Tutorial: How to Get Started with AI Computing on OmniTensor

6 min readOct 10, 2024

OmniTensor offers an easy way to share your GPU resources through its decentralized AI computing network. Follow the steps below to start contributing your GPU to advanced AI tasks and earn rewards in the form of OMNIT tokens.

Step 1: Access the OmniTensor AI Computing Page

To begin, go to the AI Computing section of the OmniTensor platform. This is where you’ll link your GPU, begin computing tasks, and monitor your contributions in real time.

Step 2: Connect Your EVM-Compatible Wallet

Once you’re on the AI Computing page, connect a wallet that supports Ethereum Virtual Machine (EVM) networks (like MetaMask). Here’s how:

Connect Wallet: Click the “Connect Wallet” button on the platform.
Switch to the OmniTensor Testnet: When prompted, allow your wallet to add and switch to the OmniTensor EVM Chain Testnet. This enables your wallet to interact with the platform.
Sign the Invitation Transaction: After switching networks, you’ll be prompted to sign a transaction. There are no transaction fees involved; the signature simply registers your invitation code and grants access to the AI Computing section.

Step 3: Start AI Computing

With your wallet connected and invitation registered, you’re ready to start contributing your GPU.

After connecting your wallet, the system will automatically detect your GPU and display its details in the “GPU Information” section.
Start Computing: Click the “Start AI Computing” button to begin. Your GPU will start working on tasks assigned by the network, and you can track your progress on the page.

Side Notes:

OmniTensor is currently in its testnet phase, meaning your GPU won’t be under a real computational load. No actual resources are used, and the points earned during this phase will convert into OMNIT tokens after the Token Generation Event (TGE).
Every AI task computed is recorded on the blockchain as a unique transaction tagged as AIComputing. You can view the details of each task by checking the transaction records in the “Details” section.

Step 4: Earn Rewards

As you compute tasks, you’ll earn OmniPoints. Each task you complete earns 0.1 points, which will later be converted into OMNIT tokens when the platform transitions beyond the testnet phase. If others are working on the same task, the task finishes faster, but points are still earned individually by each contributor.

Maximizing Your Contribution

You can invite others to join the platform using your unique invitation code. For each new user who signs up using your code, you’ll earn 10% of the points they accumulate as a bonus, increasing your rewards.
The more users contributing, the faster larger tasks are completed. While tasks are split for efficiency, points are distributed individually, ensuring fair compensation for all contributors.

OmniTensor’s AI Computing Interface Overview

OmniTensor’s AI Computing interface, available at app.omnitensor.io/computing, offers an easy-to-use platform where individuals can share their GPU’s computational power to support decentralized AI tasks. This interface allows users to join a global network, contributing unused GPU resources for AI computations. Participants help drive advanced AI projects and, in return, earn OmniPoints, which are convertible to OMNIT tokens, OmniTensor’s native cryptocurrency.

Interface Breakdown

Left Panel Menu:

The left side of the interface contains a navigation menu with several key options:

Home: Returns you to the main dashboard.
AI Computing: The primary interface where users can share their GPU resources to compute AI tasks and earn rewards.
Onchain Tasks: Offers additional tasks tied to OmniTensor’s decentralized ecosystem, likely involving smart contracts or blockchain validation activities.
AI Models: Provides access to available AI models that users can utilize or help train.
Leaderboard: Displays rankings of top contributors within the OmniTensor network, showing a sense of community and competition.

Top-Left GPU Information:

In the upper-left section of the interface, you will see technical details about your connected GPU. This includes the vendor, renderer, and specific technical details like the driver version, Direct3D version, and more. This ensures that users know precisely which hardware is being utilized and its status.

Top-Right Invitation Code Field:

In the top-right, the “Invitation Code” field allows new users to enter an invitation code to join the platform. Invitations help grow the network and allow participants to earn additional rewards. This system encourages users to share the platform with others, promoting decentralized AI growth.

Middle-Right Invitation Links:

Just below the invitation code field is a personalized invitation link and code unique to each user. This link can be shared with others to invite them to join OmniTensor’s AI Computing platform. As an incentive, users who invite others receive 10% of the points accumulated by those they invite. This adds an additional layer of gamification and reward to the platform’s community-driven nature.

Bottom-Left Task Information:

The bottom-left panel shows active AI tasks being computed by the user’s GPU. It displays:

Task Number: A unique identifier for the AI task (e.g., Task #33679).
Task Type: The specific AI activity being computed (e.g., Fine-tuning a pre-trained AI model).
Active Contributors: The number of contributors working on this task at any given time.
Completion Percentage: A progress bar showing how much of the task has been completed by all contributors.
Points Earned: The total number of points the user has accumulated for contributing to AI tasks.
Start/Stop AI Computing Button: A central control to begin or pause the AI computation process.

Bottom-Right Resource Load:

On the bottom-right of the interface is a real-time graph showing the resource load over the last 15 seconds. This graph visually represents the power usage and computational load that your GPU is contributing, allowing users to monitor how much their hardware is being utilized for AI tasks.

OmniTensor is paving the way for decentralized AI computing, enabling anyone with a GPU to contribute to global AI development while earning rewards. By following these simple steps, you can connect your GPU, start computing AI tasks, and accumulate points that will be converted into OMNIT tokens later. Whether you’re interested in exploring new decentralized technologies or just want to make use of idle GPU power, OmniTensor offers a straightforward and rewarding platform.

By bringing together a community of contributors, OmniTensor helps address the growing demand for AI computation while creating an ecosystem where participants are rewarded fairly. As the platform evolves from testnet to mainnet, this is a unique opportunity to be part of the decentralized AI movement from the ground up.

Join the OmniTensor network today and start contributing to the future of AI.