-managerial' ' Sales'
' Exec-managerial' ' Adm-clerical' ' Other-service' ' Exec-managerial'
' Exec-managerial' ' Exec-managerial' ' Adm-clerical' ' ?'
' Craft-repair' ' Craft-repair' ' Exec-managerial' ' Craft-repair']
['WeightedEnsemble_L2', 'NeuralNetMXNet_BAG_L1', 'LightGBM_BAG_L1']
You can alternatively specify a particular model to persist via the
models argument of persist_models(), or simply set
models='all' to simultaneously load every single model that was
trained during fit.
Using smaller ensemble or faster model for prediction¶
Without having to retrain any models, one can construct alternative ensembles that aggregate individual models’ predictions with different weighting schemes. These ensembles become smaller (and hence faster for prediction) if they assign nonzero weight to less models. You can produce a wide variety of ensembles with different accuracy-speed tradeoffs like this:
additional_ensembles = predictor.fit_weighted_ensemble(expand_pareto_frontier=True)
print("Alternative ensembles you can use for prediction:", additional_ensembles)
predictor.leaderboard(only_pareto_frontier=True, silent=True)
Fitting model: WeightedEnsemble_L2Best ...
0.3129 = Validation score (accuracy)
0.11s = Training runtime
0.0s = Validation runtime
Alternative ensembles you can use for prediction: ['WeightedEnsemble_L2Best']
| model | score_val | pred_time_val | fit_time | pred_time_val_marginal | fit_time_marginal | stack_level | can_infer | fit_order | |
|---|---|---|---|---|---|---|---|---|---|
| 0 | WeightedEnsemble_L2Best | 0.312883 | 1.458586 | 26.339494 | 0.000399 | 0.108314 | 2 | True | 4 |
| 1 | LightGBM_BAG_L1 | 0.296524 | 0.317307 | 7.497480 | 0.317307 | 7.497480 | 1 | True | 1 |
The resulting leaderboard will contain the most accurate model for a given inference-latency. You can select whichever model exhibits acceptable latency from the leaderboard and use it for prediction.
model_for_prediction = additional_ensembles[0]
predictions = predictor.predict(test_data, model=model_for_prediction)
predictor.delete_models(models_to_delete=additional_ensembles, dry_run=False) # delete these extra models so they don't affect rest of tutorial
Deleting model WeightedEnsemble_L2Best. All files under agModels-predictOccupation/models/WeightedEnsemble_L2Best/ will be removed.
Collapsing bagged ensembles via refit_full¶
For an ensemble predictor trained with bagging (as done above), recall
there ~10 bagged copies of each individual model trained on different
train/validation folds. We can collapse this bag of ~10 models into a
single model that’s fit to the full dataset, which can greatly reduce
its memory/latency requirements (but may also reduce accuracy). Below we
refit such a model for each original model but you can alternatively do
this for just a particular model by specifying the model argument of
refit_full().
refit_model_map = predictor.refit_full()
print("Name of each refit-full model corresponding to a previous bagged ensemble:")
print(refit_model_map)
predictor.leaderboard(test_data, silent=True)
Fitting 1 L1 models ...
Fitting model: LightGBM_BAG_L1_FULL ...
0.12s = Training runtime
Fitting 1 L1 models ...
Fitting model: NeuralNetMXNet_BAG_L1_FULL ...
0.3s = Training runtime
Fitting model: WeightedEnsemble_L2_FULL ...
0.3129 = Validation score (accuracy)
0.0s = Training runtime
0.0s = Validation runtime
Name of each refit-full model corresponding to a previous bagged ensemble:
{'LightGBM_BAG_L1': 'LightGBM_BAG_L1_FULL', 'NeuralNetMXNet_BAG_L1': 'NeuralNetMXNet_BAG_L1_FULL', 'WeightedEnsemble_L2': 'WeightedEnsemble_L2_FULL'}
| model | score_test | score_val | pred_time_test | pred_time_val | fit_time | pred_time_test_marginal | pred_time_val_marginal | fit_time_marginal | stack_level | can_infer | fit_order | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 0 | LightGBM_BAG_L1 | 0.284546 | 0.296524 | 0.913584 | 0.317307 | 7.497480 | 0.913584 | 0.317307 | 7.497480 | 1 | True | 1 |
| 1 | WeightedEnsemble_L2 | 0.278465 | 0.312883 | 23.571955 | 1.458627 | 26.340142 | 0.002269 | 0.000440 | 0.108962 | 2 | True | 3 |
| 2 | WeightedEnsemble_L2_FULL | 0.272594 | NaN | 0.382753 | NaN | 0.422620 | 0.001819 | 0.000499 | 0.000312 | 2 | True | 6 |
| 3 | LightGBM_BAG_L1_FULL | 0.269868 | NaN | 0.017992 | NaN | 0.122259 | 0.017992 | NaN | 0.122259 | 1 | True | 4 |
| 4 | NeuralNetMXNet_BAG_L1 | 0.147201 | 0.139059 | 22.656102 | 1.140880 | 18.733700 | 22.656102 | 1.140880 | 18.733700 | 1 | True | 2 |
| 5 | NeuralNetMXNet_BAG_L1_FULL | 0.052632 | NaN | 0.362942 | NaN | 0.300049 | 0.362942 | NaN | 0.300049 | 1 | True | 5 |
This adds the refit-full models to the leaderboard and we can opt to use
any of them for prediction just like any other model. Note
pred_time_test and pred_time_val list the time taken to produce
predictions with each model (in seconds) on the test/validation data.
