Age | Commit message (Collapse) | Author | |
---|---|---|---|
2012-01-27 | rename vest to dpmert (dynamic programming mert), rename variables and types ↵ | Chris Dyer | |
to correspond to standard geometric concepts | |||
2012-01-27 | migration to new metric api for vest, clean up of unsupported/not functional ↵ | Chris Dyer | |
code | |||
2012-01-27 | migrate mert to the new scorer interface | Chris Dyer | |
2011-12-20 | migrate fast_score to the new API | Chris Dyer | |
2011-11-11 | new improved distributed operation for PRO, MERT | Chris Dyer | |
2011-11-10 | better defaults for pro | Chris Dyer | |
2011-10-20 | crazy bugs in parallelize | Guest_account Guest_account prguest11 | |
2011-09-13 | get rid of bad Weights class so it no longer keeps a copy of a vector inside it | Chris Dyer | |
2011-06-08 | rudimentary support for meteor via an external process. still needs ↵ | Chris Dyer | |
configuration of path, but it should work | |||
2011-05-16 | proper sampling from unit sphere | Chris Dyer | |
2011-04-22 | fix build | Chris Dyer | |
2011-04-22 | make compatible with FastSparseVector | Chris Dyer | |
2011-04-16 | pthread fix | Chris Dyer | |
2011-03-24 | Be more paranoid and check count of topbest sents, too. Also, provide path ↵ | Jonathan Clark | |
to decoder logs when we do fail. | |||
2011-03-20 | prevent over-aggressive error checking in vest script | Chris Dyer | |
2011-03-14 | more paranoid checking when (idiot/time-crunched) user tries to define his ↵ | Jonathan Clark | |
own <seg> tags during tuning | |||
2011-03-11 | be more verbose when running each child decoder process when forking. also, ↵ | Jonathan Clark | |
avoid some non-bash errors | |||
2011-03-11 | dont fail on possibly temporary qstat errors | Jonathan Clark | |
2011-03-11 | just use grep and dont check return code | Jonathan Clark | |
2011-03-11 | fail | Jonathan Clark | |
2011-03-11 | dont die when there are no running jobs | Jonathan Clark | |
2011-03-11 | another dumb bug involving cleanup being executed preemptively | Jonathan Clark | |
2011-03-11 | fix my dumb bug that killed qsub functionality | Jonathan Clark | |
2011-03-10 | Unsuck dist-vest | Jonathan Clark | |
2011-03-10 | initial version of scons | Jonathan Clark | |
2011-02-22 | fail early in vest on decoder failure | Chris Dyer | |
2011-02-11 | vest fix | Chris Dyer | |
2011-02-11 | vest fix | Chris Dyer | |
2011-02-11 | multi-pass mert | Chris Dyer | |
2011-02-06 | parallelize vest on large shared memory machines using fork and make | Chris Dyer | |
2010-12-10 | extract kbest alignments | Chris Dyer | |
2010-11-12 | Configuration for new scheduler on the LTI cluster | jon.h.clark | |
git-svn-id: https://ws10smt.googlecode.com/svn/trunk@714 ec762483-ff6d-05da-a07a-a48fb63a330f | |||
2010-08-26 | few fixes for stricter ld in gcc 4.5.1 | redpony | |
git-svn-id: https://ws10smt.googlecode.com/svn/trunk@624 ec762483-ff6d-05da-a07a-a48fb63a330f | |||
2010-08-17 | vest generate / map vector print / read compatability | graehl | |
git-svn-id: https://ws10smt.googlecode.com/svn/trunk@584 ec762483-ff6d-05da-a07a-a48fb63a330f | |||
2010-08-17 | fix bug when vest starts with an origin = 0 | redpony | |
git-svn-id: https://ws10smt.googlecode.com/svn/trunk@578 ec762483-ff6d-05da-a07a-a48fb63a330f | |||
2010-08-12 | add 'make check' support | redpony | |
git-svn-id: https://ws10smt.googlecode.com/svn/trunk@534 ec762483-ff6d-05da-a07a-a48fb63a330f | |||
2010-08-11 | more link lne fixes | redpony | |
git-svn-id: https://ws10smt.googlecode.com/svn/trunk@521 ec762483-ff6d-05da-a07a-a48fb63a330f | |||
2010-08-11 | forgotten files | redpony | |
git-svn-id: https://ws10smt.googlecode.com/svn/trunk@520 ec762483-ff6d-05da-a07a-a48fb63a330f | |||
2010-08-11 | major refactor, break bad circular deps | redpony | |
git-svn-id: https://ws10smt.googlecode.com/svn/trunk@509 ec762483-ff6d-05da-a07a-a48fb63a330f | |||
2010-08-10 | support for running in multiple environments which are automatically detected | redpony | |
git-svn-id: https://ws10smt.googlecode.com/svn/trunk@501 ec762483-ff6d-05da-a07a-a48fb63a330f | |||
2010-08-07 | syntax | graehl@gmail.com | |
git-svn-id: https://ws10smt.googlecode.com/svn/trunk@492 ec762483-ff6d-05da-a07a-a48fb63a330f | |||
2010-08-07 | init bleumodel in generate | graehl@gmail.com | |
git-svn-id: https://ws10smt.googlecode.com/svn/trunk@491 ec762483-ff6d-05da-a07a-a48fb63a330f | |||
2010-08-07 | propagation of feature name+debug from factory, return correct features ↵ | graehl | |
array for fsa ffs git-svn-id: https://ws10smt.googlecode.com/svn/trunk@486 ec762483-ff6d-05da-a07a-a48fb63a330f | |||
2010-07-31 | fixed dist-vest bug in passing ref args after the first - explains the phil ↵ | graehl | |
complaint and my bad oracle bleu tuning git-svn-id: https://ws10smt.googlecode.com/svn/trunk@463 ec762483-ff6d-05da-a07a-a48fb63a330f | |||
2010-07-28 | vest: pass refs to generate when using oracle; expect gzipped decoder output | graehl | |
git-svn-id: https://ws10smt.googlecode.com/svn/trunk@450 ec762483-ff6d-05da-a07a-a48fb63a330f | |||
2010-07-28 | vest bleu_weight | graehl | |
git-svn-id: https://ws10smt.googlecode.com/svn/trunk@448 ec762483-ff6d-05da-a07a-a48fb63a330f | |||
2010-07-28 | minor | graehl | |
git-svn-id: https://ws10smt.googlecode.com/svn/trunk@447 ec762483-ff6d-05da-a07a-a48fb63a330f | |||
2010-07-27 | bindir | graehl | |
git-svn-id: https://ws10smt.googlecode.com/svn/trunk@440 ec762483-ff6d-05da-a07a-a48fb63a330f | |||
2010-07-27 | vest cpbin include sentclient | graehl | |
git-svn-id: https://ws10smt.googlecode.com/svn/trunk@439 ec762483-ff6d-05da-a07a-a48fb63a330f | |||
2010-07-26 | cpbin sentserver too (dep used by parallelize.pl | graehl | |
git-svn-id: https://ws10smt.googlecode.com/svn/trunk@425 ec762483-ff6d-05da-a07a-a48fb63a330f |