Atoma Docs
  • Developers
  • Introduction
    • Atoma Basics
    • Compute Layer
    • Trust and Privacy
  • Developers
    • Atoma architecture overview
    • Atoma smart contract
      • Atoma architecture
      • Interact with the Atoma smart contract
      • Current available tasks on Atoma
  • Node Operators
    • Atoma node
      • Install Sui cli
      • Install CUDA
      • Install docker
      • Firewall permissions
      • HuggingFace authentication
      • Atoma node deployment
      • Atoma node daemon
        • Node subscriptions
        • Task Management
        • Stack operations
        • Attestation/Dispute management
        • Claimed Stacks management
        • Submit new node registration transaction
        • Task subscriptions
        • Submit Stack settlements and attestations
        • Claim Stacks
Powered by GitBook
On this page
  1. Node Operators
  2. Atoma node

Install CUDA

Cuda installation for inference service

PreviousInstall Sui cliNextInstall docker

Last updated 6 months ago

In order to run the inference service on Atoma, Nvidia cuda toolkit is required to be installed locally. Please follow the official NVIDIA installation guide for you OS:

https://developer.nvidia.com/cuda-downloads?