inference.py file not called upon when deploying Tensorflow Model

[ana-pcosta]

Describe the bug
I have trained a custom tensorflow model to perform semantic segmentation on images. I would like to perform inference by passing a jpg file to the endpoint using 64base encoding. The model signature def is as described in https://github.com/aws/sagemaker-tensorflow-serving-container:

signature_def['serving_default']:
The given SavedModel SignatureDef contains the following input(s):
inputs['image_bytes'] tensor_info:
dtype: DT_STRING
shape: (-1)
name: input_tensor:0

And the inference.py is saved in the model.tar.gz file in this structure:
model
|--1
|--variables
|--saved_model.pb
code
|--inference.py
|--requirements.txt

But whenever I try to do predictor.content_type = 'application/x-image' and predictor.predict(img), where img is a bytearray of a jpg file, this error pops up:

TypeError Traceback (most recent call last)
in

~/anaconda3/envs/tensorflow2_p36/lib/python3.6/site-packages/sagemaker/tensorflow/serving.py in predict(self, data, initial_args)
118 args["CustomAttributes"] = self._model_attributes
119
--> 120 return super(Predictor, self).predict(data, args)
121
122

~/anaconda3/envs/tensorflow2_p36/lib/python3.6/site-packages/sagemaker/predictor.py in predict(self, data, initial_args, target_model, target_variant)
110 """
111
--> 112 request_args = self._create_request_args(data, initial_args, target_model, target_variant)
113 response = self.sagemaker_session.sagemaker_runtime_client.invoke_endpoint(**request_args)
114 return self._handle_response(response)

~/anaconda3/envs/tensorflow2_p36/lib/python3.6/site-packages/sagemaker/predictor.py in _create_request_args(self, data, initial_args, target_model, target_variant)
153
154 if self.serializer is not None:
--> 155 data = self.serializer(data)
156
157 args["Body"] = data

~/anaconda3/envs/tensorflow2_p36/lib/python3.6/site-packages/sagemaker/predictor.py in call(self, data)
545 return _json_serialize_from_buffer(data)
546
--> 547 return json.dumps(_ndarray_to_list(data))
548
549

~/anaconda3/envs/tensorflow2_p36/lib/python3.6/json/init.py in dumps(obj, skipkeys, ensure_ascii, check_circular, allow_nan, cls, indent, separators, default, sort_keys, **kw)
229 cls is None and indent is None and separators is None and
230 default is None and not sort_keys and not kw):
--> 231 return _default_encoder.encode(obj)
232 if cls is None:
233 cls = JSONEncoder

~/anaconda3/envs/tensorflow2_p36/lib/python3.6/json/encoder.py in encode(self, o)
197 # exceptions aren't as detailed. The list call should be roughly
198 # equivalent to the PySequence_Fast that ''.join() would do.
--> 199 chunks = self.iterencode(o, _one_shot=True)
200 if not isinstance(chunks, (list, tuple)):
201 chunks = list(chunks)

~/anaconda3/envs/tensorflow2_p36/lib/python3.6/json/encoder.py in iterencode(self, o, _one_shot)
255 self.key_separator, self.item_separator, self.sort_keys,
256 self.skipkeys, _one_shot)
--> 257 return _iterencode(o, 0)
258
259 def _make_iterencode(markers, _default, _encoder, _indent, _floatstr,

~/anaconda3/envs/tensorflow2_p36/lib/python3.6/json/encoder.py in default(self, o)
178 """
179 raise TypeError("Object of type '%s' is not JSON serializable" %
--> 180 o.class.name)
181
182 def encode(self, o):

TypeError: Object of type 'bytearray' is not JSON serializable

System information
A description of your system. Please provide:

• Tensorflow 2.0.0:
• Python 3.6:
• GPU:
• Tensorflow2 docker image:

[ajaykarpur]

Hi @ana-pcosta, please refer to the documentation on serializers: https://sagemaker.readthedocs.io/en/stable/api/inference/serializers.html

Please feel free to re-open this issue if you have further questions.