Skip to content

Commit 9d1a418

Browse files
pravali96gwang111
authored andcommitted
add in-process mode definition to docs (#1622)
Co-authored-by: Gary Wang <[email protected]>
1 parent 784a18f commit 9d1a418

File tree

2 files changed

+14
-7
lines changed

2 files changed

+14
-7
lines changed

src/sagemaker/serve/builder/model_builder.py

Lines changed: 4 additions & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -150,6 +150,7 @@ class ModelBuilder(Triton, DJL, JumpStart, TGI, Transformers, TensorflowServing,
150150
151151
* ``Mode.SAGEMAKER_ENDPOINT``: Launch on a SageMaker endpoint
152152
* ``Mode.LOCAL_CONTAINER``: Launch locally with a container
153+
* ``Mode.IN_PROCESS``: Launch locally to a FastAPI server instead of using a container.
153154
shared_libs (List[str]): Any shared libraries you want to bring into
154155
the model packaging.
155156
dependencies (Optional[Dict[str, Any]): The dependencies of the model
@@ -613,7 +614,9 @@ def _model_builder_deploy_wrapper(
613614
)
614615

615616
if "endpoint_logging" not in kwargs:
616-
kwargs["endpoint_logging"] = False
617+
kwargs["endpoint_logging"] = True
618+
kwargs.pop("mode", None)
619+
self.pysdk_model.role = kwargs.pop("role", self.pysdk_model.role)
617620
predictor = self._original_deploy(
618621
*args,
619622
instance_type=instance_type,

src/sagemaker/serve/utils/predictors.py

Lines changed: 10 additions & 6 deletions
Original file line numberDiff line numberDiff line change
@@ -333,10 +333,14 @@ def __init__(
333333

334334
def predict(self, data):
335335
"""Placeholder docstring"""
336-
return self._mode_obj._invoke_serving(
337-
self.serializer.serialize(data),
338-
self.content_type,
339-
self.deserializer.ACCEPT[0],
336+
return self.deserializer.deserialize(
337+
io.BytesIO(
338+
self._mode_obj._invoke_serving(
339+
self.serializer.serialize(data),
340+
self.content_type,
341+
self.accept[0],
342+
)
343+
)
340344
)
341345

342346
@property
@@ -357,8 +361,8 @@ def delete_predictor(self):
357361
def _get_in_process_mode_predictor(
358362
# model_server: ModelServer,
359363
mode_obj: Type[InProcessMode],
360-
serializer=IdentitySerializer(),
361-
deserializer=BytesDeserializer(),
364+
serializer=JSONSerializer(),
365+
deserializer=JSONDeserializer(),
362366
) -> Type[PredictorBase]:
363367
"""Returns Predictor for IN_PROCESS mode"""
364368
return InProcessModePredictor(

0 commit comments

Comments
 (0)