mir 0.16.0-beta1

Mir

Generic Numerical Library for Science and Machine Learning.

Contents

• mir.ndslice Multidimensional Random Access Ranges and Arrays
• mir.sparse Sparse Tensors
• Sparse - DOK format
• Different ranges for COO format
• CompressedTensor - CSR/CSC formats
• mir.sparse.blas - Sparse BLAS for CompressedTensor
• mir.model.lda.hoffman - Online variational Bayes for latent Dirichlet allocation (Online VB LDA) for sparse documents. LDA is used for topic modeling.
• mir.combinatorics Combinations, combinations with repeats, cartesian power, permutations.
• mir.las.sum Functions and Output Ranges for Summation Algorithms. Works with user-defined types.
• Precise algorithm: improved analog of Python's fsum
• Pairwise algorithm: fast version for Input Ranges
• Kahan, KBN, and KB2 algorithms
• mir.blas - this is slow snail, it is for experiments with BLAS API. Don't use for now.

In progress

• mir.random - non-uniform RNGs.
• mir.blas - BLAS in D.

Documentation

Alpha version of API is available here.

Compatibility

	Linux	Mac OS X	Windows
DMD 64
DMD 32		OS X >= 10.7 is x86-64 only
LDC 64			by ignoring issue #120
LDC 32		OS X >= 10.7 is x86-64 only	by ignoring issue #120

• DMD (reference D compiler) >= 2.070
• LDC (LLVM D Compiler) >= 1.0.0.

Installation

Fast setup with the dub package manager

Dub is the D's package manager. You can create a new project with:

dub init --format=json <project-name>

Now you need to edit the dub.json add mir as dependency and set its targetType to executable.

{
	...
	"dependencies": {
		"mir": "~><current-version>"
	},
	"targetType": "executable"
}

Now you can create a main file in the source and run your code with:

dub

You can use a different compile with dub --compiler ldc. For a performance boost, add -b release to let the compiler perform additional optimizations, inlining, removal of bound checking and assert statements.

Arch Linux

Type	Name	Version
Stable	libmir
Latest	libmir-git

Contributing

See our TODO List. Mir is very young and we are open for contributing to source code, documentation, examples and benchmarks.

Notes

• mir.ndslice is a development version of the std.experimental.ndslice package.