/MLOpsTools

MLOps tools, pipelines and frameworks for flexibility and scalability in ML workflows.

Primary LanguageShell

MLOpsTools

MLOps tools, pipelines and frameworks for flexibility and scalability in ML workflows.

Building blocks of an ML system in production:

  1. Data prep:
    1.1 Data Ingestion.
    1.2 Data Analysis.
    1.3 Data Transformation.
    1.4 Data Validation. 1.5 Data Splitting.

  2. Model development + training at scale:
    2.1 Model training.
    2.2 Model Validation.
    2.3 MOdel training at scale.

  3. Deployment and Monitoring:
    3.1 Deployment.
    3.2 Serving.
    3.1 Monitoring.
    3.2 Logging.

Pros:

  • Composability.
  • Portability.
  • Scaling.

image

Amazon EKS
Pros:

  1. Fully managed kubernetes control plane.
  2. UI.
  3. Scheduling engine for multi-step ML workflows.
  4. SDK for pipeline and components manipulation.
  5. Pre packaged optimised deep learning docker containers are offered by EC2, SageMaker, EKS. (No need for further tuning).

Kubeflow on AWS (EKS) Set up:

  1. Install kubectl locally.
  2. Install AWS-CLI.
  3. Configure the aws-cli to generate config and credential files.
  4. Install eksctl.
  5. Install the aws-iam-authenticator.
  6. Create an EKS cluster using eksctl.
    6.1 Export environment variables: cluster name, region, k8s version and EC2_instance type.
  7. Create a cluster config file for use with eksctl and confirm its creation.
  8. Install Kubeflow.
  9. Create a kubeflow project
  10. Access the kubeflow UI.