AWS Marketplace Product Usage Demonstration - Model Packages

---

This notebook’s CI test result for us-west-2 is as follows. CI test results in other regions can be found at the end of the notebook.

---

Using Model Package ARN with Amazon SageMaker APIs

This sample notebook demonstrates two new functionalities added to Amazon SageMaker: 1. Using a Model Package ARN for inference via Batch Transform jobs / Live Endpoints 2. Using a Marketplace Model Package ARN - we will use Scikit Decision Trees - Pretrained Model

Overall flow diagram

Compatibility

This notebook is compatible only with Scikit Decision Trees - Pretrained Model sample model that is published to AWS Marketplace

Set up the environment

import sagemaker as sage
from sagemaker import get_execution_role
from sagemaker.serializers import CSVSerializer

role = get_execution_role()

# S3 prefixes
common_prefix = "DEMO-scikit-byo-iris"
batch_inference_input_prefix = common_prefix + "/batch-inference-input-data"
from sagemaker.predictor import Predictor

Create the session

The session remembers our connection parameters to Amazon SageMaker. We’ll use it to perform all of our Amazon SageMaker operations.

sagemaker_session = sage.Session()

Create Model

Now we use the above Model Package to create a model

from src.scikit_product_arns import ScikitArnProvider

modelpackage_arn = ScikitArnProvider.get_model_package_arn(sagemaker_session.boto_region_name)
print("Using model package arn " + modelpackage_arn)

from sagemaker import ModelPackage

model = ModelPackage(
    role=role, model_package_arn=modelpackage_arn, sagemaker_session=sagemaker_session
)

Batch Transform Job

Now let’s use the model built to run a batch inference job and verify it works.

Batch Transform Input Preparation

The snippet below is removing the “label” column (column indexed at 0) and retaining the rest to be batch transform’s input.

NOTE: This is the same training data, which is a no-no from a statistical/ML science perspective. But the aim of this notebook is to demonstrate how things work end-to-end.

import pandas as pd

## Remove first column that contains the label
shape = pd.read_csv("data/training/iris.csv", header=None).drop([0], axis=1)

TRANSFORM_WORKDIR = "data/transform"
shape.to_csv(TRANSFORM_WORKDIR + "/batchtransform_test.csv", index=False, header=False)

transform_input = (
    sagemaker_session.upload_data(TRANSFORM_WORKDIR, key_prefix=batch_inference_input_prefix)
    + "/batchtransform_test.csv"
)
print("Transform input uploaded to " + transform_input)

import json
import uuid

transformer = model.transformer(1, "ml.m4.xlarge")
transformer.transform(transform_input, content_type="text/csv")
transformer.wait()

print("Batch Transform output saved to " + transformer.output_path)

Inspect the Batch Transform Output in S3

from urllib.parse import urlparse

parsed_url = urlparse(transformer.output_path)
bucket_name = parsed_url.netloc
file_key = "{}/{}.out".format(parsed_url.path[1:], "batchtransform_test.csv")

s3_client = sagemaker_session.boto_session.client("s3")

response = s3_client.get_object(Bucket=sagemaker_session.default_bucket(), Key=file_key)
response_bytes = response["Body"].read().decode("utf-8")
print(response_bytes)

Live Inference Endpoint

Now we demonstrate the creation of an endpoint for live inference

endpoint_name = "scikit-model"
predictor = model.deploy(1, "ml.m4.xlarge", endpoint_name=endpoint_name)

Choose some data and use it for a prediction

In order to do some predictions, we’ll extract some of the data we used for training and do predictions against it. This is, of course, bad statistical practice, but a good way to see how the mechanism works.

TRAINING_WORKDIR = "data/training"
shape = pd.read_csv(TRAINING_WORKDIR + "/iris.csv", header=None)

import itertools

a = [50 * i for i in range(3)]
b = [40 + i for i in range(10)]
indices = [i + j for i, j in itertools.product(a, b)]

test_data = shape.iloc[indices[:-1]]
test_X = test_data.iloc[:, 1:]
test_y = test_data.iloc[:, 0]

predictor = Predictor(
    endpoint_name=endpoint_name, sagemaker_session=None, serializer=CSVSerializer()
)

print(predictor.predict(test_X.values).decode("utf-8"))

Cleanup endpoint

model.sagemaker_session.delete_endpoint(endpoint_name)
model.sagemaker_session.delete_endpoint_config(endpoint_name)
model.delete_model()

Notebook CI Test Results

This notebook was tested in multiple regions. The test results are as follows, except for us-west-2 which is shown at the top of the notebook.