Publish time: 2022-03-07 00:00:00
Natural language processing plays a critical role in multiple AI applications. Based on transformer structure, we enhanced the capability of Traditional-Chinese language processing of various transformer models, including CKIP Albert LM model and CKIP Bert LM model for natural language understanding, and CKIP GPT2 LM model for natural language generation, etc. These models can be further trained for different downstream tasks. In this toolkit, besides the LM models mentioned above, we also provide the trained models for some popular downstream tasks, including word segmentation, part-of-speech tagging, and named entity recognition, which can recognize 18 types, including person, norp, facility, organization, gpe, location, product, event, work of art, law, language, date, time, percent, money, quantity, ordinal and cardinal. The toolkit is written in Python and with excellent performance, and its APIs are easy to call and integrate.
Welcome to test our online system: https://ckip.iis.sinica.edu.tw/service/transformers/.
- Based on transformer structure, we enhanced the capability of Traditional-Chinese language processing of various transformer models.
- Ckip Transformer for word segmentation significantly surpasses Jieba system.
- Rich POS types: 61 POS types in total (https://github.com/ckiplab/ckiptagger/wiki/POS-Tags)
- Rich NER types: 11 general entity types and 7 numeral types (https://github.com/ckiplab/ckiptagger/wiki/Entity-Types)
- Support user-defined word list
- Can be further trained for different downstream tasks.

Figure1 Performance Comparison
05T-1110222
Know-how
- Big-data Opinion Analysis
- Language Understanding
- Intelligent Agent
- Chatbot
- Commodity Intelligence System
Wei-Yun Ma, Mu Yang
If you are interested in this technology, please fill in the information on the webpage under the manufacturer option (Request for Technology Licensing, Material Transfer, or Sponsored Research) on the website of this office

Chinese-Enhanced Transformer Models for Natural Language Processing(CKIP Transformers)