autogluon.timeseries 1.0.1b20240325__tar.gz → 1.0.1b20240326__tar.gz

{autogluon.timeseries-1.0.1b20240325 → autogluon.timeseries-1.0.1b20240326}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: autogluon.timeseries
-Version: 1.0.1b20240325
+Version: 1.0.1b20240326
 Summary: AutoML for Image, Text, and Tabular Data
 Home-page: https://github.com/autogluon/autogluon
 Author: AutoGluon Community

{autogluon.timeseries-1.0.1b20240325 → autogluon.timeseries-1.0.1b20240326}/src/autogluon/timeseries/configs/presets_configs.py

@@ -1,4 +1,5 @@
 """Preset configurations for autogluon.timeseries Predictors"""
+
 from autogluon.timeseries.models.presets import get_default_hps
 
 # TODO: change default HPO settings when other HPO strategies (e.g., Ray tune) are available
@@ -11,18 +12,23 @@ TIMESERIES_PRESETS_CONFIGS = dict(
     fast_training={"hyperparameters": "very_light"},
     chronos_tiny={
         "hyperparameters": {"Chronos": {"model_path": "tiny"}},
+        "skip_model_selection": True,
     },
     chronos_mini={
         "hyperparameters": {"Chronos": {"model_path": "mini"}},
+        "skip_model_selection": True,
     },
     chronos_small={
         "hyperparameters": {"Chronos": {"model_path": "small"}},
+        "skip_model_selection": True,
     },
     chronos_base={
         "hyperparameters": {"Chronos": {"model_path": "base"}},
+        "skip_model_selection": True,
     },
     chronos_large={
         "hyperparameters": {"Chronos": {"model_path": "large", "batch_size": 8}},
+        "skip_model_selection": True,
     },
     chronos_ensemble={
         "hyperparameters": {

{autogluon.timeseries-1.0.1b20240325 → autogluon.timeseries-1.0.1b20240326}/src/autogluon/timeseries/learner.py

@@ -97,6 +97,7 @@ class TimeSeriesLearner(AbstractLearner):
                 target=self.target,
                 quantile_levels=self.quantile_levels,
                 verbosity=kwargs.get("verbosity", 2),
+                skip_model_selection=kwargs.get("skip_model_selection", False),
                 enable_ensemble=kwargs.get("enable_ensemble", True),
                 metadata=self.feature_generator.covariate_metadata,
                 val_splitter=val_splitter,

{autogluon.timeseries-1.0.1b20240325 → autogluon.timeseries-1.0.1b20240326}/src/autogluon/timeseries/predictor.py

@@ -425,6 +425,7 @@ class TimeSeriesPredictor(TimeSeriesPredictorDeprecatedMixin):
         refit_every_n_windows: int = 1,
         refit_full: bool = False,
         enable_ensemble: bool = True,
+        skip_model_selection: bool = False,
         random_seed: Optional[int] = 123,
         verbosity: Optional[int] = None,
     ) -> "TimeSeriesPredictor":
@@ -642,6 +643,10 @@ class TimeSeriesPredictor(TimeSeriesPredictorDeprecatedMixin):
         enable_ensemble : bool, default = True
             If True, the ``TimeSeriesPredictor`` will fit a simple weighted ensemble on top of the models specified via
             ``hyperparameters``.
+        skip_model_selection : bool, default = False
+            If True, predictor will not compute the validation score. For example, this argument is useful if we want
+            to use the predictor as a wrapper for a single pre-trained model. If set to True, then the ``hyperparameters``
+            dict must contain exactly one model without hyperparameter search spaces or an exception will be raised.
         random_seed : int or None, default = 123
             If provided, fixes the seed of the random number generator for all models. This guarantees reproducible
             results for most models (except those trained on GPU because of the non-determinism of GPU operations).
@@ -681,6 +686,7 @@ class TimeSeriesPredictor(TimeSeriesPredictorDeprecatedMixin):
             val_step_size=val_step_size,
             refit_every_n_windows=refit_every_n_windows,
             refit_full=refit_full,
+            skip_model_selection=skip_model_selection,
             enable_ensemble=enable_ensemble,
             random_seed=random_seed,
             verbosity=verbosity,
@@ -715,9 +721,10 @@ class TimeSeriesPredictor(TimeSeriesPredictorDeprecatedMixin):
         if num_val_windows == 0 and tuning_data is None:
             raise ValueError("Please set num_val_windows >= 1 or provide custom tuning_data")
 
-        train_data = self._filter_short_series(
-            train_data, num_val_windows=num_val_windows, val_step_size=val_step_size
-        )
+        if not skip_model_selection:
+            train_data = self._filter_short_series(
+                train_data, num_val_windows=num_val_windows, val_step_size=val_step_size
+            )
 
         val_splitter = ExpandingWindowSplitter(
             prediction_length=self.prediction_length, num_val_windows=num_val_windows, val_step_size=val_step_size
@@ -734,6 +741,7 @@ class TimeSeriesPredictor(TimeSeriesPredictorDeprecatedMixin):
             verbosity=verbosity,
             val_splitter=val_splitter,
             refit_every_n_windows=refit_every_n_windows,
+            skip_model_selection=skip_model_selection,
             enable_ensemble=enable_ensemble,
             random_seed=random_seed,
         )

{autogluon.timeseries-1.0.1b20240325 → autogluon.timeseries-1.0.1b20240326}/src/autogluon/timeseries/trainer/abstract_trainer.py

