def testBuildRequests_KerasModel(self):
builder = request_builder._TFServingRpcRequestBuilder(
model_name='foo',
signatures=self._GetKerasModelSignature())
builder.ReadExamplesArtifact(self._examples, num_examples=1)
result = builder.BuildRequests()
self.assertEqual(len(result), 1)
self.assertIsInstance(result[0], predict_pb2.PredictRequest)
self.assertEqual(result[0].model_spec.name, 'foo')
self.assertEqual(result[0].model_spec.signature_name, 'serving_default')
def testBuildRequests_EstimatorModel_ServingDefault(self):
builder = request_builder._TFServingRpcRequestBuilder(
model_name='foo',
signatures=self._GetEstimatorModelSignature())
builder.ReadExamplesArtifact(self._examples, num_examples=1)
result = builder.BuildRequests()
self.assertEqual(len(result), 1)
self.assertIsInstance(result[0], classification_pb2.ClassificationRequest)
self.assertEqual(result[0].model_spec.name, 'foo')
self.assertEqual(result[0].model_spec.signature_name, 'serving_default')
def testBuildRequests_EstimatorModel_Regression(self):
builder = request_builder._TFServingRpcRequestBuilder(
model_name='foo',
signatures=self._GetEstimatorModelSignature(
signature_names=['regression']))
builder.ReadExamplesArtifact(self._examples, num_examples=1)
result = builder.BuildRequests()
self.assertEqual(len(result), 1)
self.assertIsInstance(result[0], regression_pb2.RegressionRequest)
self.assertEqual(result[0].model_spec.name, 'foo')
self.assertEqual(result[0].model_spec.signature_name, 'regression')
def testBuildRequests_PredictMethod_FailOnInvalidSignature(self):
builder = request_builder._TFServingRpcRequestBuilder(
model_name='foo',
signatures={
# Signature argument is not for serialized tf.Example (i.e. dtype !=
# DT_STRING or shape != (None,)).
'serving_default':
_make_signature_def({
'method_name': 'tensorflow/serving/predict',
'inputs': {
'x': {
'name': 'serving_default_input:0',
'dtype': 'DT_FLOAT',
'tensor_shape': {
'dim': [
{
'size': -1
},
{
'size': 784
},
]
}
}
},
'outputs': {
'y': {
'name': 'StatefulPartitionedCall:0',
'dtype': 'DT_FLOAT',
'tensor_shape': {
'dim': [
{
'size': -1
},
{
'size': 10
},
]
}
}
},
})
})
builder.ReadExamplesArtifact(self._examples, num_examples=1)
with self.assertRaisesRegex(
ValueError,
'Unable to find valid input key from SignatureDef'):
builder.BuildRequests()
def testBuildRequests_PredictMethod(self):
builder = request_builder._TFServingRpcRequestBuilder(
model_name='foo',
signatures={
# Has only one argument with dtype=DT_STRING and shape=(None,).
# This is the only valid form that InfraValidator accepts today.
'serving_default':
_make_signature_def({
'method_name': 'tensorflow/serving/predict',
'inputs': {
'x': {
'name': 'serving_default_examples:0',
'dtype': 'DT_STRING',
'tensor_shape': {
'dim': [
{
'size': -1
},
]
}
}
},
'outputs': {
'y': {
'name': 'StatefulPartitionedCall:0',
'dtype': 'DT_FLOAT',
'tensor_shape': {
'dim': [
{
'size': -1
},
{
'size': 10
},
]
}
}
},
})
})
builder.ReadExamplesArtifact(self._examples, num_examples=1)
result = builder.BuildRequests()
self.assertEqual(len(result), 1)
self.assertIsInstance(result[0], predict_pb2.PredictRequest)
self.assertEqual(result[0].inputs['x'].dtype,
tf.dtypes.string.as_datatype_enum)
def testBuildRequests_EstimatorModel_Predict(self):
builder = request_builder._TFServingRpcRequestBuilder(
model_name='foo',
signatures=self._GetEstimatorModelSignature(
signature_names=['predict']))
builder.ReadExamplesArtifact(self._examples, num_examples=1)
result = builder.BuildRequests()
self.assertEqual(len(result), 1)
self.assertIsInstance(result[0], predict_pb2.PredictRequest)
self.assertEqual(result[0].model_spec.name, 'foo')
self.assertEqual(result[0].model_spec.signature_name, 'predict')
self.assertEqual(len(result[0].inputs), 1)
input_key = list(result[0].inputs.keys())[0]
self.assertEqual(result[0].inputs[input_key].dtype,
tf.dtypes.string.as_datatype_enum)
