Age | Commit message (Expand) | Author |
2014-04-25 | check for non-rescorable hypergraphs | Chris Dyer |
2014-04-08 | smarter union | Chris Dyer |
2014-04-07 | track node state for smarter union | Chris Dyer |
2012-07-28 | slight improvement to the way NTs are handled in the FST-CFG composition algo... | Chris Dyer |
2012-06-04 | Make decoder REALLY quiet when SILENT=1 | Victor Chahuneau |
2012-05-27 | clean up | Chris Dyer |
2011-10-11 | remove implicit conversion-to-double operator from LogVal<T> that caused over... | Guest_account Guest_account prguest11 |
2011-04-22 | make compatible with FastSparseVector | Chris Dyer |
2011-03-22 | fix local normalizer code for summary features | Chris Dyer |
2011-02-10 | conditional compilation of experimental code, remove prelm scoring code in pr... | Chris Dyer |
2010-08-09 | ComputeEdgeViterbi | graehl@gmail.com |
2010-08-09 | hg sort in_edges by viterbi, debug print temporarily enabled which always --s... | graehl@gmail.com |
2010-08-08 | cfg | graehl@gmail.com |
2010-08-02 | fake tdict names for non-ids, push viterbi cost to root in hg, store as featu... | graehl@gmail.com |
2010-07-25 | fixed CreateViterbiHypergraph (old impl did not work), so --show_derivation w... | graehl |
2010-07-25 | more comprehensible (but untested) edge/node filtering, awesome per-edge debu... | graehl |
2010-07-24 | FSA: simpler Scan1 ScanT1 methods, otherewise also expose edge to full Scan | graehl |
2010-07-23 | sparse_vector use google::dense_hash_map, fsa scan logging | graehl |
2010-07-16 | more scale-insensitive creep - will work even if LogVal is using float (1e-6 ... | graehl |
2010-07-12 | DEBUG_PROMISE - looks ok | graehl |
2010-07-12 | inebriated, but sped up inside/outside pruning and made cube poplimit scale w... | graehl |
2010-07-08 | --show_features | graehl |
2010-07-07 | safe hg pruning without needing additional inside reachability pass (max marg... | graehl |
2010-07-07 | get | graehl |
2010-07-06 | comment | graehl |
2010-07-06 | prune epsilon, comments, use inside pruning for probably no segfaults in vest | graehl |
2010-07-06 | hg density prune epsilon workaround comment - fundamental bug probably still ... | graehl |
2010-07-06 | epsilon distance for pruning - should alter density | graehl |
2010-07-06 | log density prune | graehl |
2010-07-06 | cdec --[prelm_]density_prune | graehl |
2010-07-05 | comment unused var names, todo | graehl |
2010-07-05 | TODO (mostly low priority) notes | graehl |
2010-07-01 | factor forest stats (show size post pruning, and portion kept) and cdec --sca... | graehl |
2010-06-25 | allow --beam_prune=0 | graehl |
2010-06-22 | initial checkin | redpony |