Installation Guide
{: .no_toc }
Complete guide to installing AiDotNet and its dependencies. {: .fs-6 .fw-300 }
Table of contents
{: .no_toc .text-delta }
- TOC {:toc}
Basic Installation
Using .NET CLI
dotnet add package AiDotNet
Using Package Manager Console
Install-Package AiDotNet
Using PackageReference
Add to your .csproj file:
<ItemGroup>
<PackageReference Include="AiDotNet" Version="*" />
</ItemGroup>
Platform Support
| Platform | Status | Notes |
|---|---|---|
| Windows x64 | ✅ Full | Recommended |
| Linux x64 | ✅ Full | Ubuntu 20.04+ |
| macOS x64 | ✅ Full | macOS 11+ |
| macOS ARM64 | ✅ Full | Apple Silicon |
.NET Version Support
| Version | Status |
|---|---|
| .NET 8.0 | ✅ Primary target |
| .NET 7.0 | ✅ Supported |
| .NET 6.0 | ✅ Supported |
| .NET Framework 4.6.2+ | ✅ Supported |
GPU Acceleration
NVIDIA CUDA
For GPU-accelerated training on NVIDIA GPUs:
Install CUDA Toolkit 12.x+ (or 11.8+ minimum)
- Download from NVIDIA CUDA Downloads
- CUDA 12.x recommended for best performance with modern GPUs
Install cuDNN 9.x+ (or 8.6+ minimum)
- Download from NVIDIA cuDNN
- cuDNN 9.x recommended for compatibility with CUDA 12.x
Verify installation
nvcc --version
OpenCL
For AMD/Intel GPU support:
- Install OpenCL runtime for your GPU vendor
- Install CLBlast for optimized BLAS operations
Optional Packages
Model Serving
For production model serving:
dotnet add package AiDotNet.Serving
Dashboard
For training visualization:
dotnet add package AiDotNet.Dashboard
Verifying Installation
Create a simple test:
using AiDotNet;
Console.WriteLine("AiDotNet installed successfully!");
// Test basic functionality
var builder = new AiModelBuilder<double, double[], double>();
Console.WriteLine("AiModelBuilder created.");
Run it:
dotnet run
Troubleshooting
Package Not Found
Ensure you have the latest NuGet sources:
dotnet nuget list source
GPU Not Detected
- Verify CUDA installation:
nvcc --version - Check GPU visibility:
nvidia-smi - Ensure CUDA path is in environment variables
Build Errors
- Clear NuGet cache:
dotnet nuget locals all --clear - Restore packages:
dotnet restore - Rebuild:
dotnet build
Next Steps
- Quick Start Tutorial - Build your first model
- GPU Setup Guide - Detailed GPU configuration