layer-inference-planner

Community

Plan layer-by-layer AirLLM inference across mesh

Author47network
Version1.0.0
Installs0

System Documentation

What problem does it solve?

Plans AirLLM-style layer-by-layer inference across mesh devices to enable running large models on constrained hardware.

Core Features & Use Cases

  • Single-device sequential inference planning
  • Multi-device distributed planning with activation transfer and latency budgeting
  • Visualization and timing estimates to compare strategies Use cases include planning 70B+ models across several devices with VRAM constraints.

Quick Start

Instruct the planner to generate a plan by providing action set to plan_single along with model_id, total_layers, activation_size_mb, and available_vram_mb.

Dependency Matrix

Required Modules

@sven/compute-mesh/layer-inference

Components

Standard package

💻 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: layer-inference-planner
Download link: https://github.com/47network/Sven/archive/main.zip#layer-inference-planner

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.