Upload README.md with huggingface_hub

README.md

@@ -1,3 +1,23 @@
+---
+license: cc-by-4.0
+task_categories:
+- image-to-text
+language:
+- en
+pretty_name: NatureCommsCorpus
+size_categories:
+- 100G<n<1000G
+tags:
+- nature
+- nature communications
+- nature sciences
+- multi-discipline
+- multi-modal
+- science understanding
+- large vision language models
+- multimodal large language models
+---
+
 ## Dataset Structure
 
 The dataset is stored in the `rawdata` directory, organized into multiple levels for ease of access and categorization. Below is a detailed explanation of the directory structure:
@@ -80,3 +100,19 @@ Below is an example of the structure of a metadata JSON file:
 
 The dataset includes various metadata for scientific articles, including abstracts, figure captions, and references, which can be used for tasks such as article summarization, figure captioning, or reference extraction. The directory structure allows researchers to access data at different levels, depending on the task.
 
+## Additional Information
+
+### Licensing Information
+
+This corpus is released under [CC BY-NC-SA 4.0](http://creativecommons.org/licenses/by-nc-sa/4.0/).
+
+### Citation Information
+
+```
+@article{li2024mmsci,
+  title={Mmsci: A multimodal multi-discipline dataset for phd-level scientific comprehension},
+  author={Li, Zekun and Yang, Xianjun and Choi, Kyuri and Zhu, Wanrong and Hsieh, Ryan and Kim, HyeonJung and Lim, Jin Hyuk and Ji, Sungyoung and Lee, Byungju and Yan, Xifeng and others},
+  journal={arXiv preprint arXiv:2407.04903},
+  year={2024}
+}
+```