mirror of
https://github.com/explosion/spaCy.git
synced 2024-12-28 10:56:31 +03:00
* whitespace
This commit is contained in:
parent
ed40a8380e
commit
65ac389191
|
@ -84,7 +84,7 @@ def _merge_sents(sents):
|
||||||
def train(Language, gold_tuples, model_dir, n_iter=15, feat_set=u'basic',
|
def train(Language, gold_tuples, model_dir, n_iter=15, feat_set=u'basic',
|
||||||
seed=0, gold_preproc=False, n_sents=0, corruption_level=0,
|
seed=0, gold_preproc=False, n_sents=0, corruption_level=0,
|
||||||
verbose=False,
|
verbose=False,
|
||||||
eta=0.01, mu=0.9, n_hidden=100,
|
eta=0.01, mu=0.9, nv_hidden=100,
|
||||||
nv_word=10, nv_tag=10, nv_label=10):
|
nv_word=10, nv_tag=10, nv_label=10):
|
||||||
dep_model_dir = path.join(model_dir, 'deps')
|
dep_model_dir = path.join(model_dir, 'deps')
|
||||||
pos_model_dir = path.join(model_dir, 'pos')
|
pos_model_dir = path.join(model_dir, 'pos')
|
||||||
|
@ -105,7 +105,7 @@ def train(Language, gold_tuples, model_dir, n_iter=15, feat_set=u'basic',
|
||||||
features=feat_set,
|
features=feat_set,
|
||||||
labels=Language.ParserTransitionSystem.get_labels(gold_tuples),
|
labels=Language.ParserTransitionSystem.get_labels(gold_tuples),
|
||||||
vector_lengths=(nv_word, nv_tag, nv_label),
|
vector_lengths=(nv_word, nv_tag, nv_label),
|
||||||
hidden_nodes=n_hidden,
|
hidden_nodes=nv_hidden,
|
||||||
eta=eta,
|
eta=eta,
|
||||||
mu=mu
|
mu=mu
|
||||||
)
|
)
|
||||||
|
@ -123,7 +123,7 @@ def train(Language, gold_tuples, model_dir, n_iter=15, feat_set=u'basic',
|
||||||
(nv_tag * len(tags)) + \
|
(nv_tag * len(tags)) + \
|
||||||
(nv_label * len(labels))
|
(nv_label * len(labels))
|
||||||
print 'Compiling'
|
print 'Compiling'
|
||||||
debug, train_func, predict_func = compile_theano_model(n_classes, n_hidden,
|
debug, train_func, predict_func = compile_theano_model(n_classes, nv_hidden,
|
||||||
n_in, 0.0, 0.0)
|
n_in, 0.0, 0.0)
|
||||||
print 'Done'
|
print 'Done'
|
||||||
return TheanoModel(
|
return TheanoModel(
|
||||||
|
@ -251,7 +251,7 @@ def main(train_loc, dev_loc, model_dir, n_sents=0, n_iter=15, out_loc="", verbos
|
||||||
gold_train = list(read_json_file(train_loc))
|
gold_train = list(read_json_file(train_loc))
|
||||||
nlp = train(English, gold_train, model_dir,
|
nlp = train(English, gold_train, model_dir,
|
||||||
feat_set='embed',
|
feat_set='embed',
|
||||||
nv_word=nv_word, nv_tag=nv_tag, nv_label=nv_label,
|
nv_word=nv_word, nv_tag=nv_tag, nv_label=nv_label, nv_hidden=nv_hidden,
|
||||||
gold_preproc=gold_preproc, n_sents=n_sents,
|
gold_preproc=gold_preproc, n_sents=n_sents,
|
||||||
corruption_level=corruption_level, n_iter=n_iter,
|
corruption_level=corruption_level, n_iter=n_iter,
|
||||||
verbose=verbose)
|
verbose=verbose)
|
||||||
|
|
Loading…
Reference in New Issue
Block a user