Since the refit-full models were trained using all of the data, there is
no internal validation score (score_val) available for them. You can
also call refit_full() with non-bagged models to refit the same
models to your full dataset (there won’t be memory/latency gains in this
case but test accuracy may improve).
Model distillation¶
While computationally-favorable, single individual models will usually
have lower accuracy than weighted/stacked/bagged ensembles. Model
Distillation offers one way to
retain the computational benefits of a single model, while enjoying some
of the accuracy-boost that comes with ensembling. The idea is to train
the individual model (which we can call the student) to mimic the
predictions of the full stack ensemble (the teacher). Like
refit_full(), the distill() function will produce additional
models we can opt to use for prediction.
student_models = predictor.distill(time_limit=30) # specify much longer time limit in real applications
print(student_models)
preds_student = predictor.predict(test_data_nolabel, model=student_models[0])
print(f"predictions from {student_models[0]}:", list(preds_student)[:5])
predictor.leaderboard(test_data)
Distilling with teacher='WeightedEnsemble_L2', teacher_preds=soft, augment_method=spunge ...
SPUNGE: Augmenting training data with 1955 synthetic samples for distillation...
Distilling with each of these student models: ['LightGBM_DSTL', 'NeuralNetMXNet_DSTL', 'RandomForestMSE_DSTL', 'CatBoost_DSTL']
Fitting 4 L1 models ...
Fitting model: LightGBM_DSTL ... Training model for up to 30.0s of the 30.0s of remaining time.
Note: model has different eval_metric than default.
-2.1722 = Validation score (soft_log_loss)
4.14s = Training runtime
0.01s = Validation runtime
Fitting model: NeuralNetMXNet_DSTL ... Training model for up to 25.74s of the 25.74s of remaining time.
Note: model has different eval_metric than default.
-2.2078 = Validation score (soft_log_loss)
7.35s = Training runtime
0.02s = Validation runtime
Fitting model: RandomForestMSE_DSTL ... Training model for up to 18.36s of the 18.36s of remaining time.
Note: model has different eval_metric than default.
-2.175 = Validation score (soft_log_loss)
1.02s = Training runtime
0.11s = Validation runtime
Fitting model: CatBoost_DSTL ... Training model for up to 17.09s of the 17.09s of remaining time.
Warning: Exception caused CatBoost_DSTL to fail during training (ImportError)... Skipping this model.
import catboost_dev failed (needed for distillation with CatBoost models). Make sure you can import catboost and then run: 'pip install catboost-dev'.Detailed info: No module named 'catboost_dev'
Completed 1/20 k-fold bagging repeats ...
Distilling with each of these student models: ['WeightedEnsemble_L2_DSTL']
Fitting model: WeightedEnsemble_L2_DSTL ... Training model for up to 30.0s of the 16.73s of remaining time.
Note: model has different eval_metric than default.
