https://github.com/rikeda71/torchcrf

An Inplementation of CRF (Conditional Random Fields) in PyTorch 1.0
https://github.com/rikeda71/torchcrf

conditional-random-fields crf named-entity-recognition ner pytorch

Last synced: 5 months ago
JSON representation

An Inplementation of CRF (Conditional Random Fields) in PyTorch 1.0

• Host: GitHub
• URL: https://github.com/rikeda71/torchcrf
• Owner: rikeda71
• License: mit
• Created: 2018-11-07T08:03:53.000Z (over 6 years ago)
• Default Branch: master
• Last Pushed: 2020-08-01T09:26:36.000Z (over 4 years ago)
• Last Synced: 2024-12-01T14:46:31.786Z (5 months ago)
• Topics: conditional-random-fields, crf, named-entity-recognition, ner, pytorch
• Language: Python
• Homepage:
• Size: 63.5 KB
• Stars: 136
• Watchers: 4
• Forks: 11
• Open Issues: 4
• Metadata Files:
  • Readme: README.ja.md
  • License: LICENSE

Awesome Lists containing this project

README

        
# Torch CRF

[![CircleCI](https://circleci.com/gh/s14t284/TorchCRF.svg?style=svg)](https://circleci.com/gh/s14t284/TorchCRF)

[![Coverage Status](https://coveralls.io/repos/github/s14t284/TorchCRF/badge.svg)](https://coveralls.io/github/s14t284/TorchCRF)

[![MIT License](https://img.shields.io/github/license/s14t284/TorchCRF)](LICENSE)

[![Python Versions](https://img.shields.io/pypi/pyversions/TorchCRF.svg)](https://pypi.org/project/TorchCRF/)

[![PyPI version](https://badge.fury.io/py/TorchCRF.svg)](https://badge.fury.io/py/TorchCRF)

PyTorch 1.0 による条件付き確率場 (CRF) の実装

## Requirements

- python3 (>=3.6)

- PyTorch 1.0

## Installation

    $ pip install TorchCRF

## Usage

```python

>>> 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)

```

### 推論

```python

>>> crf.forward(hidden, labels, mask)

tensor([-7.6204, -3.6124], grad_fn=)

```

### 系列ラベルの予測

```python

>>> crf.viterbi_decode(hidden, mask)

[[0, 2, 2], [4, 0]]

```

## License

MIT

## References

- [threelittlemonkeys/lstm-crf-pytorch](https://github.com/threelittlemonkeys/lstm-crf-pytorch)

- [kmkurn/pytorch-crf](https://github.com/kmkurn/pytorch-crf)