AutoGluon automates machine learning tasks enabling you to easily achieve strong predictive performance in your applications.
npx @tessl/cli install tessl/pypi-autogluon@1.4.0AutoGluon is a comprehensive automated machine learning (AutoML) library that enables developers to build high-accuracy ML models with minimal code across multiple data modalities including tabular, text, image, and time series data. The library provides a unified interface through various predictor classes that automatically handle feature engineering, model selection, hyperparameter optimization, and ensemble creation.
pip install autogluonimport autogluonCommon imports for specific domains:
from autogluon.tabular import TabularPredictor, InterpretableTabularPredictor, TabularDataset
from autogluon.multimodal import MultiModalPredictor
from autogluon.timeseries import TimeSeriesPredictor, TimeSeriesDataFrameCore utilities and data structures:
from autogluon.core import constants, metrics
from autogluon.common import TabularDataset, FeatureMetadata
from autogluon.features import * # Feature generators
from autogluon.eda import AnalysisState# Tabular ML (most common use case)
predictor = TabularPredictor(label='target_column')
predictor.fit('train.csv', presets='best_quality')
# Make predictions
predictions = predictor.predict('test.csv')
probabilities = predictor.predict_proba('test.csv')
# Evaluate performance
performance = predictor.evaluate('test.csv')
leaderboard = predictor.leaderboard()
# Multimodal data (text + image + tabular)
mm_predictor = MultiModalPredictor(label='label', problem_type='multiclass')
mm_predictor.fit(train_data) # pandas DataFrame with text, image paths, and numerical columns
# Time series forecasting
ts_data = TimeSeriesDataFrame.from_data_frame(df, id_column='item_id', timestamp_column='timestamp')
ts_predictor = TimeSeriesPredictor(prediction_length=24, freq='H')
ts_predictor.fit(ts_data)
# Generate forecasts
forecasts = ts_predictor.predict(ts_data)AutoGluon uses a modular architecture centered around specialized predictor classes:
This design enables easy switching between domains while maintaining consistent APIs and leveraging state-of-the-art models for each data type.
Automated machine learning for structured/tabular data supporting classification and regression tasks. Handles feature engineering, model selection, hyperparameter tuning, and ensembling with minimal configuration.
class TabularPredictor:
def __init__(self, label: str, problem_type: str = None, path: str = None, **kwargs): ...
def fit(self, train_data, presets: str = None, time_limit: int = None, **kwargs): ...
def predict(self, data): ...
def predict_proba(self, data): ...
def evaluate(self, data, **kwargs): ...
def leaderboard(self, data=None, **kwargs): ...Automated machine learning for heterogeneous data combining text, images, and tabular features. Supports classification, regression, object detection, named entity recognition, and semantic matching tasks.
class MultiModalPredictor:
def __init__(self, label: str = None, problem_type: str = None, presets: str = None, **kwargs): ...
def fit(self, train_data, **kwargs): ...
def predict(self, data, **kwargs): ...
def predict_proba(self, data, **kwargs): ...
def evaluate(self, data, **kwargs): ...
def extract_embedding(self, data, **kwargs): ...Probabilistic forecasting for univariate and multivariate time series data. Supports both statistical and deep learning models with automatic model selection and quantile predictions.
class TimeSeriesPredictor:
def __init__(self, target: str = "target", prediction_length: int = 1, freq: str = None, **kwargs): ...
def fit(self, train_data, **kwargs): ...
def predict(self, data, **kwargs): ...
def evaluate(self, data, **kwargs): ...
def leaderboard(self, data=None, **kwargs): ...Comprehensive feature generation and transformation capabilities for automated feature engineering across different data types.
class AutoMLPipelineFeatureGenerator:
def __init__(self, **kwargs): ...
def fit_transform(self, X: DataFrame, y: Series = None, **kwargs): ...
def transform(self, X: DataFrame, **kwargs): ...Shared utilities for metrics, constants, and data structures used across all AutoGluon predictors.
class TabularDataset:
def __init__(self, df: DataFrame): ...
@classmethod
def load(cls, file_path: str): ...
class Scorer:
def __init__(self, name: str, score_func: callable, **kwargs): ...# Problem type constants
BINARY = "binary"
MULTICLASS = "multiclass"
REGRESSION = "regression"
QUANTILE = "quantile"
# Common data structures
TabularDataset = pandas.DataFrame # Enhanced DataFrame with AutoGluon utilities
TimeSeriesDataFrame = pandas.DataFrame # Time series specific DataFrame structure
# Predictor types
TabularPredictor = Type[TabularPredictor]
MultiModalPredictor = Type[MultiModalPredictor]
TimeSeriesPredictor = Type[TimeSeriesPredictor]