-2.16 = Validation score (soft_log_loss)
0.29s = Training runtime
0.0s = Validation runtime
Distilled model leaderboard:
model score_val pred_time_val fit_time pred_time_val_marginal fit_time_marginal stack_level can_infer fit_order
0 NeuralNetMXNet_DSTL 0.357143 0.019804 7.354267 0.019804 7.354267 1 True 8
1 WeightedEnsemble_L2_DSTL 0.346939 0.122399 5.442422 0.001168 0.285002 2 True 10
2 LightGBM_DSTL 0.326531 0.014075 4.142191 0.014075 4.142191 1 True 7
3 RandomForestMSE_DSTL 0.306122 0.107156 1.015228 0.107156 1.015228 1 True 9
['LightGBM_DSTL', 'NeuralNetMXNet_DSTL', 'RandomForestMSE_DSTL', 'WeightedEnsemble_L2_DSTL']
predictions from LightGBM_DSTL: [' Exec-managerial', ' Exec-managerial', ' Craft-repair', ' Sales', ' Exec-managerial']
model score_test score_val pred_time_test pred_time_val fit_time pred_time_test_marginal pred_time_val_marginal fit_time_marginal stack_level can_infer fit_order
0 WeightedEnsemble_L2_DSTL 0.288320 0.346939 0.536227 0.122399 5.442422 0.002203 0.001168 0.285002 2 True 10
1 LightGBM_DSTL 0.285175 0.326531 0.354778 0.014075 4.142191 0.354778 0.014075 4.142191 1 True 7
2 NeuralNetMXNet_DSTL 0.284756 0.357143 0.346866 0.019804 7.354267 0.346866 0.019804 7.354267 1 True 8
3 LightGBM_BAG_L1 0.284546 0.296524 0.921535 0.317307 7.497480 0.921535 0.317307 7.497480 1 True 1
4 RandomForestMSE_DSTL 0.279933 0.306122 0.179247 0.107156 1.015228 0.179247 0.107156 1.015228 1 True 9
5 WeightedEnsemble_L2 0.278465 0.312883 23.650895 1.458627 26.340142 0.002828 0.000440 0.108962 2 True 3
6 WeightedEnsemble_L2_FULL 0.272594 NaN 0.447413 NaN 0.422620 0.001762 0.000499 0.000312 2 True 6
7 LightGBM_BAG_L1_FULL 0.269868 NaN 0.018223 NaN 0.122259 0.018223 NaN 0.122259 1 True 4
8 NeuralNetMXNet_BAG_L1 0.147201 0.139059 22.726533 1.140880 18.733700 22.726533 1.140880 18.733700 1 True 2
9 NeuralNetMXNet_BAG_L1_FULL 0.052632 NaN 0.427428 NaN 0.300049 0.427428 NaN 0.300049 1 True 5
| model | score_test | score_val | pred_time_test | pred_time_val | fit_time | pred_time_test_marginal | pred_time_val_marginal | fit_time_marginal | stack_level | can_infer | fit_order | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 0 | WeightedEnsemble_L2_DSTL | 0.288320 | 0.346939 | 0.536227 | 0.122399 | 5.442422 | 0.002203 | 0.001168 | 0.285002 | 2 | True | 10 |
| 1 | LightGBM_DSTL | 0.285175 | 0.326531 | 0.354778 | 0.014075 | 4.142191 | 0.354778 | 0.014075 | 4.142191 | 1 | True | 7 |
| 2 | NeuralNetMXNet_DSTL | 0.284756 | 0.357143 | 0.346866 | 0.019804 | 7.354267 | 0.346866 | 0.019804 | 7.354267 | 1 | True | 8 |
| 3 | LightGBM_BAG_L1 | 0.284546 | 0.296524 | 0.921535 | 0.317307 | 7.497480 | 0.921535 | 0.317307 | 7.497480 | 1 | True | 1 |
| 4 | RandomForestMSE_DSTL | 0.279933 | 0.306122 | 0.179247 | 0.107156 | 1.015228 | 0.179247 | 0.107156 | 1.015228 | 1 | True | 9 |
| 5 | WeightedEnsemble_L2 | 0.278465 | 0.312883 | 23.650895 | 1.458627 | 26.340142 | 0.002828 | 0.000440 | 0.108962 | 2 | True | 3 |
| 6 | WeightedEnsemble_L2_FULL | 0.272594 | NaN | 0.447413 | NaN | 0.422620 | 0.001762 | 0.000499 | 0.000312 | 2 | True | 6 |
| 7 | LightGBM_BAG_L1_FULL | 0.269868 | NaN | 0.018223 | NaN | 0.122259 | 0.018223 | NaN | 0.122259 | 1 | True | 4 |
| 8 | NeuralNetMXNet_BAG_L1 | 0.147201 | 0.139059 | 22.726533 | 1.140880 | 18.733700 | 22.726533 | 1.140880 | 18.733700 | 1 | True | 2 |
| 9 | NeuralNetMXNet_BAG_L1_FULL | 0.052632 | NaN | 0.427428 | NaN | 0.300049 | 0.427428 | NaN | 0.300049 | 1 | True | 5 |
Faster presets or hyperparameters¶
Instead of trying to speed up a cumbersome trained model at prediction
time, if you know inference latency or memory will be an issue at the
outset, then you can adjust the training process accordingly to ensure
fit() does not produce unwieldy models.
