markat1/mlops

Productionizing an ML Model

1. Initial Setup

1. Create a GitHub Project:
   • Create a new repository on GitHub.
   • Clone the repository to your local machine.
2. Install GitHub Desktop:
   • Download and install GitHub Desktop.
3. Download and Setup VSCode:
   • Download and install Visual Studio Code.
   • Open your project folder in VSCode.
   • Set up a virtual environment in VSCode.

2. AWS IAM Setup

1. Create IAM User with Proper Permissions:

   • Go to the IAM console in AWS.

   • 

   • 

   • Create a new IAM user with programmatic access.

   • Attach policies for Amazon S3, Amazon ECR, and Amazon SageMaker (e.g., AmazonS3FullAccess, AmazonEC2ContainerRegistryFullAccess, AmazonSageMakerFullAccess).

     

     

     

     

2. Configure AWS CLI with IAM User Credentials:
   • Install AWS CLI.
   • Run aws configure and input the IAM user’s access key, secret key, default region, and output format.

AWS Access Key ID:
AWS Secret Access Key:
Default region name: us-east-2
Default output format: json

3. Project Structure

1. Create Project Files and Folders:
   • main.py: This is your main code that controls the entry into training and inference scripts:
   • train.py: Your main training script.
   • requirements.txt: List of dependencies.
   • inference.py: Your main inference script.
   • data/: Directory to store your dataset.
2. Download Data from Kaggle:
   • Download the dataset from Kaggle.
   • Place the dataset in the data/ folder.
3. Install Dependencies:
   • Install the dependencies listed in requirements.txt within your virtual environment. pip install -r requirements.txt
4. Run the Training Script:
   • Ensure training.py runs successfully in your environment.

4. Dockerization

1. Install Docker:
   • Download and install Docker.
2. Create Dockerfile:
   • Write a Dockerfile to containerize your application. More detailed explanation about the Docker file can be found here
3. Build and Run Docker Image:
   • Build the Docker image.
   • Run the Docker container and verify the results.

5. S3 Integration

1. Upload Data to S3:

   • Upload your dataset to an S3 bucket.
   • Ensure your IAM user has s3:PutObject and s3:GetObject permissions.

2. Modify Code for S3 Integration:

   • Update your code to download the data from S3 at runtime using Boto3.

   import boto3
   s3 = boto3.client('s3')
   s3.download_file('your-bucket-name', 'your-dataset.csv', 'data/your-dataset.csv')

6. Docker Image on ECR

1. Setup ECR Repository:

   • Create a repository on Amazon Elastic Container Registry (ECR).

2. Push Docker Image to ECR:

   • Authenticate Docker to your ECR using IAM credentials.
   • Tag and push your Docker image to ECR.

   aws ecr get-login-password --region your-region | docker login --username AWS --password-stdin your-account-id.dkr.ecr.your-region.amazonaws.com
   docker tag your-image:latest your-account-id.dkr.ecr.your-region.amazonaws.com/your-repo:latest
   docker push your-account-id.dkr.ecr.your-region.amazonaws.com/your-repo:latest
   
   

3. Download Image from ECR:

   • Pull the Docker image from ECR and run it locally to verify.

7. SageMaker Endpoint Integration

1. Deploy Model on SageMaker:
   • Write and deploy your model to a SageMaker endpoint.
2. Call SageMaker Endpoint Locally:
   • Write code to call the SageMaker endpoint from your local environment.
3. Test the End-to-End Pipeline:
   • Test the entire workflow from local code execution to SageMaker inference.

8. GitHub Integration

1. Push Code to GitHub:

   • Commit and push your local project to GitHub.

2. Setup GitHub Actions for CI/CD:

   • Add a build workflow to automate building and pushing your Docker image to ECR.
   

3. Execute the Build Locally:

   • Run the build workflow locally to ensure it works as expected.