Skip to main content

​
Installation & Setup

Get started with Nolano.AI by choosing the deployment option that best fits your research needs and infrastructure requirements.

Nolano Cloud

Managed cloud platform - start training immediately with zero setup

Self-Hosted Installation

Install on your own infrastructure for maximum control and customization

​
Nolano Cloud (Coming Soon)

The fastest way to get started with foundation model research. Our managed cloud platform provides instant access to high-performance compute with zero infrastructure management.

​
Features

  • No installation required
  • Pre-configured environments with all dependencies
  • Immediate access to GPU clusters
  • Automatic scaling based on workload
  • High-performance A100 and H100 GPU clusters
  • Distributed training across multiple nodes
  • Automatic fault tolerance and recovery
  • Built-in monitoring and logging
  • Team workspaces and shared experiments
  • Version control for models and datasets
  • Experiment tracking and comparison
  • Secure data handling and privacy

​
Getting Started with Nolano Cloud (Coming Soon)

1

Sign Up

Create your account at nolano.ai and verify your email.
2

Choose Your Plan

Select a plan that fits your research needs:

Lab

$99/month
  • 100 GPU hours/month
  • Multi-node training
  • Priority support

Enterprise

Custom pricing
  • Unlimited usage
  • Dedicated clusters
  • Custom SLAs
3

Access Your Workspace

Once your account is set up, the Nolano team will provide you with access credentials and setup instructions for your workspace. You’ll receive:
  • CLI installation package and authentication tokens
  • Workspace setup guide tailored to your research needs
Getting Access: Reach out to the Nolano team at [email protected] to request access to Nolano Cloud and receive your personalized setup instructions.
4

Upload Your Data

The Nolano team will guide you through the data upload process, which includes:
  • Secure data transfer protocols
  • Data validation and preprocessing options
  • Integration with your existing data pipelines
The team will provide you with the appropriate tools and access methods based on your data size and security requirements.
New to foundation models? Our cloud platform includes guided tutorials and example datasets to help you get started quickly.

​
Self-Hosted Installation

Install Nolano.AI on your own infrastructure for complete control over your research environment, data privacy, and compute resources.

​
System Requirements

  • OS: Ubuntu 20.04+ or CentOS 8+
  • Python: 3.9 or higher
  • GPU: CUDA-compatible GPU with 16GB+ VRAM
  • RAM: 32GB system memory
  • Storage: 500GB available disk space
  • Network: High-bandwidth internet for model downloads

​
Installation Steps

1

Prerequisites

Install system dependencies and CUDA toolkit:
# Update system packages
sudo apt update && sudo apt upgrade -y

# Install essential build tools
sudo apt install -y build-essential git curl wget

# Install CUDA Toolkit (12.1 recommended)
wget https://developer.download.nvidia.com/compute/cuda/12.1.0/local_installers/cuda_12.1.0_530.30.02_linux.run
sudo sh cuda_12.1.0_530.30.02_linux.run

# Add CUDA to PATH
echo 'export PATH=/usr/local/cuda/bin:$PATH' >> ~/.bashrc
echo 'export LD_LIBRARY_PATH=/usr/local/cuda/lib64:$LD_LIBRARY_PATH' >> ~/.bashrc
source ~/.bashrc
2

Python Environment Setup

Create and configure a Python environment:
# Install Python 3.11 and pip
sudo apt install -y python3.11 python3.11-pip python3.11-venv

# Create virtual environment
python3.11 -m venv nolano-env
source nolano-env/bin/activate

# Upgrade pip and install essential packages
pip install --upgrade pip setuptools wheel
3

Get Nolano.AI Access

Contact the Nolano team to receive the installation package and setup instructions:

Request Self-Hosted Access

Email [email protected] with:
  • Your research institution/organization
  • Intended use case and research goals
  • Hardware specifications and cluster details
  • Timeline for your project
The team will provide you with:
  • Custom installation package tailored to your environment
  • Detailed setup guide for your specific hardware configuration
  • License keys and authentication credentials
  • Direct support during installation process
Nolano.AI is currently in private beta. Access is granted on a case-by-case basis to qualified research institutions and organizations.
4

Installation & Verification

The Nolano team will guide you through the installation process and help verify that everything is working correctly. This includes:
  • Running system diagnostics to ensure hardware compatibility
  • Configuring distributed training settings for your cluster
  • Testing GPU detection and CUDA integration
  • Validating network connectivity for multi-node setups
The team provides hands-on support during installation to ensure optimal performance for your specific research environment.
5

Configuration

The Nolano team will help you configure the system for your specific environment:
  • GPU count and memory allocation
  • Mixed precision training settings
  • Gradient checkpointing optimization
  • Hardware-specific performance tuning
  • Cache and checkpoint directory configuration
  • Data storage path optimization
  • Shared storage for multi-node clusters
  • Backup and recovery settings
  • Multi-node communication setup
  • Network backend configuration (NCCL/Gloo)
  • Fault tolerance and recovery settings
  • Load balancing across nodes
The team provides pre-configured templates based on your hardware setup and research requirements.

​
Multi-Node Setup

For distributed training across multiple machines, the Nolano team provides comprehensive cluster setup support:
The team will help you with:
  • Network Architecture Planning: Design optimal interconnect topology for your cluster
  • Shared Storage Setup: Configure high-performance distributed filesystems (NFS, Lustre, or custom solutions)
  • Node Communication: Establish secure and efficient inter-node communication protocols
  • Load Balancing: Implement dynamic workload distribution across available resources
  • Fault Tolerance: Set up automatic failover and recovery mechanisms
  • Performance Optimization: Fine-tune settings for your specific hardware configuration
Multi-node deployments require careful planning and configuration. The Nolano team ensures optimal performance and reliability for your distributed training environment.

​
Docker Installation

For containerized deployments, the Nolano team provides:

Pre-built Containers

Ready-to-use Docker images with Nolano.AI pre-installed and optimized for different GPU architectures

Custom Containerization

Tailored Docker configurations for your specific research environment and deployment requirements
  • Optimized Base Images: Pre-configured containers with all dependencies
  • GPU Integration: CUDA and cuDNN optimization for containerized environments
  • Orchestration Support: Kubernetes and Docker Swarm deployment configurations
  • Volume Management: Efficient data and model storage strategies
  • Security Configuration: Container security best practices and access controls
Contact [email protected] to request access to pre-built containers or custom containerization support.

​
Support & Troubleshooting

The Nolano team provides comprehensive support for installation and ongoing maintenance:

Get Support

For any installation or technical issues, contact the Nolano team:
  • Email: [email protected]
  • Emergency Support: Available for production deployments
  • Response Time: Within a single call to the founders phone numbers

​
Model Setup

After installation, you’re ready to start your foundation model research:

Data Preparation

Learn how to prepare and tokenize your datasets

Model Configuration

Explore different model architectures and settings

Training Configuration

See more on configuring your model
Need Help? Reach out to support for assistance with installation and setup.