One option is to specify more lightweight presets:
presets = ['good_quality_faster_inference_only_refit', 'optimize_for_deployment']
predictor_light = TabularPredictor(label=label, eval_metric=metric).fit(train_data, presets=presets, time_limit=30)
No path specified. Models will be saved in: "AutogluonModels/ag-20210831_221437/"
Presets specified: ['good_quality_faster_inference_only_refit', 'optimize_for_deployment']
Beginning AutoGluon training ... Time limit = 30s
AutoGluon will save models to "AutogluonModels/ag-20210831_221437/"
AutoGluon Version: 0.3.1b20210831
Train Data Rows: 500
Train Data Columns: 14
Preprocessing data ...
AutoGluon infers your prediction problem is: 'multiclass' (because dtype of label-column == object).
First 10 (of 15) unique label values: [' Exec-managerial', ' Other-service', ' Craft-repair', ' Sales', ' Prof-specialty', ' Protective-serv', ' ?', ' Adm-clerical', ' Machine-op-inspct', ' Tech-support']
If 'multiclass' is not the correct problem_type, please manually specify the problem_type argument in fit() (You may specify problem_type as one of: ['binary', 'multiclass', 'regression'])
Warning: Some classes in the training set have fewer than 10 examples. AutoGluon will only keep 12 out of 15 classes for training and will not try to predict the rare classes. To keep more classes, increase the number of datapoints from these rare classes in the training data or reduce label_count_threshold.
Fraction of data from classes with at least 10 examples that will be kept for training models: 0.978
Train Data Class Count: 12
Using Feature Generators to preprocess the data ...
Fitting AutoMLPipelineFeatureGenerator...
Available Memory: 21664.1 MB
Train Data (Original) Memory Usage: 0.28 MB (0.0% 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.
Stage 2 Generators:
Fitting FillNaFeatureGenerator...
Stage 3 Generators:
Fitting IdentityFeatureGenerator...
Fitting CategoryFeatureGenerator...
Fitting CategoryMemoryMinimizeFeatureGenerator...
Stage 4 Generators:
Fitting DropUniqueFeatureGenerator...
Types of features in original data (raw dtype, special dtypes):
('int', []) : 6 | ['age', 'fnlwgt', 'education-num', 'capital-gain', 'capital-loss', ...]
('object', []) : 8 | ['workclass', 'education', 'marital-status', 'relationship', 'race', ...]
Types of features in processed data (raw dtype, special dtypes):
('category', []) : 6 | ['workclass', 'education', 'marital-status', 'relationship', 'race', ...]
('int', []) : 6 | ['age', 'fnlwgt', 'education-num', 'capital-gain', 'capital-loss', ...]
('int', ['bool']) : 2 | ['sex', 'class']
0.1s = Fit runtime
14 features in original data used to generate 14 features in processed data.
Train Data (Processed) Memory Usage: 0.03 MB (0.0% of available memory)
Data preprocessing and feature engineering runtime = 0.08s ...
AutoGluon will gauge predictive performance using evaluation metric: 'accuracy'
To change this, specify the eval_metric argument of fit()
Fitting 11 L1 models ...
Fitting model: NeuralNetFastAI_BAG_L1 ... Training model for up to 29.92s of the 29.92s of remaining time.
Ran out of time, stopping training early. (Stopping on epoch 0)
No improvement since epoch 7: early stopping
No improvement since epoch 4: early stopping
No improvement since epoch 9: early stopping
0.229 = Validation score (accuracy)
5.36s = Training runtime
0.08s = Validation runtime
Fitting model: LightGBMXT_BAG_L1 ... Training model for up to 24.47s of the 24.47s of remaining time.
0.3476 = Validation score (accuracy)
2.73s = Training runtime
0.04s = Validation runtime
Fitting model: LightGBM_BAG_L1 ... Training model for up to 21.69s of the 21.69s of remaining time.
0.3395 = Validation score (accuracy)
3.35s = Training runtime
0.03s = Validation runtime
Fitting model: RandomForestGini_BAG_L1 ... Training model for up to 18.3s of the 18.29s of remaining time.
