Skip to content
This repository has been archived by the owner on Sep 14, 2023. It is now read-only.

danieldk/golinear

Repository files navigation

Introduction

Report card GoDoc

golinear is a package for training and using linear classifiers in the Go programming language (golang).

Installation

To use this package, you need the liblinear library. On Mac OS X, you can install this library with Homebrew:

brew install liblinear

Ubuntu and Debian provide packages for liblinear. However, at the time of writing (July 2, 2014), these were serverly outdated. This package requires version 1.9 or later.

This latest API-stable version (v1) can be installed with the go command:

go get gopkg.in/danieldk/golinear.v1

or included in your source code:

import "gopkg.in/danieldk/golinear.v1"

The package documentation is available at: http://godoc.org/gopkg.in/danieldk/golinear.v1

OpenMP

If you wish to use liblinear with OpenMP support for multicore processing, please use this command to install the package:

CGO_LDFLAGS="-lgomp" CGO_CFLAGS="-DCV_OMP" go get github.com/danieldk/golinear

Plans

  1. Port classification to Go.
  2. Port training to Go.

We will take a pragmatic approach to porting code to Go: if the performance penalty is minor, ported code will flow to the main branch. Otherwise, we will keep it around until the performance is good enough.

Examples

Examples for using golinear can be found at:

https://github.com/danieldk/golinear-examples