Frequently Asked Questions

Gemini Models

If you’re running older Gemini models (like 1.5 Flash or Pro), consider switching to newer versions if available and listed as active on the Dria Edge AI Dashboard. Not all models receive tasks equally. To earn points, you generally need to run models that are actively being requested by users on the network. Check the dashboard to see which models are currently processing tasks.

Gemini may be unavailable in your region — if so, there’s no fix for that at the moment.

If you are using a Free Gemini account, there might be rate limits imposed by Google. If your node stops earning due to API limits, you may need to wait for the limits to reset or consider a paid tier if usage is high.

Running Multiple Nodes

You can run multiple nodes under the same IP address or Wi-Fi network, but you must use a unique private key (wallet) for each node instance. Using the same private key for multiple active nodes can lead to conflicts and issues with task assignment and rewards.

Avoid Using WSL

If you’re using WSL (Windows Subsystem for Linux), please avoid it for running the Dria node launcher. It is known to cause compatibility issues.

Instead, open a native Windows terminal (like CMD or PowerShell) as Administrator and run:

powershell -c "irm https://dria.co/launcher.ps1 | iex"

System Requirements

There isn’t a strict minimum, as you can run API-based models (Gemini, OpenAI, etc.) which rely on external services, not your local hardware. However, if you want to run local models using Ollama, your system needs to be powerful enough to achieve at least 15 Eval TPS for the specific model you choose. Refer to the Selecting Models page for performance benchmarks on various hardware configurations.

Node Status

Visit the Dria Edge AI Dashboard and log in with the wallet associated with your node. The dashboard shows your node’s status (Online/Offline), recent activity, and earned points. If the status is “Online” and you see points accumulating (even if intermittently), your node is working.

Earning Points

There could be several reasons:

No tasks: There might be low demand for the specific model(s) your node is running. Check the dashboard and consider switching to more active models.
Node offline: Ensure your node is running and connected to the internet. Check the dashboard for its status.
Performance issues (Local Models): If running an Ollama model, ensure your system meets the performance requirements (>= 15 Eval TPS). Use dkn-compute-launcher measure.
API Key Issues (API Models): If running API models, ensure your API keys are correct, valid, and have sufficient quota/credits.
Network Issues: Firewalls or network configurations might prevent your node from communicating effectively. Ensure the required ports (default 4001) are open.

Refer to the Rewards page for more details on the earning mechanism.

Ollama Requirements

No. If you only plan to serve models via APIs (like OpenAI, Gemini, OpenRouter), you do not need to install Ollama. You only need Ollama if you intend to run models locally on your own hardware.

Points Updates

Points shown on the Dria Edge AI Dashboard are updated in near real-time as your node successfully completes tasks.

Performance Metrics

TPS stands for Tokens Per Second. It’s a common metric used to measure the processing speed of AI models. In the context of Dria and the measure command, we primarily look at Eval TPS to determine if a local Ollama model is fast enough to participate effectively in the network (generally >= 15 TPS).

Getting Started

Help

Gemini Models

Running Multiple Nodes

Avoid Using WSL

System Requirements

Node Status

Earning Points

Ollama Requirements

Points Updates

Performance Metrics

Getting Started

Help

​Gemini Models

​Running Multiple Nodes

​Avoid Using WSL

​System Requirements

​Node Status

​Earning Points

​Ollama Requirements

​Points Updates

​Performance Metrics

Gemini Models

Running Multiple Nodes

Avoid Using WSL

System Requirements

Node Status

Earning Points

Ollama Requirements

Points Updates

Performance Metrics