Develop Reference

debug and deploy featurizer (data processor for imodel inference) of sagemaker endpoint

I am looking at this example to implement the data processing of incoming raw data for a sagemaker endpoint prior to model inference/scoring. This is all great but I have 2 questions:
How can one debug this (e.g can I invoke endpoint without it being exposed as restful API and then use Sagemaker debugger)
Sagemaker can be used "remotely" - e.g. via VSC. Can such a script be uploaded programatically?
Thanks.

Sagemaker Debugger is only to monitor the training jobs.
https://docs.aws.amazon.com/sagemaker/latest/dg/train-debugger.html
I dont think you can use it on Endpoints.
The script that you have provided is used both for training and inference. The container used by the estimator will take care of what functions to run. So it is not possible to debug the script directly. But what are you debugging in the code ? Training part or the inference part ?
While creating the estimator we need to give either the entry_point or the source directory. If you are using the "entry_point" then the value should be relative path to the file, if you are using "source_dir" then you should be able to give an S3 path. So before running the estimator, you can programmatically tar the files and upload it to S3 and then use the S3 path in the estimator.

Updating Sagemaker Endpoint with new Endpoint Configuration

A bit confused with automatisation of Sagemaker retraining the model.
Currently I have a notebook instance with Sagemaker LinearLerner model making the classification task. So using Estimator I'm making training, then deploying the model creating Endpoint. Afterwards using Lambda function for invoke this endpoint, I add it to the API Gateway receiving the api endpoint which can be used for POST requests and sending back response with class.
Now I'm facing with the problem of retraining. For that I use serverless approach and lambda function getting environment variables for training_jobs. But the problem that Sagemaker not allow to rewrite training job and you can only create new one. My goal is to automatise the part when the new training job and the new endpoint config will apply to the existing endpoint that I don't need to change anything in API gateway. Is that somehow possible to automatically attach new endpoint config with existing endpoint?
Thanks

Yes, use the UpdateEndpoint endpoint. However, if you are using the Python Sagemaker SDK, be aware, there might be some documentation floating around asking you to call
model.deploy(..., update_endpoint=True)
This is apparently now deprecated in v2 of the Sagemaker SDK:
You should instead use the Predictor class to perform this update:
from sagemaker.predictor import Predictor
predictor = Predictor(endpoint_name="YOUR-ENDPOINT-NAME", sagemaker_session=sagemaker_session_object)
predictor.update_endpoint(instance_type="ml.t2.large", initial_instance_count=1)

If I am understanding the question correctly, you should be able to use CreateEndpointConfig near the end of the training job, then use UpdateEndpoint:
Deploys the new EndpointConfig specified in the request, switches to using newly created endpoint, and then deletes resources provisioned for the endpoint using the previous EndpointConfig (there is no availability loss).
If the API Gateway / Lambda is routed via the endpoint ARN, that should not change after using UpdateEndpoint.

Making a Prediction Sagemaker Pytorch

I have trained and deployed a model in Pytorch with Sagemaker. I am able to call the endpoint and get a prediction. I am using the default input_fn() function (i.e. not defined in my serve.py).
model = PyTorchModel(model_data=trained_model_location,
role=role,
framework_version='1.0.0',
entry_point='serve.py',
source_dir='source')
predictor = model.deploy(initial_instance_count=1, instance_type='ml.m4.xlarge')
A prediction can be made as follows:
input ="0.12787057, 1.0612601, -1.1081504"
predictor.predict(np.genfromtxt(StringIO(input), delimiter=",").reshape(1,3) )
I want to be able to serve the model with REST API and am HTTP POST using lambda and API gateway. I was able to use invoke_endpoint() for this with an XGBOOST model in Sagemaker this way. I am not sure what to send into the body for Pytorch.
client = boto3.client('sagemaker-runtime')
response = client.invoke_endpoint(EndpointName=ENDPOINT ,
ContentType='text/csv',
Body=???)
I believe I need to understand how to write the customer input_fn to accept and process the type of data I am able to send through invoke_client. Am I on the right track and if so, how could the input_fn be written to accept a csv from invoke_endpoint?

