Customized Layout Detection for Scientific PDFs with LayoutParser and Label Studio
Автор: Label Studio
Загружено: 2022-02-09
Просмотров: 25039
LayoutParser is a deep learning-based tool that helps identify documents’ visual layouts and extract structured information from them. However, end users commonly encounter unique layouts that require data annotation and fine-tuning a customized model in order to achieve an ideal accuracy. In this webinar, we will show how to build one such model for scientific PDF extraction as an example. We will use label-studio for creating the training dataset, and we will fine-tune and use the layout detection model in LayoutParser.
As always, submit questions and participate in Q&A by joining our Label Studio Slack Community #webinars channel here: http://slack.labelstud.io.s3-website-...
More about LayoutParser:
https://github.com/Layout-Parser/layo...
More about Label Studio:
Open Source: https://labelstud.io
Enterprise: https://www.heartex.com
Newsletter: https://labelstudio.substack.com/
GitHub: https://github.com/heartexlabs/label-...
Доступные форматы для скачивания:
Скачать видео mp4
-
Информация по загрузке:
![Как общаться с PDF-файлами, используя локальные большие языковые модели [Ollama RAG]](https://imager.clipsaver.ru/ztBJqzBU5kc/max.jpg)
