Matthias Feurer: Dummy implementation of a multi-objective ensemble. (#1523)

Author: Github Actions

development/_downloads/baf53fc945368668a0cd202acebc6220/example_parallel_manual_spawning_cli.py

@@ -193,7 +193,7 @@ def cli_start_worker(scheduler_file_name):
         delete_tmp_folder_after_terminate=False,
         time_left_for_this_task=30,
         per_run_time_limit=10,
-        memory_limit=1024,
+        memory_limit=2048,
         tmp_folder=tmp_folder,
         seed=777,
         # n_jobs is ignored internally as we pass a dask client.

development/_downloads/bc82bea3a5dd7bdba60b65220891d9e5/examples_python.zip

@@ -123,7 +123,7 @@
       },
       "outputs": [],
       "source": [
-        "if __name__ == \"__main__\":\n    X, y = sklearn.datasets.load_breast_cancer(return_X_y=True)\n    X_train, X_test, y_train, y_test = sklearn.model_selection.train_test_split(\n        X, y, random_state=1\n    )\n\n    automl = AutoSklearnClassifier(\n        delete_tmp_folder_after_terminate=False,\n        time_left_for_this_task=30,\n        per_run_time_limit=10,\n        memory_limit=1024,\n        tmp_folder=tmp_folder,\n        seed=777,\n        # n_jobs is ignored internally as we pass a dask client.\n        n_jobs=1,\n        # Pass a dask client which connects to the previously constructed cluster.\n        dask_client=client,\n    )\n    automl.fit(X_train, y_train)\n\n    automl.fit_ensemble(\n        y_train,\n        task=MULTICLASS_CLASSIFICATION,\n        dataset_name=\"digits\",\n        ensemble_kwargs={\"ensemble_size\": 20},\n        ensemble_nbest=50,\n    )\n\n    predictions = automl.predict(X_test)\n    print(automl.sprint_statistics())\n    print(\"Accuracy score\", sklearn.metrics.accuracy_score(y_test, predictions))"
+        "if __name__ == \"__main__\":\n    X, y = sklearn.datasets.load_breast_cancer(return_X_y=True)\n    X_train, X_test, y_train, y_test = sklearn.model_selection.train_test_split(\n        X, y, random_state=1\n    )\n\n    automl = AutoSklearnClassifier(\n        delete_tmp_folder_after_terminate=False,\n        time_left_for_this_task=30,\n        per_run_time_limit=10,\n        memory_limit=2048,\n        tmp_folder=tmp_folder,\n        seed=777,\n        # n_jobs is ignored internally as we pass a dask client.\n        n_jobs=1,\n        # Pass a dask client which connects to the previously constructed cluster.\n        dask_client=client,\n    )\n    automl.fit(X_train, y_train)\n\n    automl.fit_ensemble(\n        y_train,\n        task=MULTICLASS_CLASSIFICATION,\n        dataset_name=\"digits\",\n        ensemble_kwargs={\"ensemble_size\": 20},\n        ensemble_nbest=50,\n    )\n\n    predictions = automl.predict(X_test)\n    print(automl.sprint_statistics())\n    print(\"Accuracy score\", sklearn.metrics.accuracy_score(y_test, predictions))"
       ]
     },
     {