ShannonAI/glyce
A PyTorch-based toolkit that enhances Chinese character representation by incorporating visual glyph information into BERT models.
Not currently ranked — collecting fresh signals.
star history
Glyce is a Chinese character representation toolkit that leverages visual glyph information to improve NLP task performance. It extends BERT to incorporate character-level visual features and provides pre-trained Glyce-BERT models trained on Chinese text. The approach combines visual representations of Chinese characters with traditional embedding techniques to boost performance across sequence labeling, sentence classification, semantic role labeling, and dependency parsing tasks.
Frequently asked
- What is ShannonAI/glyce?
- A PyTorch-based toolkit that enhances Chinese character representation by incorporating visual glyph information into BERT models.
- Is glyce open source?
- Yes — ShannonAI/glyce is open source, released under the Apache-2.0 license.
- What language is glyce written in?
- ShannonAI/glyce is primarily written in Python.
- How popular is glyce?
- ShannonAI/glyce has 424 stars on GitHub.
- Where can I find glyce?
- ShannonAI/glyce is on GitHub at https://github.com/ShannonAI/glyce.