fortytwo-node

Fortytwo Node Operator FOR Inference with CPU Mode

banner

This repository contains installation instructions and helper scripts for setting up the FortyTwo Network console application on Linux systems

Example Dashboard If Need Buidl

Sample index.html https://arcxteam.github.io/fortytwo-node/example/

1. Get Realtime Node by Telegram Bot

Use tool-bot monitoring FortyTwo node operator in performs swarm model inference! Get a detail dashboard ranking, received token, daily report & tracking any transaction FOR 42Tokens in participant network

Got bot here… https://greyscope.xyz/x/fortytwo

Key Bot Features:

2. Installing Setup Node

Update Services & Depedency

sudo apt update && sudo apt upgrade -y &&
sudo apt install -y \
automake autoconf build-essential clang curl \
gcc git htop iptables jq libatomic1 libblas3 libclang-dev \
libgbm1 liblapack3 liblapack-dev libleveldb-dev libomp-dev \
libopenblas-dev libgomp1 libopenmpi-dev libssl-dev lz4 make nano \
ncdu ninja-build nvme-cli pkg-config \
python3-pip screen tar tmux unzip wget

Create Folder

mkdir -p ~/FortytwoCLI && cd ~/FortytwoCLI

Download Execute by Officially

curl -L -o fortytwo-console-app.zip https://github.com/Fortytwo-Network/fortytwo-console-app/archive/refs/heads/main.zip
unzip fortytwo-console-app.zip
cd fortytwo-console-app-main
# will delete official script linux.sh
rm linux.sh

Use P2P a Peers Network (optional)

wget https://raw.githubusercontent.com/arcxteam/fortytwo-node/refs/heads/main/.p2p_known_peers.json

Redirect Installation Script or Cloning Repo

wget https://raw.githubusercontent.com/arcxteam/fortytwo-node/refs/heads/main/linux.sh

Create a screen & Running Operator

screen -S fortytwo

chmod +x linux.sh && ./linux.sh

# back to screen log
screen -r fortytwo

Note; After installation, the FortyTwo console application be ready run. If not yet received 42T after a few minutes or hourly.. check your logs, if get like here

Solutions

Error Cause & Custom LLModel

The script downloaded FortytwoCapsule-linux-amd64-cuda124 (GPU version), which requires CUDA libraries libcuda.so.1. On CPU-only servers without NVIDIA drivers, this library is missing, causing the load failure.

Why It Happened

Fix

  1. Manual Download Capsule & Replace: ```bash cd ~/FortytwoCLI/fortytwo-console-app-main/FortytwoNode

rm -f FortytwoCapsule

LATEST=$(curl -s https://download.swarminference.io/capsule/latest)

wget “https://download.swarminference.io/capsule/v${LATEST}/FortytwoCapsule-linux-amd64”
-O FortytwoCapsule

chmod +x FortytwoCapsule


2. **Rerun Script**:
```bash
cd ~/FortytwoCLI/fortytwo-console-app-main
rm linux.sh
wget https://raw.githubusercontent.com/arcxteam/fortytwo-node/refs/heads/main/linux.sh
chmod +x linux.sh && ./linux.sh
  1. Another Custom Model GGUF
    • Costum model by import plz select 1
    • Need use LLM_HF_REPO and LLM_HF_MODEL_NAME
    • Visit Huggingface
    • Check repo file model need a parameter key is GGUF & llama.cpp
    • Select low LLModel medium at 1GB-3GB file and use token 0.5B-3B with quantity param Q3_K_M or Q4_K_M or Q5_K_M
    • A list custom-parameter GGUF LLModels for Text, Code & Math https://arcxteam.github.io/fortytwo-node/llmodel.html

    Example my custom LLModel

    LLM_HF_REPO="unsloth/DeepSeek-R1-Distill-Qwen-1.5B-GGUF"
LLM_HF_MODEL_NAME="DeepSeek-R1-Distill-Qwen-1.5B-Q4_K_M.gguf"
NODE_NAME="⬢ MATH EQUATIONS & REASONING: DeepSeek-R1-Distill-Qwen-1.5B Q4"

    image

Notes