templates updated

model_catalog_examples/artifact_boilerplate/runtime.yaml

@@ -1,7 +1,20 @@
+# This is template runtime.yaml. It has default conda environment value. You can replace these values.
 MODEL_ARTIFACT_VERSION: '3.0'
 MODEL_DEPLOYMENT:
   INFERENCE_CONDA_ENV:
-    INFERENCE_ENV_PATH: <INSERT_INFERENCE_ENV_PATH>
-    INFERENCE_ENV_SLUG: <INSERT_INFERENCE_ENV_SLUG>
-    INFERENCE_ENV_TYPE: <INSERT_INFERENCE_ENV_TYPE>
-    INFERENCE_PYTHON_VERSION: <INSERT_INFERENCE_PYTHON_VERSION>
+    # Replace with the object storage path of the conda environment you want to use.
+    # Go to https://docs.oracle.com/en-us/iaas/data-science/using/conda_environ_list.htm for a list of data_science conda environments you can use out-of-the-box
+    INFERENCE_ENV_PATH: oci://service-conda-packs@id19sfcrra6z/service_pack/cpu/Data Exploration and Manipulation for CPU Python 3.7/2.0/dataexpl_p37_cpu_v2
+
+    # Replace with the slug of the environment you want to use.
+    # Slugs for data_science environment can be found for each environment either in the notebook session Environment Explorer or here:
+    # https://docs.oracle.com/en-us/iaas/data-science/using/conda_environ_list.htm
+    INFERENCE_ENV_SLUG: dataexpl_p37_cpu_v2
+
+    # Replace with the type of environment. Either published or data_science.
+    # Published environments are environments that you create and store in your own object storage bucket.
+    # For more information: https://docs.oracle.com/en-us/iaas/data-science/using/conda_publishs_object.htm
+    INFERENCE_ENV_TYPE: data_science
+
+    # Provide the Python version of the environment.
+    INFERENCE_PYTHON_VERSION: 3.7

model_catalog_examples/artifact_boilerplate/score.py

@@ -6,17 +6,20 @@
 from cloudpickle import cloudpickle
 
 """
-    Replace with your own model object and your own serialization library (e.g. pickle, onnx, etc).
+    You can provide your own model object and your own serialization library (e.g. pickle, onnx, etc).
+    If no model is specified then predict() by default will return 'Hello World!'
 """
-model_name = 'model.pkl'
 
+"""
+    model_name = 'model.pkl'
+"""
 
 """
    Inference script. This script is used for prediction by scoring server when schema is known.
 """
 
 
-def load_model(model_file_name=model_name):
+def load_model(model_file_name=None):
     """
     Loads model from the serialized format
     WARNING: Please use the same library to load the model which was used to serialise it.
@@ -25,6 +28,8 @@ def load_model(model_file_name=model_name):
     -------
     model:  a model instance on which predict API can be invoked
     """
+    if not model_file_name:
+        return None
     model_dir = os.path.dirname(os.path.realpath(__file__))
     contents = os.listdir(model_dir)
     # --------------------------WARNING-------------------------
@@ -50,6 +55,8 @@ def predict(data, model=load_model()):
     predictions: Output from scoring server
         Format: {'prediction':output from model.predict method}
     """
+    if model is None or len(data) == 0:
+        return {'prediction':'Hello world!'}
     from pandas import read_json, DataFrame
     from io import StringIO
     data = read_json(StringIO(data)) if isinstance(data, str) else DataFrame.from_dict(data)