0.2986 = Validation score (accuracy)
0.61s = Training runtime
0.1s = Validation runtime
Fitting model: RandomForestEntr_BAG_L1 ... Training model for up to 17.58s of the 17.58s of remaining time.
0.3067 = Validation score (accuracy)
0.61s = Training runtime
0.1s = Validation runtime
Fitting model: CatBoost_BAG_L1 ... Training model for up to 16.87s of the 16.87s of remaining time.
Time limit exceeded... Skipping CatBoost_BAG_L1.
Fitting model: ExtraTreesGini_BAG_L1 ... Training model for up to 12.98s of the 12.98s of remaining time.
0.3027 = Validation score (accuracy)
0.71s = Training runtime
0.1s = Validation runtime
Fitting model: ExtraTreesEntr_BAG_L1 ... Training model for up to 12.16s of the 12.16s of remaining time.
0.319 = Validation score (accuracy)
0.61s = Training runtime
0.1s = Validation runtime
Fitting model: XGBoost_BAG_L1 ... Training model for up to 11.45s of the 11.45s of remaining time.
0.3476 = Validation score (accuracy)
4.33s = Training runtime
0.04s = Validation runtime
Fitting model: NeuralNetMXNet_BAG_L1 ... Training model for up to 7.05s of the 7.05s of remaining time.
Ran out of time, stopping training early. (Stopping on epoch 14)
Ran out of time, stopping training early. (Stopping on epoch 15)
Ran out of time, stopping training early. (Stopping on epoch 15)
Ran out of time, stopping training early. (Stopping on epoch 17)
Ran out of time, stopping training early. (Stopping on epoch 21)
0.18 = Validation score (accuracy)
6.75s = Training runtime
0.11s = Validation runtime
Fitting model: LightGBMLarge_BAG_L1 ... Training model for up to 0.18s of the 0.18s of remaining time.
Ran out of time, early stopping on iteration 1. Best iteration is:
[1] train_set's multi_error: 0.759591 valid_set's multi_error: 0.816327
Time limit exceeded... Skipping LightGBMLarge_BAG_L1.
Completed 1/20 k-fold bagging repeats ...
Fitting model: WeightedEnsemble_L2 ... Training model for up to 29.92s of the -0.0s of remaining time.
0.3476 = Validation score (accuracy)
0.28s = Training runtime
0.0s = Validation runtime
AutoGluon training complete, total runtime = 30.29s ...
Fitting 1 L1 models ...
Fitting model: LightGBMXT_BAG_L1_FULL ...
0.3s = Training runtime
Deleting model NeuralNetFastAI_BAG_L1. All files under AutogluonModels/ag-20210831_221437/models/NeuralNetFastAI_BAG_L1/ will be removed.
Deleting model LightGBMXT_BAG_L1. All files under AutogluonModels/ag-20210831_221437/models/LightGBMXT_BAG_L1/ will be removed.
Deleting model LightGBM_BAG_L1. All files under AutogluonModels/ag-20210831_221437/models/LightGBM_BAG_L1/ will be removed.
Deleting model RandomForestGini_BAG_L1. All files under AutogluonModels/ag-20210831_221437/models/RandomForestGini_BAG_L1/ will be removed.
Deleting model RandomForestEntr_BAG_L1. All files under AutogluonModels/ag-20210831_221437/models/RandomForestEntr_BAG_L1/ will be removed.
Deleting model ExtraTreesGini_BAG_L1. All files under AutogluonModels/ag-20210831_221437/models/ExtraTreesGini_BAG_L1/ will be removed.
Deleting model ExtraTreesEntr_BAG_L1. All files under AutogluonModels/ag-20210831_221437/models/ExtraTreesEntr_BAG_L1/ will be removed.
Deleting model XGBoost_BAG_L1. All files under AutogluonModels/ag-20210831_221437/models/XGBoost_BAG_L1/ will be removed.
Deleting model NeuralNetMXNet_BAG_L1. All files under AutogluonModels/ag-20210831_221437/models/NeuralNetMXNet_BAG_L1/ will be removed.
Deleting model WeightedEnsemble_L2. All files under AutogluonModels/ag-20210831_221437/models/WeightedEnsemble_L2/ will be removed.
TabularPredictor saved. To load, use: predictor = TabularPredictor.load("AutogluonModels/ag-20210831_221437/")
Another option is to specify more lightweight hyperparameters:
predictor_light = TabularPredictor(label=label, eval_metric=metric).fit(train_data, hyperparameters='very_light', time_limit=30)
No path specified. Models will be saved in: "AutogluonModels/ag-20210831_221508/"
Beginning AutoGluon training ... Time limit = 30s
AutoGluon will save models to "AutogluonModels/ag-20210831_221508/"
AutoGluon Version: 0.3.1b20210831
Train Data Rows: 500
Train Data Columns: 14
Preprocessing data ...
