SciSharp/NumSharp

High Performance Computation for N-D Tensors in .NET, identical API to NumPy.

What it solves

NumSharp provides a native .NET implementation of the NumPy API, allowing developers to perform high-performance scientific computing and machine learning operations directly in C# or F# without needing to embed or interface with the Python runtime.

How it works

It implements an N-dimensional array (NDArray) structure that mimics NumPy's behavior, including broadcasting, slicing, and dtype-aware operations. The library achieves high performance by using runtime-generated Intermediate Language (IL) code and SIMD (Single Instruction, Multiple Data) CPU acceleration, which translates numerical operations into highly optimized machine code at runtime.

Who it’s for

This project is for .NET developers building machine learning infrastructure, scientific applications, or numerical utilities who require the familiar NumPy programming model within a managed, native .NET environment.

Highlights

  • Native .NET implementation of NumPy-style array operations (no Python dependency).
  • Support for N-dimensional arrays with advanced features like broadcasting and view-based slicing.
  • High-performance execution via runtime-specialized IL generation and SIMD CPU acceleration.
  • Extensive API coverage aiming for compatibility with NumPy 2.x.
  • Part of the broader SciSharp ecosystem for scientific and machine learning development in .NET.

Related

  • Project
    NVIDIA/deepops

    Infrastructure automation tools that use Ansible to deploy and optimize GPU-enabled Kubernetes and Slurm clusters, specifically for NVIDIA DGX systems.

  • Project
    Lightning-AI/lightning-thunder

    A source-to-source compiler for PyTorch that optimizes models for training and inference through a Pythonic intermediate representation and extensible transformation plugins.

  • Project
    sb-ai-lab/LightAutoML

    An automated machine learning framework that allows users to create ML models via ready-to-use presets or by building custom pipelines for tabular, text, image, and time series data.

  • Project
    MaxHalford/prince

    Prince is a Python library that provides efficient implementations of multivariate exploratory data analysis methods, such as PCA and correspondence analysis, using a familiar scikit-learn style API.