An Implementation of Conditional Random Fields in pytorch
Project description
Torch CRF
Implementation of CRF (Conditional Random Fields) in PyTorch 1.0
Requirements
- python3 (>=3.6)
- PyTorch 1.0
Installation
$ pip install TorchCRF
Usage
>>> import torch
>>> from TorchCRF import CRF
>>> batch_size = 2
>>> sequence_size = 3
>>> num_labels = 5
>>> mask = torch.FloatTensor([[1, 1, 1], [1, 1, 0]]) # (batch_size. sequence_size)
>>> labels = torch.LongTensor([[0, 2, 3], [1, 4, 1]]) # (batch_size, sequence_size)
>>> hidden = torch.randn((batch_size, sequence_size, num_labels), requires_grad=True)
>>> crf = CRF(num_labels)
Computing log-likelihood (used where forward)
>>> crf.forward(hidden, labels, mask)
tensor([-7.6204, -3.6124], grad_fn=<ThSubBackward>)
Decoding (predict labels of sequences)
>>> crf.viterbi_decode(hidden, mask)
[[0, 2, 2], [4, 0]]
License
MIT
References
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
TorchCRF-1.0.4.tar.gz
(5.8 kB
view hashes)