AutoGluon infers your prediction problem is: 'multiclass' (because dtype of label-column == object).
First 10 (of 15) unique label values: [' Exec-managerial', ' Other-service', ' Craft-repair', ' Sales', ' Prof-specialty', ' Protective-serv', ' ?', ' Adm-clerical', ' Machine-op-inspct', ' Tech-support']
If 'multiclass' is not the correct problem_type, please manually specify the problem_type argument in fit() (You may specify problem_type as one of: ['binary', 'multiclass', 'regression'])
Warning: Some classes in the training set have fewer than 10 examples. AutoGluon will only keep 12 out of 15 classes for training and will not try to predict the rare classes. To keep more classes, increase the number of datapoints from these rare classes in the training data or reduce label_count_threshold.
Fraction of data from classes with at least 10 examples that will be kept for training models: 0.978
Train Data Class Count: 12
Using Feature Generators to preprocess the data ...
Fitting AutoMLPipelineFeatureGenerator...
Available Memory: 19507.87 MB
Train Data (Original) Memory Usage: 0.28 MB (0.0% 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.
Stage 2 Generators:
Fitting FillNaFeatureGenerator...
Stage 3 Generators:
Fitting IdentityFeatureGenerator...
Fitting CategoryFeatureGenerator...
Fitting CategoryMemoryMinimizeFeatureGenerator...
Stage 4 Generators:
Fitting DropUniqueFeatureGenerator...
Types of features in original data (raw dtype, special dtypes):
('int', []) : 6 | ['age', 'fnlwgt', 'education-num', 'capital-gain', 'capital-loss', ...]
('object', []) : 8 | ['workclass', 'education', 'marital-status', 'relationship', 'race', ...]
Types of features in processed data (raw dtype, special dtypes):
('category', []) : 6 | ['workclass', 'education', 'marital-status', 'relationship', 'race', ...]
('int', []) : 6 | ['age', 'fnlwgt', 'education-num', 'capital-gain', 'capital-loss', ...]
('int', ['bool']) : 2 | ['sex', 'class']
0.1s = Fit runtime
14 features in original data used to generate 14 features in processed data.
Train Data (Processed) Memory Usage: 0.03 MB (0.0% of available memory)
Data preprocessing and feature engineering runtime = 0.09s ...
AutoGluon will gauge predictive performance using evaluation metric: 'accuracy'
To change this, specify the eval_metric argument of fit()
Automatically generating train/validation split with holdout_frac=0.2, Train Rows: 391, Val Rows: 98
Fitting 6 L1 models ...
Fitting model: NeuralNetFastAI ... Training model for up to 29.91s of the 29.91s of remaining time.
0.3367 = Validation score (accuracy)
0.55s = Training runtime
0.02s = Validation runtime
Fitting model: LightGBM ... Training model for up to 29.34s of the 29.34s of remaining time.
0.3673 = Validation score (accuracy)
0.7s = Training runtime
0.01s = Validation runtime
Fitting model: LightGBMXT ... Training model for up to 28.62s of the 28.62s of remaining time.
0.3571 = Validation score (accuracy)
0.6s = Training runtime
0.01s = Validation runtime
Fitting model: CatBoost ... Training model for up to 27.98s of the 27.98s of remaining time.
0.3571 = Validation score (accuracy)
5.29s = Training runtime
0.01s = Validation runtime
Fitting model: XGBoost ... Training model for up to 22.68s of the 22.68s of remaining time.
0.398 = Validation score (accuracy)
0.76s = Training runtime
0.01s = Validation runtime
Fitting model: NeuralNetMXNet ... Training model for up to 21.82s of the 21.82s of remaining time.
0.3265 = Validation score (accuracy)
3.48s = Training runtime
0.02s = Validation runtime
Fitting model: WeightedEnsemble_L2 ... Training model for up to 29.91s of the 18.05s of remaining time.
0.4082 = Validation score (accuracy)
0.14s = Training runtime
0.0s = Validation runtime
AutoGluon training complete, total runtime = 12.1s ...
