Service Provider on NuNet
Please read NuNets Disclaimer before installing any software on your devices.
Basic Steps to Run an ML Job on NuNet as a Service Provider:
Download and install the Device Management Service (DMS): Install the DMS on your machine, which will enable you to request ML jobs by connecting you to compute provider machines that are onboarded on the NuNet platform.
Download and install the Service Provider Dashboard (SPD): Install the Service Provider Dashboard on your local machine, making it accessible for you to submit and monitor your ML jobs. No sign-up or NuNet account is needed.
Obtain compute resources: Open your preferred browser and visit
localhost:9991to specify the type of compute resources you require (CPU or GPU) and the amount of resources needed for your ML job.Define your ML job: Specify your ML model URL, and provide any additional dependencies required for the execution of the ML job.
Set up your wallet: Connect your Cardano wallet (Eternl or Nami wallet) to the Service Provider Dashboard. Ensure that your wallet has enough ADA and NTX tokens for running the ML job and covering transaction fees. For Public Alpha we are using the PreProd Cardano Network.
Set a budget: Determine the maximum amount of NTX tokens you are willing to spend on the ML job. This budget will be locked in the smart contract on the PreProd Cardano Network as a guarantee for the compute providers.
Submit your ML job: Review your job configuration and submit the ML job to the NuNet platform. The platform will automatically match your job with suitable compute providers based on the resources needed.
Monitor your job: Track the progress of your ML job through the Service Provider Dashboard. You can check the log outputs of your job every 2 minutes since your job begins. Please wait for around 5 minutes for the first log to appear when deploying new ML jobs.
Review the results: Once the ML job is completed, you can download the output data and review the results. The locked NTX tokens will be released and distributed as a reward to the compute providers who contributed resources to your job.
Repeat the process (optional): If you have more ML jobs to run, you can follow the same steps to execute them on the NuNet platform, utilizing its decentralized computing resources.
[Add YouTube video on running the Public Alpha as a Service Provider.]
Last updated