Bright Wire is an open source machine learning library for .NET with GPU support (via CUDA)
Clone or download
Jack Dermody
Latest commit 82c231e Oct 18, 2018

README.md

Bright Wire

Bright Wire is an extensible machine learning library for .NET with GPU support (via CUDA).

Getting Started

Bright Wire runs "out of the box" for CPU based computation on .net 4.6 and above and .net standard 2. For GPU based computation, you will need to install NVIDIA CUDA Toolkit 10 (and have a Kepler or better NVIDIA GPU).

To enable higher performance CPU based computation, Bright Wire also supports the Intel Math Kernel Library (MKL) via the Numerics.Net Wrapper.

Tutorials

Nuget Installation

To install the .net4 version (no CUDA support, any CPU) use:

Install-Package BrightWire.Net4

To add CUDA support (.net4, x64 only) use:

Install-Package BrightWire.CUDA.Net4.x64

Note: When using the CUDA version, make sure that the /cuda/brightwire.ptx file is copied to the output directory (Properties/Copy To Output Directory).

To install the .net standard version for use with Xamarin or UWP use:

Install-Package BrightWire

Recompiling the PTX

It's highly likely that your GPU supports different CUDA capabilities than the precompiled brightwire.ptx in this repository. You can find what is your capability level here. It's a number, ex. 3.0, 3.5, that you use for specifying compute_XX and sm_XX parameters.

If you get an ErrorNoBinaryForGPU exception, that means you have to recompile. The instructions are here.

Example command for NVIDIA GeForce GTX770M (CUDA 3.0)

nvcc kernel.cu -use_fast_math -ptx -m 64 -arch compute_30 -code sm_30 -o kernel.ptx

Linux Support

Without CUDA

You can use Bright Wire on Mono (tested with 4.6.2/Fedora 25) out of the box, no additional setting up is needed.

With CUDA

Bright Wire can also work with CUDA on Mono. When you build your solution, you will need to extract ConfigForLinux.zip archive from here to your output path. That way, CUDA won't look for nvcuda on Linux, but for libcuda shared object. You can even run on your Optimus enabled laptop (tested with GTX770M with Bumblebee) with optirun mono [binary_name].

Another issue you may have is that protobuf library complains that it is already referencing NETStandard library. NuGet version is a bit older on Mono, so please try with the latest NuGet binary from their website. That way, all the libraries get pulled correctly.

Features

Connectionist aka "Deep Learning"

  • Feed Forward, Convolutional and Bidirectional network architectures
  • LSTM, GRU, Simple, Elman and Jordan recurrent neural networks
  • L2, Dropout and DropConnect regularisation
  • Relu, LeakyRelu, Sigmoid, Tanh and SoftMax activation functions
  • Gaussian, Xavier and Identity weight initialisation
  • Cross Entropy, Quadratic and Binary cost functions
  • Momentum, NesterovMomentum, Adagrad, RMSprop and Adam gradient descent optimisations

Bayesian

  • Naive Bayes
  • Multinomial Bayes
  • Multivariate Bernoulli
  • Markov Models

Unsupervised

  • K Means clustering
  • Hierachical clustering
  • Non Negative Matrix Factorisation
  • Random Projection

Linear

  • Regression
  • Logistic Regression
  • Multinomial Logistic Regression

Tree Based

  • Decision Trees
  • Random Forest

Ensemble Methods

  • Stacking

Other

  • K Nearest Neighbour classification
  • In-memory and file based data processing

Dependencies