mvsusp’s gists

mvsusp / dataset.ipynb

Created December 10, 2018 22:56

Sorry, something went wrong. Reload?

Sorry, we cannot display this file.

Sorry, this file is invalid so it cannot be displayed.

mvsusp / predict.java

Last active January 27, 2022 22:09

How to make predictions against a SageMaker endpoint

	package org.example.basicapp;

	import com.amazonaws.services.sagemakerruntime.AmazonSageMakerRuntime;
	import com.amazonaws.services.sagemakerruntime.AmazonSageMakerRuntimeClientBuilder;
	import com.amazonaws.services.sagemakerruntime.model.InvokeEndpointRequest;
	import com.amazonaws.services.sagemakerruntime.model.InvokeEndpointResult;

	import java.nio.ByteBuffer;

	public class App

mvsusp / create-endpoint.sh

Last active July 22, 2018 19:39

How to create an AWS SageMaker endpoint for predictions - creating the endpoint

	#!/usr/bin/env bash

	MODEL_NAME=half-plus-three-v1

	ENDPOINT_CONFIG_NAME=half-plus-three-config-v1

	ENDPOINT_NAME=half-plus-three

	PRODUCTION_VARIANTS="VariantName=Default,ModelName=${MODEL_NAME},"\
	"InitialInstanceCount=1,InstanceType=ml.c4.large"

mvsusp / create-model.sh

Last active July 22, 2018 19:30

How to create an AWS SageMaker endpoint for predictions - creating a model

	#!/usr/bin/env bash

	MODEL_NAME=half-plus-three-v1

	# the role named created with
	# https://gist.github.com/mvsusp/599311cb9f4ee1091065f8206c026962
	ROLE_NAME=SageMakerRole

	# the name of the image created with
	# https://gist.github.com/mvsusp/07610f9cfecbec13fb2b7c77a2e843c4

mvsusp / create-role.sh

Last active July 22, 2018 18:41

How to create a SageMaker Execution role

	#!/usr/bin/env bash

	# This script creates a role named SageMakerRole
	# that can be used by SageMaker and has Full access to S3.

	ROLE_NAME=SageMakerRole

	# WARNING: this policy gives full S3 access to container that
	# is running in SageMaker. You can change this policy to a more
	# restrictive one, or create your own policy.

mvsusp / push.sh

Last active July 25, 2018 04:08

How to push a Docker image to AWS ECS

	#!/usr/bin/env bash

	IMAGE_NAME=sagemaker-tf-serving
	REGION=$(aws configure get region)
	ACCOUNT=$(aws sts get-caller-identity --query Account --output text)

	# If the ECS repository doesn't exist, creates it.
	aws ecr create-repository --repository-name ${IMAGE_NAME} > /dev/null 2>&1

	# ECR requires the image name to be in this format:

mvsusp / build.sh

Last active July 21, 2018 22:02

How to host a TensorFlow or Keras model in AWS SageMaker - Building the image and testing the container

	#!/usr/bin/env bash

	docker build -t sagemaker-byoc-tf-serving .

mvsusp / nginx.conf

Last active July 21, 2018 21:55

How to host a TensorFlow or Keras model in AWS SageMaker - nginx.conf

	events {
	# determines how many requests can simultaneously be served
	# https://www.digitalocean.com/community/tutorials/how-to-optimize-nginx-configuration
	# for more information
	worker_connections 2048;
	}

	http {
	server {
	# configures the server to listen to the port 8080

mvsusp / Dockerfile

Created July 21, 2018 20:58

Creating the Docker container

	# Using the official tensorflow serving image from docker hub as base image
	FROM tensorflow/serving

	# Installing NGINX, used to rever proxy the predictions from SageMaker to TF Serving
	RUN apt-get update && apt-get install -y --no-install-recommends nginx git

	# Copy our model folder to the container
	COPY saved_model_half_plus_three /saved_model_half_plus_three

	# Copy NGINX configuration to the container

mvsusp / download_saved_model_half_plus_three.sh

Last active July 21, 2018 22:12

Downloading saved_model_half_plus_three

	git clone --single-branch -b r1.9 https://github.com/tensorflow/serving.git /tmp/serving

	cp -r \
	/tmp/serving/tensorflow_serving/servables/tensorflow/testdata/saved_model_half_plus_three \
	saved_model_half_plus_three

Marcio Vinicius dos Santos mvsusp