將超參數微調升級至 SDK v2
在 SDK v2 中，微調超參數會合併到作業中。
工作具有類型。 大部分作業都是執行 command的命令作業，例如 python main.py。 工作中執行的內容與任何程式設計語言無關，因此您可以執行 bash 指令碼、叫用 python 解譯器、執行一堆 curl 命令或任何其他項目。
掃掠作業是另一種作業類型，其定義掃掠設定，並可藉由呼叫命令的掃掠方法來起始。
若要升級，您必須變更程序代碼，以定義和提交超參數微調實驗至 SDK v2。 您在作業內執行的內容不需要升級至 SDK v2。 不過，建議您從模型定型腳本中移除 Azure 機器學習 特定的任何程序代碼。 此區隔可讓您更輕鬆地在本機和雲端之間進行轉換，並被視為成熟 MLOps 的最佳做法。 實際上，這表示移除 azureml.* 程式代碼行。 模型記錄和追蹤程式代碼應該取代為 MLflow。 如需詳細資訊，請參閱 如何在 v2 中使用 MLflow。
本文提供 SDK v1 和 SDK v2 中案例的比較。
在實驗中執行超參數微調
SDK v1
from azureml.core import ScriptRunConfig, Experiment, Workspace from azureml.train.hyperdrive import RandomParameterSampling, BanditPolicy, HyperDriveConfig, PrimaryMetricGoal from azureml.train.hyperdrive import choice, loguniform dataset = Dataset.get_by_name(ws, 'mnist-dataset') # list the files referenced by mnist dataset dataset.to_path() #define the search space for your hyperparameters param_sampling = RandomParameterSampling( { '--batch-size': choice(25, 50, 100), '--first-layer-neurons': choice(10, 50, 200, 300, 500), '--second-layer-neurons': choice(10, 50, 200, 500), '--learning-rate': loguniform(-6, -1) } ) args = ['--data-folder', dataset.as_named_input('mnist').as_mount()] #Set up your script run src = ScriptRunConfig(source_directory=script_folder, script='keras_mnist.py', arguments=args, compute_target=compute_target, environment=keras_env) # Set early stopping on this one early_termination_policy = BanditPolicy(evaluation_interval=2, slack_factor=0.1) # Define the configurations for your hyperparameter tuning experiment hyperdrive_config = HyperDriveConfig(run_config=src, hyperparameter_sampling=param_sampling, policy=early_termination_policy, primary_metric_name='Accuracy', primary_metric_goal=PrimaryMetricGoal.MAXIMIZE, max_total_runs=20, max_concurrent_runs=4) # Specify your experiment details experiment = Experiment(workspace, experiment_name) hyperdrive_run = experiment.submit(hyperdrive_config) #Find the best model best_run = hyperdrive_run.get_best_run_by_primary_metric()SDK v2
from azure.ai.ml import MLClient from azure.ai.ml import command, Input from azure.ai.ml.sweep import Choice, Uniform, MedianStoppingPolicy from azure.identity import DefaultAzureCredential # Create your command command_job_for_sweep = command( code="./src", command="python main.py --iris-csv ${{inputs.iris_csv}} --learning-rate ${{inputs.learning_rate}} --boosting ${{inputs.boosting}}", environment="AzureML-lightgbm-3.2-ubuntu18.04-py37-cpu@latest", inputs={ "iris_csv": Input( type="uri_file", path="https://azuremlexamples.blob.core.windows.net/datasets/iris.csv", ), #define the search space for your hyperparameters "learning_rate": Uniform(min_value=0.01, max_value=0.9), "boosting": Choice(values=["gbdt", "dart"]), }, compute="cpu-cluster", ) # Call sweep() on your command job to sweep over your parameter expressions sweep_job = command_job_for_sweep.sweep( compute="cpu-cluster", sampling_algorithm="random", primary_metric="test-multi_logloss", goal="Minimize", ) # Define the limits for this sweep sweep_job.set_limits(max_total_trials=20, max_concurrent_trials=10, timeout=7200) # Set early stopping on this one sweep_job.early_termination = MedianStoppingPolicy(delay_evaluation=5, evaluation_interval=2) # Specify your experiment details sweep_job.display_name = "lightgbm-iris-sweep-example" sweep_job.experiment_name = "lightgbm-iris-sweep-example" sweep_job.description = "Run a hyperparameter sweep job for LightGBM on Iris dataset." # submit the sweep returned_sweep_job = ml_client.create_or_update(sweep_job) # get a URL for the status of the job returned_sweep_job.services["Studio"].endpoint # Download best trial model output ml_client.jobs.download(returned_sweep_job.name, output_name="model")