Yes you are on the right track. You can send csv-serialized input to the endpoint without using the predictor from the SageMaker SDK, and using other SDKs such as boto3 which is installed in lambda:
import boto3
runtime = boto3.client('sagemaker-runtime')
payload = '0.12787057, 1.0612601, -1.1081504'
response = runtime.invoke_endpoint(
EndpointName=ENDPOINT_NAME,
ContentType='text/csv',
Body=payload.encode('utf-8'))
result = json.loads(response['Body'].read().decode())
This will pass to the endpoint a csv-formatted input, that you may need to reshape back in the input_fn to put in the appropriate dimension expected by the model.
for example:
def input_fn(request_body, request_content_type):
if request_content_type == 'text/csv':
return torch.from_numpy(
np.genfromtxt(StringIO(request_body), delimiter=',').reshape(1,3))
Note: I wasn't able to test the specific input_fn above with your input content and shape but I used the approach on Sklearn RandomForest couple times, and looking at the Pytorch SageMaker serving doc the above rationale should work.
Don't hesitate to use endpoint logs in Cloudwatch to diagnose any inference error (available from the endpoint UI in the console), those logs are usually much more verbose that the high-level logs returned by the inference SDKs

How can I load a third-party ML model into a serverless function

I have built an ML model (using the sklearn module), and I want to serve it predictions via AWS API Gateway + Lambda function.
My problems are:
I can't install sklearn + numpy etc. because of the lambda capacity limitations. (the bundle is greater than 140MB)
Maybe that a silly question, but, do you know if there are better ways to do that task?
I've tried this tutorial, in order to reduce the bundle size. However, it raises an exception because of the --use-wheel flag.
https://serverlesscode.com/post/scikitlearn-with-amazon-linux-container/
bucket = s3.Bucket(os.environ['BUCKET'])
model_stream = bucket.Object(os.environ['MODEL_NAME'])
model = pickle.loads(model_stream)
model.predict(z_features)[0]
Where z_features are my features after using a scalar

Just figure it out!
The solution basically stands on top of the AWS Lambda Layers.
I created a sklearn layer that contains only the relevant compiled libraries.
Then, I run sls package in order to pack a bundle which contains those files together with my own handler.py code.
The last step was to run
sls deploy --package .serverless
Hope it'll be helpful to others.

If you simply want to serve your sklearn model, you could skip the hassle of setting up a lambda function and tinkering with the API Gateway - just upload your model as a pkl file to FlashAI.io which will serve your model automatically for free. It handles high traffic environments and unlimited inference requests. For sklearn models specifically, just check out the user guide and within 5 minutes you'll have your model available as an API.
Disclaimer: I'm the author of this service

How to deploy a Keras CNN Model to webservice?

I am currently in the process of trying to deploy a Keras Convolutional Neural Network for a webservice.
I had tried converting my saved keras hdf5 model to a tensorflow.js model and deploying that but it slowed down the client-side app as the model is relatively robust and thus, takes a large amount of space in the client memory.
Thus, I am trying to figure out a way to deploy the model in the cloud and make predictions through a request from the web-app with an image, and then receive a response which holds the prediction tensor. I know that gcloud may have some similar abilities or feature but I am unsure of how to get started.
Essentially, I am asking if there is any service that will allow me to deploy a pre-trained and saved convolutional neural network model to which I can send images in a request and use the model to return a predicted tensor?

You can export a trained Keras model and serve it with TensorFlow Serving. TF Serving allows to host models and call them via either gRPC or REST requests. You could deploy a flask app with an endpoint that accepts an image, wraps it as a payload and calls your model via the requests module.
Your code to export the model as a servable would look like this:
import tensorflow as tf
# The export path contains the name and the version of the model
model = keras.models.load_model('./mymodel.h5')
# Feth the Keras session and save the model
with keras.backend.get_session() as sess:
tf.saved_model.simple_save(
sess,
export_dir,
inputs={'images': model.input},
outputs={t.name:t for t in model.outputs})
This will store the files necessary for TF Serving. From this directory, you can host the model as follows:
tensorflow_model_server --model_base_path=$(pwd) --rest_api_port=9000 --model_name=MyModel
Your request would then look like this:
requests.post('http://ip:9000/v1/models/MyModel:predict', json=payload)
Where payload is a dictionary that contains your request image.

If you want a click-to-deploy solution for serving your model on Google Cloud, consider using Cloud ML Engine's Online Prediction service. First, follow the instructions in #sdcbr's response to export your SavedModel. Copy the model to GCS then you simply create a model and a version:
gcloud ml-engine models create "my_image_model"
gcloud ml-engine versions create "v1"\
--model "my_image_model" --origin $DEPLOYMENT_SOURCE
Or, even easier, use Cloud Console to do the above with a few clicks!
You will get a serverless REST endpoint that includes authentication and authorization, autoscaling (including scale to zero) as well as logging and monitoring, without having to write or maintain a line of code.