⚙️ Installation RL-Swarm

🛠️ Hardware Requirements

RL Swarm is an open source system for peer-to-peer reinforcement learning over the internet. Running a swarm node allows you to train your personal model against the swarm intelligence. Each swarm performs RL reasoning as a group, with a gossiping system (Hivemind) for collaborative improvement between models. You can also connect your node to the Gensyn Testnet, to receive an on-chain identity that tracks your progress over time.

RL Swarm is fully open and permissionless, meaning you can run it on a basic consumer laptop at home or on a powerful GPU in the cloud. You can also experiment with different models to see which ones perform best.

Components	RAM	Recommendation
CPU	Minimum 16GB RAM	More RAM (e.g., 32GB+) recommended for larger models or datasets
GPU	CUDA-supported devices: NVIDIA RTX 3090 NVIDIA RTX 4090 NVIDIA A100 NVIDIA H100	Optional: node can run in CPU-only mode without a GPU

Additional requirements

You will install all additional utilities following the guide.

• Nodejs

• Screen

• Yarn

• Docker and Docker-Compose

• Python >=3.10 (for Mac, you will likely need to upgrade)

• Unoccupied port 3000

📝 Manual installation

1️⃣ Installation packeges and dependencies

Install dependencies for building from source

sudo apt-get update && sudo apt-get upgrade -y

Install Nodejs

curl -sSL https://raw.githubusercontent.com/zunxbt/installation/main/node.sh | bash

Install other dependencies

sudo apt update && sudo apt install -y python3 python3-venv python3-pip curl screen git yarn && curl -sS https://dl.yarnpkg.com/debian/pubkey.gpg | sudo apt-key add - && echo "deb https://dl.yarnpkg.com/debian/ stable main" | sudo tee /etc/apt/sources.list.d/yarn.list && sudo apt update && sudo apt install -y yarn

Install docker

curl -fsSL https://download.docker.com/linux/ubuntu/gpg | sudo gpg --dearmor -o /usr/share/keyrings/docker-archive-keyring.gpg

echo "deb [arch=$(dpkg --print-architecture) signed-by=/usr/share/keyrings/docker-archive-keyring.gpg] https://download.docker.com/linux/ubuntu $(lsb_release -cs) stable" | sudo tee /etc/apt/sources.list.d/docker.list > /dev/null

sudo apt-get update
sudo apt-get install docker-ce docker-ce-cli containerd.io
docker version

Install docker-compose

VER=$(curl -s https://api.github.com/repos/docker/compose/releases/latest | grep tag_name | cut -d '"' -f 4)

curl -L "https://github.com/docker/compose/releases/download/"$VER"/docker-compose-$(uname -s)-$(uname -m)" -o /usr/local/bin/docker-compose

chmod +x /usr/local/bin/docker-compose
docker-compose --version

Add Python

sudo apt-get install python3 python3-pip
sudo apt install python3.10-venv

Install Yarn

curl -o- -L https://yarnpkg.com/install.sh | sh
export PATH="$HOME/.yarn/bin:$HOME/.config/yarn/global/node_modules/.bin:$PATH"
export PATH="$HOME/.yarn/bin:$HOME/.config/yarn/global/node_modules/.bin:$PATH"
source ~/.bashrc

Check Node JS, Yarn and Docker version before start

docker version
node -v
yarn -v

2️⃣ Create huggin face token

Go to the https://huggingface.co/ and create account

After that, go to the settings - https://huggingface.co/settings/tokens and create a token

Select Write, enter a name, and create a token

Save this token, we will use it later

3️⃣ Install RL Swarm node

Clone the repo

git clone https://github.com/gensyn-ai/rl-swarm/
cd rl-swarm

Run the swarm in background using screen

screen -S swarm

Install swarm

python3 -m venv .venv
source .venv/bin/activate
./run_rl_swarm.sh

Once started, noda will ask to go to the URL (http://localhost:3000) to login.

Waiting for it shows to Waiting for userData.json to be created..

Waiting for userdata.json will be until you log in to the browser, so for now we need to minimize the screen with the key combination Ctrl + A + D, then we will use Ngrok to log in, and then return to this screen

4️⃣Login to the dashboard

Go to your browser and enter it in the address bar:

VPSIP - IP of the server from which you are installing

http://VPSIP:3000/

After you have logged in, return to the terminal

If you cannot log in or do not receive an access code use Ngrok:

Downoload Ngrok

wget https://bin.equinox.io/c/bNyj1mQVY4c/ngrok-v3-stable-linux-amd64.tgz

tar -xvzf ngrok-v3-stable-linux-amd64.tgz
sudo mv ngrok /usr/local/bin/

Go to ngrok.com and register

Next, go to your personal account: Ngrok dashboard

Copy the token

Run the command with your authtoken in the server terminal:

ngrok config add-authtoken YOUR_TOKEN

We're throwing localhost:3000

ngrok http 3000

Log out of the Ngrok interface you see in front of you now - after successful login through the browser

Then click Visit Site

Click Login

Log in with Google, if you have any problems then you can enter your mail above, get the code and log in

If everything is fine, you will see this

After that, you can return to the terminal, exit the Ngrok interface with the key combination Ctrl + C, and return to the screen with the command: screen -r swarm

You will see:

Would you like to push models you train in the RL swarm to the Hugging Face Hub? [y/N] Enter y

You will need to enter your Huggin Face token that you created earlier

Next, save these 3 words - this is the name of your node

After that, you can see the following: your model is training - but you need to wait until the progress reaches from 0/20 to 20/20

and you will see the following:

The process will be repeated, so keep an eye on the node's performance, and follow the announcements in the discord for updates and possible maintenance, etc.

5️⃣ Backup

1- Node name

• Now your node started running, Find your name after word Hello, like mine is whistling hulking armadillo (You can use CTRL+SHIFT+F to search Hello in terminal)

2- Node .pem file

• Save swarm.pem file in this directory: $HOME/rl-swarm/

6️⃣ Check Gensyn Dashboard

1. https://dashboard.gensyn.ai/ You can follow the link to the dashboard and find your node in the leaderboard by 3 words that you saved earlier (But at the moment, this function does not work very well)