pyveda: Python tools for The Veda Machine Learning API

pyveda is a package for working with machine learning data, models and inference sets stored in Veda.

Warning

pyveda is beta software, and these docs are a work in progress

Veda

Veda is DigitalGlobe’s platform for machine learning data. Veda’s goals are:

• Generate sample data from satellite imagery
• Share and distribute organized sets of data
• Share and distribute machine learning models
• Use the models and data to perform inference

pyveda is MIT licenced.

Installation:

pip install pyveda

For general information on the Veda platform and API visit http://github.com/DigitalGlobe/veda.

• Getting started
  • Getting access to Veda
  • Installing pyveda
• Accessing Data from Veda
  • Veda’s Data Storage
  • Finding Collections with pyveda
  • Accessing Collections
  • Working with the Samples
  • Using Samples with Models
  • Collection Groups
  • Batching and Transforming Data
  • Streaming Remote Data
  • Creating Releases
• Creating Data
  • Creating New Training Data
  • Sample Data
  • Creating a Collection from Spatial Data
  • Creating a Collection using the Bulk Import API
  • Adding Data to Existing Collections
  • Creating Samples from Scratch
• Features and Labels
  • Label Formats
• Model Training
• Models
  • Saving a Model
  • Uploading a Model to Veda
  • Deploying a Model
  • Searching for Models
  • Downloading a Model
  • Predicting with Veda Models
• Releases
  • Releases for Importing
• API Reference
  • Main Interface
  • Batching

Notebook Documentation

• Accessing Data from Veda
  • Veda’s Data Storage
  • Samples
  • Collections
  • Finding Collections with pyveda
• Accessing Collections
• Working with the Samples
  • Using Samples with Models
  • Collection Groups
• Batching and Transforming Data
  • Streaming Remote Data
• Creating Releases