FRAME-FM Technical Documentation¶

Framework for the Rapid Development of Environmental Foundation Models.

FRAME-FM is an open-source framework for building scalable machine learning foundation models on large environmental datasets, including archives from the UK NERC Environmental Data Service (EDS).

The framework provides structured pipelines for:

Dataset Ingestion
Data Transformation
Model Training
Experiment Tracking
Configuration Management

It is designed for both experienced machine learning practitioners and scientists who are new to ML but want to work with large environmental datasets.

Architecture¶

Configs
Data Wrangling
Observability
Transforms

API Reference¶

Autogenerated documentation for FRAME-FM modules.

FRAME_FM
- FRAME_FM package

Project Links¶

GitHub Repository: https://github.com/NERC-EDS/FRAME-FM
Technical Documentation: https://nerc-eds.github.io/FRAME-FM/

Indices and tables¶

Index
Module Index
Search Page