Getting started with Amazon SageMaker Studio | Amazon Web Services
Getting Started with SageMaker Studio
Overview of Amazon SageMaker Studio and its features for machine learning.
Introduction to SageMaker Studio
- Emily introduces the session on getting started with Amazon SageMaker Studio, a fully managed IDE designed for machine learning.
- SageMaker Studio separates the user interface from applications, allowing users to browse foundation models and explore various tools.
- Users can access multiple environments such as Jupyter Lab, SageMaker Studio Classic, VS Code, and RStudio within the platform.
Exploring JumpStart in SageMaker
- Emily navigates to the JumpStart feature in SageMaker Studio, showcasing available foundation models from leading providers.
- She selects the Mixrtal model and highlights options for training, deploying, and evaluating this specific model.
Testing Inference with Mixrtal Model
- Emily demonstrates testing an inference request using the Mixrtal 8x7B model that she has already deployed.
- She describes sending a JSON payload asking for a recipe for mayonnaise to her deployed endpoint on a multi-GPU machine (G5.48xl).
Evaluating Inference Results
- The results of the inference request are received promptly; Emily shares that she receives a simple recipe for homemade mayonnaise as a response.