Installation

Follow this guide to get your node up and running

Introduction

This guide will walk you through the installation process for running a Nesa node.


Download and Execute the Bootstrap Script

To download and run the bootstrap script, use the following command:

# For all operating systems
bash <(curl -s https://raw.githubusercontent.com/nesaorg/bootstrap/master/bootstrap.sh)

Hint: Ensure Docker is running before executing the script.

View Your Node's Performance

Visit https://node.nesa.ai and enter your Node ID.

You can find your Node ID by re-running the bootstrap script (above) and reading the Node ID from the script's header.


Node Configuration

During the script execution, you will be prompted to configure your node:

  1. Moniker: Choose a unique name for your node.

    # Example prompt
    Choose a moniker for your node: <Your Node Name>
  2. Node Type: Select whether your node will be a Validator or a Miner.

    # Example prompt
    What type(s) of node is <Your Node Name>?
    [ ] Validator
    [ ] Miner
    • Validator: Enter the private key for the validator.

      # Example prompt
      Validator's Private Key: <Your Private Key>
    • Miner:

      • Miner Type: Choose between Distributed Miner or Non-Distributed Miner.

        # Example prompt
        What type of miner will <Your Node Name> be?
        [ ] Distributed Miner
        [ ] Non-Distributed Miner
      • Model Selection:

        • For Distributed Miner: Select an existing swarm or start a new one.

          # Example prompt for Distributed Miner
          Would you like to join an existing swarm or start a new one?
          [ ] Join existing swarm
          [ ] Start a new swarm

          If starting a new swarm:

          # Example prompt
          Which model would you like to run? (meta-llama/Llama-2-13b-Chat-Hf)
        • For Non-Distributed Miner: Enter the model name to run.

          # Example prompt for Non-Distributed Miner
          Which model would you like to run? (meta-llama/Llama-2-13b-Chat-Hf)

Understanding Swarms

A swarm in the context of Nesa is an orchestrator with one or more miners working collaboratively to handle inference tasks. Distributed miners can join an existing swarm to contribute to an ongoing effort or start a new swarm to initiate a new collaborative mining process.

Swarm Configuration Process:

  1. Select an Existing Swarm:

    • You will be presented with a list of available swarms.

    • Select the swarm you want to join.

    • The script will configure your node to connect to the selected swarm.

  2. Start a New Swarm:

    • Enter the model name you want to run.

    • The script will set up a new swarm for this model.

    • Your node will act as the initial member of this new swarm, and others can join later.


Final Setup

After completing the configuration:

  1. The script will set up the working directory and clone the necessary repositories.

  2. Docker containers will be started using docker-compose.

To check the status of your node:

# Check Docker containers status
docker ps

Note: Ensure all required Docker containers are running.


Additional Help/Support

If you need any additional help or support, please visit the Nesa Discord for community support and discussion. You can also explore additional documentation to deepen your understanding of Nesa and its features.

Last updated