explosion/spaCy
1
1
Fork 0
mirror of https://github.com/explosion/spaCy.git synced 2025-10-29 23:17:59 +03:00
Code Issues Packages Projects Releases Wiki Activity

Stream the gold data during training, to reduce memory

Browse Source
This commit is contained in:
Matthew Honnibal 2018-03-10 22:32:32 +01:00
parent 7c0823ad76
commit c3d168509a
1 changed files with 3 additions and 4 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

7
spacy/gold.pyx
View File

@ -13,7 +13,7 @@ from . import _align
from .syntax import nonproj from .syntax import nonproj
from .tokens import Doc from .tokens import Doc
from . import util from . import util
from .util import minibatch from .util import minibatch, itershuffle
def tags_to_entities(tags): def tags_to_entities(tags):
@ -133,15 +133,14 @@ class GoldCorpus(object):
def train_docs(self, nlp, gold_preproc=False, def train_docs(self, nlp, gold_preproc=False,
projectivize=False, max_length=None, projectivize=False, max_length=None,
noise_level=0.0): noise_level=0.0):
train_tuples = list(self.train_tuples)
if projectivize: if projectivize:
train_tuples = nonproj.preprocess_training_data( train_tuples = nonproj.preprocess_training_data(
self.train_tuples, label_freq_cutoff=30) self.train_tuples, label_freq_cutoff=30)
random.shuffle(train_tuples) random.shuffle(self.train_locs)
gold_docs = self.iter_gold_docs(nlp, train_tuples, gold_preproc, gold_docs = self.iter_gold_docs(nlp, train_tuples, gold_preproc,
max_length=max_length, max_length=max_length,
noise_level=noise_level) noise_level=noise_level)
yield from gold_docs yield from itershuffle(gold_docs, bufsize=100)
def dev_docs(self, nlp, gold_preproc=False): def dev_docs(self, nlp, gold_preproc=False):
gold_docs = self.iter_gold_docs(nlp, self.dev_tuples, gold_preproc) gold_docs = self.iter_gold_docs(nlp, self.dev_tuples, gold_preproc)