Sergio0694/NeuralNetwork.NET
🏗️ FrameworkSergio0694
A high-performance, TensorFlow-inspired neural network library built from scratch for the .NET ecosystem.
NeuralNetwork.NET is an open-source library that brings deep learning capabilities to C# developers. Built from scratch to target .NET Standard 2.0, it offers a familiar, TensorFlow-inspired API for constructing neural network architectures. The library features a modular design that supports various activation functions, loss functions, and optimization algorithms, including standard backpropagation and gradient descent. A standout feature is its native support for GPU acceleration; by integrating with NVIDIA's cuDNN, the library offloads intensive tensor computations to the GPU, significantly reducing training and inference times for large-scale models. It is highly optimized for performance, making it a viable choice for developers who need to integrate AI features into desktop, web, or cloud-based .NET applications while maintaining strict type safety and performance standards.
💡Highlights
- ├─Native C# neural network library
- ├─GPU acceleration via NVIDIA cuDNN
- └─TensorFlow-inspired API design
🎯For
- ├─C# Developers
- └─Machine Learning Engineers