Image Studio reference
This page walks through the key components of using Image Studio for your image classification applications:
- Prerequisites to get started
- Exploring your data
- Annotating your data
- Creating labeling functions
- Training models
- Exporting your data
Prerequisite: Enable dedicated resources
Because image data is large and computationally expensive, you must first enable dedicated resources on your application before developing labeling functions (LFs) and models. Snorkel Flow will then allocate compute resources so that you have faster and more performant backend operations on much larger dataset sizes. This allows you to scale up to 500K data points in a split sample. For step-by-step instructions on how to enable dedicated resources on your application, see Using dedicated resources to speed up applications with large datasets.
Snorkel may take time to cache the dataset upon initial load.
When you are finished, be sure to de-allocate dedicated resources to free up memory for others to use.