# Install crfsuite with Homebrew, MacPorts

Fast implementation of conditional random fields. Version 0.12 via Homebrew; verified 2026-07-10. Also installable with macports: sudo port install crfsuite.

## Install

```sh
sudo av install brew:crfsuite
```

Additional install commands:

### macOS

- Homebrew (100%):

```sh
brew install crfsuite
```

  Evidence: local Homebrew formula metadata

- MacPorts (94%):

```sh
sudo port install crfsuite
```

  Evidence: MacPorts ports tree: math/crfsuite/Portfile from https://api.github.com/repos/macports/macports-ports/git/trees/master?recursive=1

## Package facts

- **Package key:** brew:crfsuite
- **Package manager:** Homebrew
- **Version:** 0.12
- **Source summary:** Fast implementation of conditional random fields
- **Homepage:** <https://www.chokkan.org/software/crfsuite/>
- **Repository:** <https://github.com/chokkan/crfsuite>
- **Last updated:** 2026-07-10T13:05:07-04:00
- **Generated:** 2026-08-03T19:37:03+00:00

## Executables

- crfsuite (alias)

## Install behavior

- Bottle: not available

## Freshness

- Page generated: 2026-08-03
- Package-manager version: 0.12
## Project history and usage

CRFsuite is Naoaki Okazaki's fast implementation of Conditional Random Fields for labeling sequential data. It provides command-line training/tagging, C++ and SWIG APIs, and training algorithms such as L-BFGS, OWL-QN, SGD, averaged perceptron, passive aggressive, and AROW.

### Project history

The official changelog records internal releases beginning with CRFsuite 0.1 on 2007-10-29 and a first public release, 0.4, on 2008-03-05. The 0.4 release added the website, documentation, a tutorial, and a CoNLL 2000 chunking performance comparison.

### Adoption history

The project page presents CRFsuite as a faster and more flexible CRF package than older template-oriented tools, with a simple data format, benchmark results, and model storage based on CQDB. Version 0.12 added Python SWIG modules and sample programs, including named entity recognition and part-of-speech tagging examples.

### How it is used

CRFsuite users train and tag sequence-labeling models with a feature-per-line data format. The official documentation emphasizes speed, training-method choice, performance evaluation during training, and programmatic use through C++ and SWIG APIs.

### Why package nerds care

For package maintainers, CRFsuite is a compact C/C++ NLP toolkit with historical source and binary releases, a BSD license, and a small command-line surface. It is also useful as a contrast point with CRF++ because its official page calls out data-format flexibility that CRF++ lacks.

### Timeline

- 2007-10-29: CRFsuite 0.1 internal release
- 2008-03-05: CRFsuite 0.4 first public release
- 2009-03-07: Version 0.6 added SGD and reduced training memory usage
- 2011-08-11: Version 0.12 optimized training, added more algorithms, revised APIs, and added Python SWIG support

### Related projects

- The official page links CRFsuite's data-format comparison to CRF++, and also points to libLBFGS and CQDB as implementation-related software.

### Sources

- <https://www.chokkan.org/software/crfsuite>
- <https://github.com/chokkan/crfsuite/blob/master/ChangeLog>


## Security Notes

No matching local secret-handling manifest was found for crfsuite. Nucleus package metadata is still published here so future coverage has a stable package URL.


## Other Package-Manager Records

- MacPorts - crfsuite: normalized package name match | MacPorts ports tree: math/crfsuite/Portfile from https://api.github.com/repos/macports/macports-ports/git/trees/master?recursive=1


## Combined YAML source

View the package source record on GitHub. [combined/crfsuite.yml](https://github.com/mxcl/pkgdb/blob/main/combined/crfsuite.yml)


## Sources

- pkg.so package database
- curated package history
- pkgdb category and tag curation
- external package-manager database matches
- cross-ecosystem install command graph