explosion/spaCy
1
1
Fork 0
mirror of https://github.com/explosion/spaCy.git synced 2025-02-11 00:50:33 +03:00
Code Issues Packages Projects Releases Wiki Activity

* Use the gold sense labels for training

Browse Source
This commit is contained in:
Matthew Honnibal 2015-07-03 05:45:42 +02:00
parent 4a60b68a24
commit 5d933eec8e
1 changed files with 7 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

9
spacy/sense_tagger.pyx
View File

@ -207,11 +207,16 @@ cdef class SenseTagger:
eg = Example(self.model.n_classes, CONTEXT_SIZE, self.model.n_feats+1, eg = Example(self.model.n_classes, CONTEXT_SIZE, self.model.n_feats+1,
self.model.n_feats+1) self.model.n_feats+1)
cdef int i cdef int i
for i, ssenses in enumerate(gold.ssenses):
if ssenses:
gold.c.ssenses[i] = encode_sense_strs(ssenses)
else:
gold.c.ssenses[i] = pos_senses(&tokens.data[i])
cdef int cost = 0 cdef int cost = 0
for i in range(tokens.length): for i in range(tokens.length):
if tokens.data[i].lex.senses == 0: if tokens.data[i].lex.senses == 0 or tokens.data[i].lex.senses == 1:
continue continue
self._set_costs(<bint*>eg.c.is_valid, eg.c.costs, pos_senses(&tokens.data[i])) self._set_costs(<bint*>eg.c.is_valid, eg.c.costs, gold.c.ssenses[i])
fill_context(eg.c.atoms, &tokens.data[i]) fill_context(eg.c.atoms, &tokens.data[i])
self.model.train(eg) self.model.train(eg)