Haidra-Org/AI-Horde

A crowdsourced distributed cluster for AI art and text generation

What it solves

AI Horde is a community-powered distributed inference cluster that makes advanced AI generation (text and images) accessible to everyone, regardless of their hardware capabilities or financial means. It solves the problem of high compute costs and hardware requirements for running large AI models like Stable Diffusion or Llama, by allowing volunteers to share their idle GPU power.

How it works

The system operates as middleware that connects users making requests for AI content with volunteer "workers" who provide the compute resources. When a user submits a request, the system finds an available worker capable of handling the job. To maintain fairness and encourage contribution, it uses a "kudos" system: workers earn kudos for processing requests, and users with more kudos receive higher priority in the generation queues. Kudos cannot be bought or sold.

Who it’s for

  • Users: Individuals who want to generate AI art or text without needing a powerful GPU.
  • Workers: People with GPUs who want to contribute their computing power to the community.
  • Developers: Those looking to integrate AI generation capabilities into their apps, games, or scripts via a REST API.
  • Enterprises: Organizations thating that want to deploy a private, scalable ML-Ops distributed inference cluster.

Highlights

  • Community-Driven: Completely free and community-run, ensuring equitable access to AI technology.
  • Multimodal: Supports both image and text generation.
  • Scalable: Designed as an enterprise-level ML-Ops cluster that can scale effortlessly to accommodate varying workloads.
  • Fair Priority: A kudos-based economy that rewards contributors while keeping the service free for anonymous users.

Related

  • Project
    Mesh-LLM/mesh-llm

    A distributed LLM serving system that pools GPUs and memory across multiple machines to run large models via an OpenAI-compatible API.

  • Project
    hyperspaceai/hyperspace-node

    A decentralized P2P AI inference network that allows users to contribute GPU, CPU, and storage resources to run AI models and earn rewards.

  • Project
    robertvoy/ComfyUI-Distributed

    A ComfyUI extension that enables distributed and parallel processing across multiple GPUs and machines to accelerate image and video generation and upscaling.

  • Project
    Project-HAMi/HAMi

    A Kubernetes-native virtualization middleware that enables sharing and isolating GPU and other AI accelerator resources across multiple workloads to increase hardware utilization.

  • Project
    hpcaitech/ColossalAI

    A distributed deep learning framework that makes training and inference for large AI models faster and cheaper through advanced parallelism and memory management.