@@ -76,6 +76,8 @@ class SimpleAbstractTrainer:
         models = self.get_model_names()
         if not models:
             raise ValueError("Trainer has no fit models that can predict.")
+        if len(models) == 1:
+            return models[0]
         model_performances = self.get_models_attribute_dict(attribute="val_score")
         performances_list = [(m, model_performances[m]) for m in models if model_performances[m] is not None]
 
@@ -256,6 +258,7 @@ class AbstractTimeSeriesTrainer(SimpleAbstractTrainer):
         eval_metric: Union[str, TimeSeriesScorer, None] = None,
         eval_metric_seasonal_period: Optional[int] = None,
         save_data: bool = True,
+        skip_model_selection: bool = False,
         enable_ensemble: bool = True,
         verbosity: int = 2,
         val_splitter: Optional[AbstractWindowSplitter] = None,
@@ -270,7 +273,9 @@ class AbstractTimeSeriesTrainer(SimpleAbstractTrainer):
         self.target = kwargs.get("target", "target")
         self.metadata = kwargs.get("metadata", CovariateMetadata())
         self.is_data_saved = False
-        self.enable_ensemble = enable_ensemble
+        self.skip_model_selection = skip_model_selection
+        # Ensemble cannot be fit if val_scores are not computed
+        self.enable_ensemble = enable_ensemble and not skip_model_selection
         self.ensemble_model_type = TimeSeriesGreedyEnsemble
 
         self.verbosity = verbosity
@@ -487,7 +492,7 @@ class AbstractTimeSeriesTrainer(SimpleAbstractTrainer):
             fit_end_time = time.time()
             model.fit_time = model.fit_time or (fit_end_time - fit_start_time)
 
-            if val_data is not None:
+            if val_data is not None and not self.skip_model_selection:
                 model.score_and_cache_oof(val_data, store_val_score=True, store_predict_time=True)
 
             self._log_scores_and_times(model.val_score, model.fit_time, model.predict_time)
@@ -557,6 +562,17 @@ class AbstractTimeSeriesTrainer(SimpleAbstractTrainer):
 
         logger.info(f"Models that will be trained: {list(m.name for m in models)}")
 
+        if self.skip_model_selection:
+            if len(models) > 1:
+                raise ValueError(
+                    "When `skip_model_selection=True`, only a single model must be provided via `hyperparameters` "
+                    f"but {len(models)} models were given"
+                )
+            if contains_searchspace(models[0].get_user_params()):
+                raise ValueError(
+                    "When `skip_model_selection=True`, model configuration should contain no search spaces."
+                )
+
         num_base_models = len(models)
         model_names_trained = []
         for i, model in enumerate(models):
@@ -646,7 +662,8 @@ class AbstractTimeSeriesTrainer(SimpleAbstractTrainer):
         try:
             best_model = self.get_model_best()
             logger.info(f"Best model: {best_model}")
-            logger.info(f"Best model score: {self.get_model_attribute(best_model, 'val_score'):.4f}")
+            if not self.skip_model_selection:
+                logger.info(f"Best model score: {self.get_model_attribute(best_model, 'val_score'):.4f}")
         except ValueError as e:
             logger.error(str(e))
 

{autogluon.timeseries-1.0.1b20240325 → autogluon.timeseries-1.0.1b20240326}/src/autogluon/timeseries/trainer/auto_trainer.py

@@ -27,7 +27,8 @@ class AutoTimeSeriesTrainer(AbstractTimeSeriesTrainer):
             target=self.target,
             metadata=self.metadata,
             excluded_model_types=kwargs.get("excluded_model_types"),
-            multi_window=multi_window,
+            # if skip_model_selection = True, we skip backtesting
+            multi_window=multi_window and not self.skip_model_selection,
         )
 
     def fit(

{autogluon.timeseries-1.0.1b20240325 → autogluon.timeseries-1.0.1b20240326}/src/autogluon/timeseries/version.py

@@ -1,3 +1,3 @@
 """This is the autogluon version file."""
-__version__ = '1.0.1b20240325'
+__version__ = '1.0.1b20240326'
 __lite__ = False

{autogluon.timeseries-1.0.1b20240325 → autogluon.timeseries-1.0.1b20240326}/src/autogluon.timeseries.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: autogluon.timeseries
-Version: 1.0.1b20240325
+Version: 1.0.1b20240326
 Summary: AutoML for Image, Text, and Tabular Data
 Home-page: https://github.com/autogluon/autogluon
 Author: AutoGluon Community

{autogluon.timeseries-1.0.1b20240325 → autogluon.timeseries-1.0.1b20240326}/src/autogluon.timeseries.egg-info/requires.txt

@@ -16,13 +16,13 @@ utilsforecast<0.0.11,>=0.0.10
 tqdm<5,>=4.38
 orjson~=3.9
 tensorboard<3,>=2.9
-autogluon.core[raytune]==1.0.1b20240325
-autogluon.common==1.0.1b20240325
-autogluon.tabular[catboost,lightgbm,xgboost]==1.0.1b20240325
+autogluon.core[raytune]==1.0.1b20240326
+autogluon.common==1.0.1b20240326
+autogluon.tabular[catboost,lightgbm,xgboost]==1.0.1b20240326
 
 [all]
-optimum[nncf,openvino]<1.18,>=1.17
 optimum[onnxruntime]<1.18,>=1.17
+optimum[nncf,openvino]<1.18,>=1.17
 
 [chronos-onnx]
 optimum[onnxruntime]<1.18,>=1.17