ScarCity
A framework for training ML models without data.
v0.4.0-beta / Active
Development
Overview
High-level introduction to the SCARCITY v1.0.0 framework.
Architecture
The layered architecture of SCARCITY.
Core Components
Module structure and key exports.
Engine (MPIE)
Multi-Path Inference Engine documentation.
Federation
Decentralized learning infrastructure.
Meta-Learning
Cross-domain adaptation and optimization.
Simulation
Agent-based modeling and visualization.
Resource Governor
Dynamic resource management and adaptation.
Stream Processing
Adaptive data ingestion and rate control.
Runtime System
Event bus and asynchronous architecture.
FMI
High-level orchestration for federated learning workflows.
Math Foundations
Algorithms and theoretical basis of SCARCITY.
API Reference
Quick reference for key exports and usage.
Implementation
Performance characteristics and requirements.