TabularPredictor saved. To load, use: predictor = TabularPredictor.load("AutogluonModels/ag-20210831_221508/")
Here you can set hyperparameters to either ‘light’, ‘very_light’,
or ‘toy’ to obtain progressively smaller (but less accurate) models and
predictors. Advanced users may instead try manually specifying
particular models’ hyperparameters in order to make them faster/smaller.
Finally, you may also exclude specific unwieldy models from being trained at all. Below we exclude models that tend to be slower (K Nearest Neighbors, Neural Network, models with custom larger-than-default hyperparameters):
excluded_model_types = ['KNN', 'NN', 'custom']
predictor_light = TabularPredictor(label=label, eval_metric=metric).fit(train_data, excluded_model_types=excluded_model_types, time_limit=30)
No path specified. Models will be saved in: "AutogluonModels/ag-20210831_221520/"
Beginning AutoGluon training ... Time limit = 30s
AutoGluon will save models to "AutogluonModels/ag-20210831_221520/"
AutoGluon Version: 0.3.1b20210831
Train Data Rows: 500
Train Data Columns: 14
Preprocessing data ...
AutoGluon infers your prediction problem is: 'multiclass' (because dtype of label-column == object).
First 10 (of 15) unique label values: [' Exec-managerial', ' Other-service', ' Craft-repair', ' Sales', ' Prof-specialty', ' Protective-serv', ' ?', ' Adm-clerical', ' Machine-op-inspct', ' Tech-support']
If 'multiclass' is not the correct problem_type, please manually specify the problem_type argument in fit() (You may specify problem_type as one of: ['binary', 'multiclass', 'regression'])
Warning: Some classes in the training set have fewer than 10 examples. AutoGluon will only keep 12 out of 15 classes for training and will not try to predict the rare classes. To keep more classes, increase the number of datapoints from these rare classes in the training data or reduce label_count_threshold.
Fraction of data from classes with at least 10 examples that will be kept for training models: 0.978
Train Data Class Count: 12
Using Feature Generators to preprocess the data ...
Fitting AutoMLPipelineFeatureGenerator...
Available Memory: 19492.59 MB
Train Data (Original) Memory Usage: 0.28 MB (0.0% 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.
Stage 2 Generators:
Fitting FillNaFeatureGenerator...
Stage 3 Generators:
Fitting IdentityFeatureGenerator...
Fitting CategoryFeatureGenerator...
Fitting CategoryMemoryMinimizeFeatureGenerator...
Stage 4 Generators:
Fitting DropUniqueFeatureGenerator...
Types of features in original data (raw dtype, special dtypes):
('int', []) : 6 | ['age', 'fnlwgt', 'education-num', 'capital-gain', 'capital-loss', ...]
('object', []) : 8 | ['workclass', 'education', 'marital-status', 'relationship', 'race', ...]
Types of features in processed data (raw dtype, special dtypes):
('category', []) : 6 | ['workclass', 'education', 'marital-status', 'relationship', 'race', ...]
('int', []) : 6 | ['age', 'fnlwgt', 'education-num', 'capital-gain', 'capital-loss', ...]
('int', ['bool']) : 2 | ['sex', 'class']
0.1s = Fit runtime
14 features in original data used to generate 14 features in processed data.
Train Data (Processed) Memory Usage: 0.03 MB (0.0% of available memory)
Data preprocessing and feature engineering runtime = 0.08s ...
AutoGluon will gauge predictive performance using evaluation metric: 'accuracy'
To change this, specify the eval_metric argument of fit()
Automatically generating train/validation split with holdout_frac=0.2, Train Rows: 391, Val Rows: 98
Excluded Model Types: ['KNN', 'NN', 'custom']
Found 'NN' model in hyperparameters, but 'NN' is present in excluded_model_types and will be removed.
Found 'KNN' model in hyperparameters, but 'KNN' is present in excluded_model_types and will be removed.
Found 'KNN' model in hyperparameters, but 'KNN' is present in excluded_model_types and will be removed.
Fitting 10 L1 models ...
Fitting model: NeuralNetFastAI ... Training model for up to 29.92s of the 29.92s of remaining time.
No improvement since epoch 5: early stopping
0.3163 = Validation score (accuracy)
0.48s = Training runtime
0.02s = Validation runtime
Fitting model: LightGBMXT ... Training model for up to 29.42s of the 29.41s of remaining time.
0.3571 = Validation score (accuracy)
0.64s = Training runtime
0.01s = Validation runtime
Fitting model: LightGBM ... Training model for up to 28.74s of the 28.74s of remaining time.
0.3673 = Validation score (accuracy)
0.7s = Training runtime
0.01s = Validation runtime
Fitting model: RandomForestGini ... Training model for up to 28.02s of the 28.02s of remaining time.
