HISTORY

18/12/2013
  Release v1.5.0: Update mode and bug fixes

    Add precision specifier for model dumping
    Add update mode to modify a model
    Lots of english corrections in the manual
    Fix bug in model format compatibility
    Fix memory allocation with large models
    Fix small memory bug in quark database
    Fix bug with bigram features in raw mode

23/04/2012
  Release v1.4.0: Forced decoding, optimizer state, and bug fixes

    Add forced decoding to partialy decode sequences
    Add optimizer state saving for L-BFGS and R-PROP
    Switched to elapsed time instead of wall time in progress
    Fix bug in Makefile (thanks to Lars Buitinck)
    Fix local normalization decoding in MEMM (thanks to Anoop Deoras)
    Fix bad handling of objwin option
    Fix bug in reader for single letter obs and lbl

03/11/2011
  Release v1.3.0: MEMM, faster gradient and bug fixes

    Added support for Maximum Entropy Markov Models.
    Added use of atomic operation in gradient computation
    Improved RProp numerical stability
    Fix bug with unseen features in raw mode (thanks to Jurgen Van Gael)
    Fix bug discarding some features in maxent mode (thanks to George Foster)
    Switched code to stdint, should resolve some issue with size_t
      on exotic systems.

07/07/2011
  Release v1.2.0: RProp improvements and bug fix

    Switch from splay-trees to critbit-tries
    Add RProp+ and RProp- variants of the RProp algorithm
    Add a new projection scheme for RProp with l1
    Make maxent work with sequences
    Add space matching in regexp
    Fix a few small bugs

19/03/2011
  Release v1.1.3: Some small improvements

    New option --jobsize for fine grained multi threading
    Improved SGD index construction : a lot faster
    Fix a small bug in sparse multi-threaded gradient

12/11/2010
  Release v1.1.2: Bug fix release

    Fix a small bug in L-BFGS/OWL-QN, should improve a bit
      convergence speed in some case.
    Fix a bug in multi-thread job system thanks to Alexander Fraser,
      should fix error rates and training speed on large dataset.
    Fix two small memory leaks.
    Some improvment in quark database handling.

24/09/2010
  Release v1.1.1: Mainly multi-threading improvements

    RPROP algorithm is now fully multi-threaded.
    Error rate estimation during training is now multi-threaded.
    Better jobs scheduling in multi-threaded gradient.
    Multi-threading code can be disabled (compilation on Windows should
      be simpler).
    Fixed bug in L1 optimization with RPROP (should improve stability).

08/09/2010
  Release v1.1.0: A few new features

    Added maxent mode.
    Added decoding through posteriors, this should improve accuracy
      at the price of computational time.
    Added the RPROP optimization algorithm.
    Added absolute indexing in patterns.
    Changed the scored output format as the posterior decoding
      provide normalized score at each position. The output is now
      compatible with CRF++.
      Some code cleanup.

29/07/2010
  Release v1.0.2: Mainly a bug fix version.

    Fixed some memory leaks, thanks to David Keeler
    Fixed argument processing to be more user friendly
    Fixed small bug in model compaction
    Added reading of raw files
    Spell corrections in man page

18/06/2010
  Release v1.0.0: Initial public version.