summaryrefslogtreecommitdiff
path: root/training
AgeCommit message (Collapse)Author
2012-11-20fixes for 2011 optimizerChris Dyer
2012-11-19merge upstream/masterPatrick Simianer
2012-11-19merge upstream/masterPatrick Simianer
2012-11-18forgotten makefileChris Dyer
2012-11-18major restructure of the training codeChris Dyer
2012-10-22Update search, make it compileKenneth Heafield
2012-10-16c++11 compatibilityChris Dyer
2012-10-11Merge branch 'master' of https://github.com/redpony/cdecChris Dyer
2012-10-11produce alignments for a test set (which can be stdin)Chris Dyer
2012-10-02cdec cleanup, remove bayesian stuff, parsing stuffChris Dyer
2012-09-20rename model1 to fast_aligner, since it does way more than model1 but is ↵Chris Dyer
mostly just fast
2012-08-13another fixChris Dyer
2012-07-29fix grammar converter to remove edges that cannot exist in any valid derivationChris Dyer
2012-07-02[python] Explicit candidate set; memory leak fixesVictor Chahuneau
2012-06-29add option for entropy optimizationChris Dyer
2012-06-24minimum risk training, not completely ready for primetimeChris Dyer
2012-06-24compute risk / gradient of riskChris Dyer
2012-06-22single processor fixChris Dyer
2012-06-19compute held-out ppl in mpi_batch_optimizeChris Dyer
2012-06-18add non-const iterators to sparse vector, speed up model1 codeChris Dyer
2012-06-01fix potential build breakChris Dyer
2012-05-29add support to rampion for accumulating k-best listsChris Dyer
2012-05-28cache metric computation in proChris Dyer
2012-05-27fix mapper to use common candidate set codeChris Dyer
2012-05-27kbest repoChris Dyer
2012-05-20Rename decoder library, assimilate mira directoryKenneth Heafield
2012-05-17Training jamfileChris Dyer
2012-05-17jamfiles for cregChris Dyer
2012-05-13put creg in its own top-level folderChris Dyer
2012-05-13fast creg training code for univariate linear and logistic regressionChris Dyer
2012-05-08better loggingChris Dyer
2012-05-08switch to liblbfgs for proChris Dyer
2012-05-08add liblbfgs, which is much less crappy than the current oneChris Dyer
2012-03-20make c++11 compatibleChris Dyer
2012-03-07more mpi fixesChris Dyer
2012-03-07better logging with MPIChris Dyer
2012-02-29mpi fixesChris Dyer
2012-02-29cleanup, mpi-ify lblmodelChris Dyer
2012-02-24load embeddings from fileChris Dyer
2012-02-22add regularizationChris Dyer
2012-02-21use lbfgsChris Dyer
2012-02-21basic lbl model, nothing to see hereChris Dyer
2012-02-19lbl preliminary clean upChris Dyer
2012-02-18initial lbl_model stubChris Dyer
2012-02-18fix diagonal modelChris Dyer
2012-02-17boost version checking, check for Eigen, get rid of old digamma stuffChris Dyer
2012-02-10move atools to utils directoryChris Dyer
2012-02-10clean up alignment toolsChris Dyer
2012-01-24more modelsChris Dyer
2012-01-20'pseudo model 2' that strictly favors a diagonal, with tunable parameters ↵Chris Dyer
for p(null) and how sharp/flat the alignment distribution is around the diagonal