在管線中執行超參數微調
SDK v1
tf_env = Environment.get(ws, name='AzureML-TensorFlow-2.0-GPU') data_folder = dataset.as_mount() src = ScriptRunConfig(source_directory=script_folder, script='tf_mnist.py', arguments=['--data-folder', data_folder], compute_target=compute_target, environment=tf_env) #Define HyperDrive configs ps = RandomParameterSampling( { '--batch-size': choice(25, 50, 100), '--first-layer-neurons': choice(10, 50, 200, 300, 500), '--second-layer-neurons': choice(10, 50, 200, 500), '--learning-rate': loguniform(-6, -1) } ) early_termination_policy = BanditPolicy(evaluation_interval=2, slack_factor=0.1) hd_config = HyperDriveConfig(run_config=src, hyperparameter_sampling=ps, policy=early_termination_policy, primary_metric_name='validation_acc', primary_metric_goal=PrimaryMetricGoal.MAXIMIZE, max_total_runs=4, max_concurrent_runs=4) metrics_output_name = 'metrics_output' metrics_data = PipelineData(name='metrics_data', datastore=datastore, pipeline_output_name=metrics_output_name, training_output=TrainingOutput("Metrics")) model_output_name = 'model_output' saved_model = PipelineData(name='saved_model', datastore=datastore, pipeline_output_name=model_output_name, training_output=TrainingOutput("Model", model_file="outputs/model/saved_model.pb")) #Create HyperDriveStep hd_step_name='hd_step01' hd_step = HyperDriveStep( name=hd_step_name, hyperdrive_config=hd_config, inputs=[data_folder], outputs=[metrics_data, saved_model]) #Find and register best model conda_dep = CondaDependencies() conda_dep.add_pip_package("azureml-sdk") rcfg = RunConfiguration(conda_dependencies=conda_dep) register_model_step = PythonScriptStep(script_name='register_model.py', name="register_model_step01", inputs=[saved_model], compute_target=cpu_cluster, arguments=["--saved-model", saved_model], allow_reuse=True, runconfig=rcfg) register_model_step.run_after(hd_step) #Run the pipeline pipeline = Pipeline(workspace=ws, steps=[hd_step, register_model_step]) pipeline_run = exp.submit(pipeline)SDK v2
train_component_func = load_component(path="./train.yml") score_component_func = load_component(path="./predict.yml") # define a pipeline @pipeline() def pipeline_with_hyperparameter_sweep(): """Tune hyperparameters using sample components.""" train_model = train_component_func( data=Input( type="uri_file", path="wasbs://datasets@azuremlexamples.blob.core.windows.net/iris.csv", ), c_value=Uniform(min_value=0.5, max_value=0.9), kernel=Choice(["rbf", "linear", "poly"]), coef0=Uniform(min_value=0.1, max_value=1), degree=3, gamma="scale", shrinking=False, probability=False, tol=0.001, cache_size=1024, verbose=False, max_iter=-1, decision_function_shape="ovr", break_ties=False, random_state=42, ) sweep_step = train_model.sweep( primary_metric="training_f1_score", goal="minimize", sampling_algorithm="random", compute="cpu-cluster", ) sweep_step.set_limits(max_total_trials=20, max_concurrent_trials=10, timeout=7200) score_data = score_component_func( model=sweep_step.outputs.model_output, test_data=sweep_step.outputs.test_data ) pipeline_job = pipeline_with_hyperparameter_sweep() # set pipeline level compute pipeline_job.settings.default_compute = "cpu-cluster" # submit job to workspace pipeline_job = ml_client.jobs.create_or_update( pipeline_job, experiment_name="pipeline_samples" ) pipeline_job
SDK v1 和 SDK v2 中的主要功能對應
| SDK v1 中的功能 | SDK v2 中的粗略對應 |
|---|---|
| HyperDriveRunConfig（） | SweepJob（） |
| hyperdrive 套件 | 掃掠套件 |
下一步
如需詳細資訊，請參閱