mirror of
https://github.com/explosion/spaCy.git
synced 2025-02-11 17:10:36 +03:00
Fix Corpus
This commit is contained in:
parent
450c6fe39c
commit
396dd60b3a
|
@ -68,9 +68,11 @@ class Corpus:
|
||||||
i += 1
|
i += 1
|
||||||
return n
|
return n
|
||||||
|
|
||||||
def train_dataset(self, nlp, **kwargs):
|
def train_dataset(self, nlp, shuffle=True, **kwargs):
|
||||||
ref_docs = self.read_docbin(nlp.vocab, self.walk_corpus(self.train_loc))
|
ref_docs = self.read_docbin(nlp.vocab, self.walk_corpus(self.train_loc))
|
||||||
examples = list(self.make_examples(nlp, ref_docs, **kwargs))
|
examples = self.make_examples(nlp, ref_docs, **kwargs)
|
||||||
|
if shuffle:
|
||||||
|
examples = list(examples)
|
||||||
random.shuffle(examples)
|
random.shuffle(examples)
|
||||||
yield from examples
|
yield from examples
|
||||||
|
|
||||||
|
|
Loading…
Reference in New Issue
Block a user