Age | Commit message (Expand) | Author |
2010-08-07 | test | graehl |
2010-08-07 | forgot to add | graehl |
2010-08-07 | apply fsa models (so far only by bottom up) in cdec | graehl |
2010-08-07 | propagation of feature name+debug from factory, return correct features array... | graehl |
2010-08-07 | usage for FSA features | graehl |
2010-08-07 | simple test decodes | graehl |
2010-08-07 | dynamic fsa ff, factory for fsa and ff shares code, factory moved to ff_facto... | graehl |
2010-08-04 | fix to compile with early versions of boost | redpony |
2010-08-04 | allow for wider set of NT strings | trevor.cohn |
2010-08-04 | alloca note | graehl@gmail.com |
2010-08-02 | inside_outside | graehl@gmail.com |
2010-08-02 | disable undefined token to string | 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 |
2010-07-31 | shorten_left comment | graehl |
2010-07-31 | define | graehl |
2010-07-28 | debugging print - still no idea on .05% difference scoring 3gram using phrases | graehl |
2010-07-28 | fsa lm phrase mystery remains, but bool fsa::simple_phrase_score indicates wh... | graehl |
2010-07-27 | disable fsa lm phrase due to slight (.05%) diff in scores - solve it later | graehl |
2010-07-27 | set fsa features() properly | graehl |
2010-07-27 | sample fsa ffs compile | graehl |
2010-07-27 | fsa feature templated Accum interface, phrase interface allows exceeding mark... | graehl |
2010-07-26 | kbest compile | graehl |
2010-07-26 | cpbin sentserver too (dep used by parallelize.pl | graehl |
2010-07-26 | fix name collision | redpony |
2010-07-26 | comment cleanup | graehl |
2010-07-26 | include contextBOW prob when shortening context | graehl |
2010-07-26 | shorten fsa lm state using contextID | graehl |
2010-07-26 | LanguageModelFsa works. TODO: sri context shortening? | graehl |
2010-07-26 | stateful ff_from_fsa works - fixed bug that never moved right state from righ... | graehl |
2010-07-25 | USE_INFO_EDGE 0 include | graehl |
2010-07-25 | USE_INFO_EDGE works for exh, cube | graehl |
2010-07-25 | being ensuring edge debug info is copied when apply_models - new hg.AddEdge m... | graehl |
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 |
2010-07-25 | method that might help phil compile his local stuff (trunk was fine) | graehl |
2010-07-25 | cleaned up kbest, new USE_INFO_EDGE 1 logs per edge, --show_derivation (needs... | graehl |
2010-07-25 | defer to subclass Scan1 | graehl |
2010-07-24 | compile | graehl |
2010-07-24 | FSA: simpler Scan1 ScanT1 methods, otherewise also expose edge to full Scan | graehl |
2010-07-24 | sparse_vector non-const [], consistently use default ctor, operator bool inst... | graehl |
2010-07-24 | minor - disable hash sparse_vector for now | graehl |
2010-07-23 | deterministic FF_FSA debug in constructor | graehl |
2010-07-23 | strip debug param from feature spec; debug info on from_fsa Init | graehl |
2010-07-23 | strip debug param from feature spec; debug info on from_fsa Init | graehl |
2010-07-23 | comment about unary (1 var) goal rule having no target terminals - seems to b... | graehl |
2010-07-23 | prettier | graehl |
2010-07-23 | pretty | graehl |
2010-07-23 | pretty | graehl |
2010-07-23 | more FF_FSA_DEBUG logs | graehl |
2010-07-23 | implicit first param to ffs: "debug". fsa final traversal set feature=0 firs... | graehl |