Module 1: Introduction to HPE AI Essentials Software

• Introduction to HPE AI Essentials:

  • Features and capabilities
  • Data science components
  • Navigation
  • How to get started

Module 2: Build and Train Models in Notebooks or Pipelines Using Kubeflow

• Using Kubeflow Notebooks on HPE AI Essentials
• Define hyperparameter tuning
• Configure Katib on HPE AI Essentials
• Create and manage Kubeflow notebook servers on HPE AI Essentials
• Hyperparameter tuning for news recommendation (Ray Tune)

Module 3: Workflows Using Feast

• Feast and its components
• Design Workflows using Feast in HPE AI Essentials
• Feast ride sharing use case

Module 4: HPE ML Development Environment (MLDE) and GPU MIG Partitioning

• The HPE ML Development Environment (MLDE)
• Configuring MLDE for air-gapped environment
• GPU MIG partitioning on HPE AI Essentials

Module 5: Manage ML Lifecycle

• Using Notebooks with MLflow on HPE AI Essentials
• Defining KServe
• MLflow bike sharing use case

Module 6: Scalable AI/ML Applications

• Ray support on HPE AI Essentials
• Ray components and terminologies
• Designing anomaly detection application on HPE AI Essentials
• Rent forecasting model
• Question answering model

Lab 1: Accessing the Lab Environment

• Task 1: Initial vLabs access
• Task 2: Add user

Lab 2: Working with Kubeflow Notebooks

• Task 1: Using Kubeflow Notebooks
• Task 2: Creating and managing Kubeflow Notebook servers

  • Creating a notebook server
  • Managing a notebook server
• Task 3: Designing workflows using Feast

Lab 3: Working with ML Algorithms and Data Science in Jupyter Notebook

• Task 1: Intro to ML with multiple algorithms
• Task 2: Explanatory data analysis in data science
• Task 3: Unsupervised clustering
• Task 4: Supervised regression
• Task 5: Demonstrate ML using financial data
• Task 6: Execute digit recognition using MNIST database
• Task 7: Using various training operators

Lab 4: Working with Generative AI models

• Task 1: Pull and run custom LLM model using Docker
• Task 2: Invoke custom LLM model in Kubeflow Notebook
• Task 3: Retrieval Augmented Generation (RAG) using Langchain