spaCy/website/usage/_vectors-similarity/_gpu.jade

31 lines
1.2 KiB
Plaintext
Raw Normal View History

2017-10-03 15:26:20 +03:00
//- 💫 DOCS > USAGE > VECTORS & SIMILARITY > GPU
p
| If you're using a GPU, it's much more efficient to keep the word vectors
| on the device. You can do that by setting the
| #[+api("vectors#attributes") #[code Vectors.data]] attribute to a
| #[code cupy.ndarray] object if you're using spaCy
| or #[+a("https://chainer.org") Chainer], or a
| #[code torch.Tensor] object if you're using
| #[+a("http://pytorch.org") PyTorch]. The #[code data] object just needs
| to support #[code __iter__] and #[code __getitem__], so if you're using
| another library such as #[+a("https://www.tensorflow.org") TensorFlow],
| you could also create a wrapper for your vectors data.
+code("spaCy, Thinc or Chainer").
import cupy.cuda
from spacy.vectors import Vectors
vector_table = numpy.zeros((3, 300), dtype='f')
vectors = Vectors([u'dog', u'cat', u'orange'], vector_table)
with cupy.cuda.Device(0):
vectors.data = cupy.asarray(vectors.data)
+code("PyTorch").
import torch
from spacy.vectors import Vectors
vector_table = numpy.zeros((3, 300), dtype='f')
vectors = Vectors([u'dog', u'cat', u'orange'], vector_table)
vectors.data = torch.Tensor(vectors.data).cuda(0)