summaryrefslogtreecommitdiff
AgeCommit message (Collapse)Author
2010-02-18add generative word alignment model and primitive EM trainer. Model 1 and ↵Chris Dyer
HMM are supported, without NULL source words
2010-02-01word aligner cleanup, new featuresChris Dyer
2010-01-29word aligner checkinChris Dyer
2010-01-27add binary model2-like featuresChris Dyer
2010-01-25fix minChris Dyer
2010-01-25fix constChris Dyer
2010-01-24simplest possible build instructionsChris Dyer
2010-01-24more autoconf fixes- use version of boost m4 macros which are much, much, ↵Chris Dyer
much better
2010-01-24Support building without gtestChris Dyer
Now the only dependence is boost, which most modern linux distros have.
2010-01-18add alignment visualization toolChris Dyer
2009-12-26add inside algorithm pass in pruning (prevents parent nodes that are ↵Chris Dyer
underivable from hanging around)
2009-12-26minor fix for fast intersection codeChris Dyer
2009-12-26increase intersection speed by a couple orders of magnitude for linear chain ↵Chris Dyer
graphs
2009-12-19few fixesChris Dyer
2009-12-19cool new alignment featureChris Dyer
2009-12-19fix rounding error failureChris Dyer
2009-12-18add symmetrization heuristics to atools, add null word configurationChris Dyer
2009-12-18fix test failures due to changing feature orderChris Dyer
2009-12-18add support for freezing the feature set to a user-specified list, even if ↵Chris Dyer
feature detectors create additional features
2009-12-17make node order explicit in graphChris Dyer
2009-12-17added non-pruning intersection and a CRF taggerChris Dyer
- the linear-chain tagger is more of a proof of concept than a real tagger-- the context-free assumptions made in a number of places mean that the algorithms used may not be as efficient as they could be, but the model is as powerful as any CRF - it would be easy to add latent variables or semi-CRF support (or both!) - i've added a couple basic features that are often used for POS tagging - non-pruning intersection is useful for lexical word alignment models and the tagger - a sample POS tagger model will be committed later
2009-12-15fix typoChris Dyer
2009-12-14few small fixes of alignment tools, add new orthographic similarity feature ↵Chris Dyer
for word aligner, final naming of directories, libraries in cdec
2009-12-08LICENSE fixes, full support of lattice decodingChris Dyer
2009-12-07minor fixChris Dyer
2009-12-07add compound splitterChris Dyer
2009-12-07add support for generating pruned lattices when in compound splitting modeChris Dyer
2009-12-07minor changes to the way the phonotactic prob is calculatedChris Dyer
2009-12-07fix pathChris Dyer
2009-12-06minor fixes, better loggingChris Dyer
2009-12-06add compound splitting logic and features (Dyer 2008, NAACL)Chris Dyer
2009-12-06forgot one file, still getting used to gitChris Dyer
2009-12-06add missing test data for unit testsChris Dyer
2009-12-05Update build instructionsAdam Lopez
2009-12-05Hacked configure to avoid gtest dependence on libtoolAdam Lopez
2009-12-04oops, forgot to add testsChris Dyer
2009-12-04git ignore addedChris Dyer
2009-12-04git ignores various auto-generated filesAdam Lopez
2009-12-04Should be able to install with autoreconf -i; configure [with options]; makeAdam Lopez
2009-12-04rename a few filesChris Dyer
2009-12-03fix descChris Dyer
2009-12-03small fixes, learning to use git...Chris Dyer
2009-12-03minimal docsChris Dyer
2009-12-03minimal docsChris Dyer
2009-12-03fix build in new layoutChris Dyer
2009-12-03initial check inChris Dyer