Inference.net Devnet Epoch 3

Distributed GPU network for AI inference. Epoch 3 features Solana integration, stake-weighted routing, and enhanced operator rewards.

Overview

Inference.net distributes AI inference workloads (DeepSeek R1, Gemma 3, Llama 3.3 70B) across GPU operators worldwide. Operators earn $INT points based on uptime, performance, and GPU contributions.

Requirements

• GPU: 16GB+ VRAM
• RAM: 16GB+
• Storage: 30GB free
• Network: 100 Mbps+ (test speed)

• RTX 4090, A100, H100

• Ubuntu 22.04 or 24.04 LTS with CUDA 12.2+
• Docker pre-installed

Prerequisites

• Spheron account (sign up)
• Payment method configured
• SSH key (how to generate)

Step 1: Deploy GPU on Spheron

1. Sign up at app.spheron.ai
2. Add credits - Click Credits button → Add funds (card/crypto)
3. Deploy:
   • Click Deploy in sidebar
   • Select GPU: RTX 4090, A100, or H100 (16GB+ VRAM)
   • Region: Closest to you
   • OS: Ubuntu 22.04 or 24.04 with CUDA 12.2+
   • Select your SSH key or upload a new one
   • Click Deploy Instance

Instance ready in 30-60 seconds.

Step 2: Connect to Instance

ssh root@your-instance-ip

Step 3: Verify Docker

docker --version

4. Install NVIDIA Container Toolkit

sudo apt-get install -y nvidia-container-toolkit
sudo nvidia-ctk runtime configure --runtime=docker
sudo systemctl restart docker

Step 4: Register on Inference Devnet

1. Visit devnet.inference.net
2. Sign in with Google or GitHub
3. Click Create Worker
4. Select Docker tab

Step 5: Run Worker Command

Copy command from Step 3 on Inference.net site and run in your terminal:

# Use actual command from Inference.net site
docker run -d --gpus all ...

Wait for Docker image download and setup (few minutes).

Step 6: Verify Worker Status

1. Return to devnet.inference.net
2. Click Check Status
3. Confirm: Worker is running! ✅
4. Click Next Step

Step 7: Join Community & Access Dashboard

• Join official channels (Discord, Telegram)
• Click Continue to Dashboard

Your node is now active and earning $INT points!

Verification

Check worker status:

• Inference.net dashboard shows active worker
• Logs display successful initialization
• $INT points accumulating

Troubleshooting

# Check Docker logs
docker ps -a
docker logs <container-id>
 
# Verify GPU access
docker run --rm --gpus all nvidia/cuda:12.2.0-base-ubuntu22.04 nvidia-smi

# Reinstall NVIDIA toolkit
sudo apt-get install -y nvidia-container-toolkit
sudo nvidia-ctk runtime configure --runtime=docker
sudo systemctl restart docker

• Verify network speed: fast.com
• Check firewall allows required ports
• Ensure stable internet connection

• Check instance is running on Spheron dashboard
• Restart Docker: sudo systemctl restart docker
• Rerun worker command from Inference.net

Managing Your Node

• Check Inference.net dashboard for $INT points
• View worker logs for activity
• Monitor GPU usage: nvidia-smi -l 1

docker stop <container-id>

• Rerun worker command from Inference.net dashboard

Additional Resources

• Inference.net Docs
• Getting Started - Spheron deployment guide
• SSH Connection - SSH setup
• Docker - Startup scripts with Docker

Epoch 3 Updates: Solana blockchain integration, improved stake-weighted routing, enhanced rewards, simplified operator experience.