0.3061 = Validation score (accuracy)
0.71s = Training runtime
0.11s = Validation runtime
Fitting model: RandomForestEntr ... Training model for up to 27.18s of the 27.18s of remaining time.
0.2857 = Validation score (accuracy)
0.61s = Training runtime
0.11s = Validation runtime
Fitting model: CatBoost ... Training model for up to 26.44s of the 26.44s of remaining time.
0.3571 = Validation score (accuracy)
5.3s = Training runtime
0.01s = Validation runtime
Fitting model: ExtraTreesGini ... Training model for up to 21.13s of the 21.13s of remaining time.
0.2959 = Validation score (accuracy)
0.71s = Training runtime
0.11s = Validation runtime
Fitting model: ExtraTreesEntr ... Training model for up to 20.29s of the 20.29s of remaining time.
0.2959 = Validation score (accuracy)
0.71s = Training runtime
0.11s = Validation runtime
Fitting model: XGBoost ... Training model for up to 19.45s of the 19.45s of remaining time.
0.398 = Validation score (accuracy)
0.74s = Training runtime
0.01s = Validation runtime
Fitting model: LightGBMLarge ... Training model for up to 18.6s of the 18.6s of remaining time.
0.2959 = Validation score (accuracy)
2.75s = Training runtime
0.01s = Validation runtime
Fitting model: WeightedEnsemble_L2 ... Training model for up to 29.92s of the 14.99s of remaining time.
0.4082 = Validation score (accuracy)
0.21s = Training runtime
0.0s = Validation runtime
AutoGluon training complete, total runtime = 15.24s ...
TabularPredictor saved. To load, use: predictor = TabularPredictor.load("AutogluonModels/ag-20210831_221520/")
If you encounter memory issues¶
To reduce memory usage during training, you may try each of the following strategies individually or combinations of them (these may harm accuracy):
In
fit(), setnum_bag_sets = 1(can also try values greater than 1 to harm accuracy less).In
fit(), setexcluded_model_types = ['KNN', 'XT' ,'RF'](or some subset of these models).Try different
presetsinfit().In
fit(), sethyperparameters = 'light'orhyperparameters = 'very_light'.Text fields in your table require substantial memory for N-gram featurization. To mitigate this in
fit(), you can either: (1) add'ignore_text'to yourpresetslist (to ignore text features), or (2) specify the argument:
from sklearn.feature_extraction.text import CountVectorizer
from autogluon.features.generators import AutoMLPipelineFeatureGenerator
feature_generator = AutoMLPipelineFeatureGenerator(vectorizer=CountVectorizer(min_df=30, ngram_range=(1, 3), max_features=MAX_NGRAM, dtype=np.uint8))
where MAX_NGRAM = 1000 say (try various values under 10000 to reduce
the number of N-gram features used to represent each text field)
In addition to reducing memory usage, many of the above strategies can also be used to reduce training times.
To reduce memory usage during inference:
If trying to produce predictions for a large test dataset, break the test data into smaller chunks as demonstrated in FAQ.
If models have been previously persisted in memory but inference-speed is not a major concern, call
predictor.unpersist_models().If models have been previously persisted in memory, bagging was used in
fit(), and inference-speed is a concern: callpredictor.refit_full()and use one of the refit-full models for prediction (ensure this is the only model persisted in memory).
If you encounter disk space issues¶
To reduce disk usage, you may try each of the following strategies individually or combinations of them:
Make sure to delete all
predictor.pathfolders from previousfit()runs! These can eat up your free space if you callfit()many times. If you didn’t specifypath, AutoGluon still automatically saved its models to a folder called: “AutogluonModels/ag-[TIMESTAMP]”, where TIMESTAMP records whenfit()was called, so make sure to also delete these folders if you run low on free space.Call
predictor.save_space()to delete auxiliary files produced duringfit().Call
predictor.delete_models(models_to_keep='best', dry_run=False)if you only intend to use this predictor for inference going forward (will delete files required for non-prediction-related functionality likefit_summary).In
fit(), you can add'optimize_for_deployment'to thepresetslist, which will automatically invoke the previous two strategies after training.Most of the above strategies to reduce memory usage will also reduce disk usage (but may harm accuracy).
References¶
The following paper describes how AutoGluon internally operates on tabular data:
Erickson et al. AutoGluon-Tabular: Robust and Accurate AutoML for Structured Data. Arxiv, 2020.