Meta-Learning Suite (MAML/FOMAML/Reptile + MAML++ Enhancements)

Community

Meta-learning testbed for rapid adaptation.

Authorsovr610
Version1.0.0
Installs0

System Documentation

What problem does it solve?

The Meta-Learning Suite standardizes the learn-to-adapt pipeline, providing a self-contained framework for designing, evaluating, and reproducing gradient-based meta-learning algorithms (MAML, FOMAML, Reptile) with MAML++ enhancements.

Core Features & Use Cases

  • Supports second-order and first-order inner loops through pluggable backends (torch.func, higher, or custom SGD).
  • Implements MAML++ enhancements: per-layer per-step learning rates (LSLR), multi-step loss (MSL), derivative-order annealing, and per-step batch-norm handling.
  • Deterministic episodic sampling, comprehensive checkpointing, and a consistent evaluation interface for research and prototyping.

Quick Start

Define a model, configure a meta-algorithm (MAML or FOMAML), and run a training loop to observe the inner-loop adaptations and outer meta-gradients.

Dependency Matrix

Required Modules

torchpytest

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: Meta-Learning Suite (MAML/FOMAML/Reptile + MAML++ Enhancements)
Download link: https://github.com/sovr610/refffiy/archive/main.zip#meta-learning-suite-maml-fomaml-reptile-maml-enhancements

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.