我有一个已注册的数据集(基于泰坦尼克号),并且用azure automl训练了一个模型:
from azureml.train.automl import AutoMLConfig
automl_config = AutoMLConfig(name='Automated ML Experiment on titanic dataset',
task='classification',
compute_target=compute_target,
enable_local_managed=True,
training_data = train_ds,
validation_data = test_ds,
label_column_name='Survived',
iterations=6,
primary_metric = 'AUC_weighted',
max_concurrent_iterations=4,
featurization='auto'
)
print("Ready for Auto ML run.")
from azureml.core.experiment import Experiment
from azureml.widgets import RunDetails
print('Submitting Auto ML experiment...')
automl_experiment = Experiment(ws, 'titanic_automl')
automl_run = automl_experiment.submit(automl_config)
RunDetails(automl_run).show()
automl_run.wait_for_completion(show_output=True)
best_run, fitted_model = automl_run.get_output()
from azureml.core import Model
# Register model
best_run.register_model(model_path='outputs/model.pkl', model_name='titanic_model_automl',
tags={'Training context':'Auto ML'},
properties={'AUC': best_run_metrics['AUC_weighted'], 'Accuracy': best_run_metrics['accuracy']})
然后我想通过使用以下命令在另一个笔记本中加载该模型:
import pickle
filename = 'model/azure-automl.pkl'
model = pickle.load(open(filename, 'rb'))
但是我无法确定它可以在本地还是在Azure计算实例上工作。