CRFpp Package

A CRFpp Package is a CRF Package developed by Taku Kudo.

• AKA: CRF++, CRFpp Program.
• Context:
  • It can redefine Predictor Features (with CRF Feature Templates?).
  • It is written in C++(Cpp) with Cpp STL.
  • It can perform fast training by using LBFGS (a quasi-newton algorithm for large scale numerical optimization problem).
  • It can make efficient use of memory both in training and testing.
  • It can encode/decode in practical time
  • It can perform n-best outputs.
  • It can perform single-best MIRA training.
  • It can output marginal probabilities for all candidates.
  • It is available as an open source software.
  • It cannot support Semi-Markov CRFs.
  • It cannot support Piece-wise CRFs.
  • It does not provide an (C++/C) API.
• See: Taku Kudo.

---

---

References

• http://crfpp.sourceforge.net/
  • CRF++ is a simple, customizable, and open source implementation of Conditional Random Fields (CRFs) for segmenting/labeling sequential data. CRF++ is designed for generic purpose and will be applied to a variety of NLP tasks, such as Named Entity Recognition, Information Extraction and Text Chunking.
  • Features
    • It can redefine feature sets
    • Written in C++ with STL
    • Fast training based on LBFGS, a quasi-newton algorithm for large scale numerical optimization problem.
    • Less memory usage both in training and testing
    • encoding/decoding in practical time
    • It can perform n-best outputs
    • It can perform single-best MIRA training
    • It can output marginal probabilities for all candidates
    • Available as an open source software
  • To Do
    • Support semi-Markov CRF
    • Support piece-wise CRF
    • Provide useful C++/C API (Currently no APIs are available)

---