Make inputs optional for hyperparameter tuning job.

CHANGELOG.rst

@@ -11,6 +11,7 @@ CHANGELOG
 * doc-fix: Fix typos in tensorflow serving documentation
 * doc-fix: Add estimator base classes to API docs
 * feature: HyperparameterTuner: add support for Automatic Model Tuning's Warm Start Jobs
+* feature: HyperparameterTuner: Make input channels optional
 
 1.14.2
 ======

setup.py

@@ -53,7 +53,7 @@ def read(fname):
       ],
 
       # Declare minimal set for installation
-      install_requires=['boto3>=1.9.38', 'numpy>=1.9.0', 'protobuf>=3.1', 'scipy>=0.19.0',
+      install_requires=['boto3>=1.9.45', 'numpy>=1.9.0', 'protobuf>=3.1', 'scipy>=0.19.0',
                         'urllib3 >=1.21', 'PyYAML>=3.2', 'protobuf3-to-dict>=0.1.5',
                         'docker-compose>=1.23.0'],
 

src/sagemaker/session.py

@@ -348,13 +348,15 @@ def tune(self, job_name, strategy, objective_type, objective_metric_name,
                     'TrainingInputMode': input_mode,
                 },
                 'RoleArn': role,
-                'InputDataConfig': input_config,
                 'OutputDataConfig': output_config,
                 'ResourceConfig': resource_config,
                 'StoppingCondition': stop_condition,
             }
         }
 
+        if input_config is not None:
+            tune_request['TrainingJobDefinition']['InputDataConfig'] = input_config
+
         if warm_start_config:
             tune_request['WarmStartConfig'] = warm_start_config
 

src/sagemaker/tuner.py

@@ -327,7 +327,7 @@ def _prepare_for_training(self, job_name=None, include_cls_metadata=True):
                 self.estimator.__class__.__name__)
             self.static_hyperparameters[self.SAGEMAKER_ESTIMATOR_MODULE] = json.dumps(self.estimator.__module__)
 
-    def fit(self, inputs, job_name=None, include_cls_metadata=True, **kwargs):
+    def fit(self, inputs=None, job_name=None, include_cls_metadata=True, **kwargs):
         """Start a hyperparameter tuning job.
 
         Args:

tests/unit/test_tuner.py

@@ -15,6 +15,7 @@
 import copy
 import json
 
+import os
 import pytest
 from mock import Mock
 
@@ -26,6 +27,8 @@
     HyperparameterTuner, _TuningJob, WarmStartConfig, create_identical_dataset_and_algorithm_tuner, \
     create_transfer_learning_tuner, WarmStartTypes
 from sagemaker.mxnet import MXNet
+
+DATA_DIR = os.path.join(os.path.dirname(__file__), '..', 'data')
 MODEL_DATA = "s3://bucket/model.tar.gz"
 
 JOB_NAME = 'tuning_job'
@@ -488,6 +491,22 @@ def test_delete_endpoint(tuner):
     tuner.sagemaker_session.delete_endpoint.assert_called_with(JOB_NAME)
 
 
+def test_fit_no_inputs(tuner, sagemaker_session):
+    script_path = os.path.join(DATA_DIR, 'mxnet_mnist', 'failure_script.py')
+    tuner.estimator = MXNet(entry_point=script_path,
+                            role=ROLE,
+                            framework_version=FRAMEWORK_VERSION,
+                            train_instance_count=TRAIN_INSTANCE_COUNT,
+                            train_instance_type=TRAIN_INSTANCE_TYPE,
+                            sagemaker_session=sagemaker_session)
+
+    tuner.fit()
+
+    _, _, tune_kwargs = sagemaker_session.tune.mock_calls[0]
+
+    assert tune_kwargs['input_config'] is None
+
+
 def test_identical_dataset_and_algorithm_tuner(sagemaker_session):
     job_details = copy.deepcopy(TUNING_JOB_DETAILS)
     sagemaker_session.sagemaker_client.describe_hyper_parameter_tuning_job = Mock(name='describe_tuning_job',
@@ -523,6 +542,8 @@ def test_transfer_learning_tuner(sagemaker_session):
     assert parent_tuner.warm_start_config.type == WarmStartTypes.TRANSFER_LEARNING
     assert parent_tuner.warm_start_config.parents == {tuner.latest_tuning_job.name, "p1", "p2"}
     assert parent_tuner.estimator == tuner.estimator
+
+
 #################################################################################
 # _ParameterRange Tests