jiesutd/LatticeLSTM
A PyTorch implementation of Lattice LSTM for Chinese Named Entity Recognition, achieving state-of-the-art 93.18% F1 on MSRA dataset.
Not currently ranked — collecting fresh signals.
star history
This repository implements a Lattice LSTM model for Chinese NER, combining character-level LSTM with word (lattice) embeddings as input. The model uses bidirectional LSTM layers followed by a CRF layer for sequence labeling, taking advantage of both character sequences and pre-defined word boundaries in Chinese text. It supports pretrained character and word embeddings for improved performance.
Frequently asked
- What is jiesutd/LatticeLSTM?
- A PyTorch implementation of Lattice LSTM for Chinese Named Entity Recognition, achieving state-of-the-art 93.18% F1 on MSRA dataset.
- Is LatticeLSTM open source?
- Yes — jiesutd/LatticeLSTM is an open-source project tracked on heatdrop.
- What language is LatticeLSTM written in?
- jiesutd/LatticeLSTM is primarily written in Python.
- How popular is LatticeLSTM?
- jiesutd/LatticeLSTM has 1.8k stars on GitHub.
- Where can I find LatticeLSTM?
- jiesutd/LatticeLSTM is on GitHub at https://github.com/jiesutd/LatticeLSTM.