Continual Learning Guard (EWC / SI / Progressive Networks / Replay)

Community

Guard continual learning across tasks.

Authorsovr610
Version1.0.0
Installs0

System Documentation

What problem does it solve?

This skill standardizes a modular continual-learning guard that preserves prior-task performance while incorporating new knowledge across sequential tasks.

Core Features & Use Cases

  • Elastic Weight Consolidation (EWC) and online variants to anchor important parameters after each task.
  • Synaptic Intelligence (SI), Progressive Networks, Experience Replay, and PackNet-style strategies to balance plasticity and stability.
  • Real-world use cases include sequential ML deployments, robotics, and adaptive content systems that must remember prior skills.

Quick Start

Instantiate and configure the Continual Learning Guard on a model to protect prior task performance while learning new tasks.

Dependency Matrix

Required Modules

torch

Components

scriptsreferencesassets

💻 Claude Code Installation

Recommended: Let Claude install automatically. Simply copy and paste the text below to Claude Code.

Please help me install this Skill:
Name: Continual Learning Guard (EWC / SI / Progressive Networks / Replay)
Download link: https://github.com/sovr610/refffiy/archive/main.zip#continual-learning-guard-ewc-si-progressive-networks-replay

Please download this .zip file, extract it, and install it in the .claude/skills/ directory.
View Source Repository

Agent Skills Search Helper

Install a tiny helper to your Agent, search and equip skill from 471,000+ vetted skills library on demand.