| | --- |
| | license: apache-2.0 |
| | task_categories: |
| | - image-classification |
| | - text-to-image |
| | - token-classification |
| | language: |
| | - zh |
| | tags: |
| | - chinese |
| | - ancient-script |
| | pretty_name: 'CHUBS: A Large-Scale Dataset of Chu Bamboo Slip Script' |
| | size_categories: |
| | - 100K<n<1M |
| | --- |
| | |
| | # CHUBS: A Large-Scale Dataset of Chu Bamboo Slip Script |
| |
|
| | <div align="center"> |
| | <a href="https://github.com/THUNLP/Chujian">Code</a> | <a href="">Paper (upcoming)</a> |
| | </div> |
| |
|
| | ## Introduction |
| |
|
| | This is a large-scale dataset of Chu bamboo slip (CBS, Chinese: 楚简, chujian) script, an ancient Chinese script used during the Spring and Autumn period over 2,000 years ago. This dataset consists of two parts: |
| |
|
| | 1. The main dataset where each example is an image and the corresponding text label. This part is contained in the `glyphs.zip` ZIP file. |
| | 2. A character detection dataset where each example is a bamboo slip and the corresponding bounding boxes of each CBS character. This part is contained in the `detection-data.zip` ZIP file. |
| | 3. A POS tagging dataset where each example is a sentence and POS annotations (using the BIO format). This part is contained under the `pos-tagging-data` directory. |
| |
|
| | This data is compiled to foster applications of NLP methods in the study of Chu bamboo slip script, and possibly other forms of ancient Chinese. |
| |
|
| | ## Data Source |
| |
|
| | The sources for this dataset are shown in the following list. |
| |
|
| | | **Source name** | **Chinese name** | **# documents** | **# slips** | **# characters** | |
| | |----------------------------|------------------|-----------------|-------------|------------------| |
| | | Tsinghua University Slips | 清华简 | 50 | 1,402 | 31,468 | |
| | | Shanghai Museum Slips | 上博简 | 60 | 881 | 25,795 | |
| | | Baoshan Slips | 包山简 | 4 | 337 | 12,647 | |
| | | Guodian Slips | 郭店简 | 18 | 705 | 11,865 | |
| | | Geling Slips | 葛陵简 | 8 | 743 | 6,209 | |
| | | Zenghouyi Slips | 曾侯乙简 | 4 | 198 | 6,016 | |
| | | Jiudian Slips | 九店简 | 2 | 232 | 2,956 | |
| | | Wangshan Slips | 望山简 | 3 | 273 | 2,218 | |
| | | Changtaiguan Slips | 长台关简 | 3 | 148 | 1,504 | |
| | | Zidanku Silk | 子弹库帛 | 7 | 7 | 1,471 | |
| | | Yangtianhu Slips | 仰天湖简 | 1 | 42 | 335 | |
| | | Wulipai Slips | 五里牌简 | 1 | 18 | 109 | |
| | | Xiyangpo Slips | 夕阳坡简 | 1 | 2 | 54 | |
| | | Ynagjiawan Slips | 杨家湾简 | 1 | 38 | 41 | |
| | | Caojiagang Slips | 曹家岗简 | 1 | 7 | 34 | |
| | | **Total** | | **164** | **5,033** | **102,722** | |
| |
|
| | ## Copyright |
| |
|
| | As mentioned in the paper, the main CHUBS dataset is extracted and processed from existing openly available textbooks and quality-checked by our team. Compiling and analyzing such data is a common practice within the paleography field in ancient Chinese script, and we have consulted with legal experts to make sure that this data can be released under a permissive license such as Apache 2.0. The POS tagging data is annotated by our team, and also has no legal concerns related to its usage. |
