Gensyn Testnet Node Guide

Buy cheap GPU at : ACCESS

Follow Me: https://x.com/WhalePiz

Telegram Group: https://t.me/Nexgenexplore

PLEASE SELECT THE HARDWARE YOU WANT TO RUN ON

💻 System Requirements

Requirement	Details
CPU Architecture	arm64 or amd64
Recommended RAM	25 GB
CUDA Devices (Recommended)	RTX 3090, RTX 4090, A100, H100
Python Version	Python >= 3.10 (For Mac, you may need to upgrade)

💻 Run node

1. Install sudo

apt update && apt upgrade -y

2. Install other dependencies

apt install screen curl iptables build-essential git wget lz4 jq make gcc nano automake autoconf tmux htop nvme-cli libgbm1 pkg-config libssl-dev libleveldb-dev tar clang bsdmainutils ncdu unzip libleveldb-dev  -y

3. Install Python

apt install python3 python3-pip python3-venv python3-dev -y

4.Install Node.js and yarn

apt update
curl -fsSL https://deb.nodesource.com/setup_22.x | bash -
apt install -y nodejs
node -v
npm install -g yarn
yarn -v

5. Install Yarn

curl -o- -L https://yarnpkg.com/install.sh | bash

export PATH="$HOME/.yarn/bin:$HOME/.config/yarn/global/node_modules/.bin:$PATH"

source ~/.bashrc

6. Create a screen session

screen -S gensyn

7. Clone the Repository and Run

cd $HOME && rm -rf gensyn-testnet && git clone https://github.com/whalepiz/gensyn-testnet.git && chmod +x gensyn-testnet/gensyn.sh && ./gensyn-testnet/gensyn.sh

• You will get a link like this slick-cars-shake.loca.lt
• Access the website and enter your IP address as I have highlighted in red
• Then log in with your email and return to the original terminal to continue the process

Answer prompts:

• Would you like to push models you train in the RL swarm to the Hugging Face Hub? [y/N] >>> Press N to join testnet
  • HuggingFace needs 2GB upload bandwidth for each model you train, you can press Y, and enter your access-token.
• Enter the name of the model you want to use in huggingface repo/name format, or press [Enter] to use the default model. >>> For default model, press Enter or choose one of these (More model parameters (B) need more vRAM):
  • Gensyn/Qwen2.5-0.5B-Instruct
  • Qwen/Qwen3-0.6B
  • nvidia/AceInstruct-1.5B
  • dnotitia/Smoothie-Qwen3-1.7B
  • Gensyn/Qwen2.5-1.5B-Instruct

Backup Instructions for swarm.pem

VPS:

1. Use Mobaxterm to establish a connection to your VPS.
2. Once connected, transfer the swarm.pem file from the following directory to your local machine:

   /root/rl-swarm/swarm.pem
   

WSL (Windows Subsystem for Linux):

1. Open Windows Explorer and search for \wsl.localhost to access your Ubuntu directories.
2. The primary directories are:
   • If installed under a specific username:

     \wsl.localhost\Ubuntu\home\<your_username>\rl-swarm
     

   • If installed under root:

     \wsl.localhost\Ubuntu\root\rl-swarm
     

3. Locate the swarm.pem file within the rl-swarm folder.

GPU Servers (e.g., Hyperbolic):

1. Open Windows PowerShell and execute the following command to connect to your GPU server:

   sftp -P PORT ubuntu@xxxx.hyperbolic.xyz
   

   • Replace ubuntu@xxxx.hyperbolic.xyz with your GPU server's hostname.
   • Replace PORT with the specific port from your server's SSH connection details.
   • The username might vary (e.g., root), depending on your server configuration.

2. After establishing the connection, you will see the sftp> prompt.

3. To navigate to the folder containing swarm.pem, use the cd command:

   cd /home/ubuntu/rl-swarm
   

4. To download the file, use the get command:

   get swarm.pem
   

   The file will be saved in the directory where you ran the sftp command, typically:

   • If you executed the command in PowerShell, the file will be stored in C:\Users\<your_pc_username>.

5. Once the download is complete, type exit to close the connection.

---

You've successfully backed up the swarm.pem file from your VPS, WSL, or GPU server.