LayoutLM Healthcare NLP Dataset
Dataset combining medical data and NLP models (LLM, transformers LayoutLM) for the analysis and classification of semi-structured documents such as patient forms.
Approximately 11,100 files with 14 columns, JSON/CSV formats adapted to medical NLP and LayoutLM transformers
CC BY-SA 4.0
Description
The dataset LayoutLM Healthcare NLP includes data and models for a variety of medical NLP tasks, such as summary, Q&A, token labeling, and classification, with a focus on semi-structured documents like patient forms.
What is this dataset for?
- Train specialized health NLP models on semi-structured documents
- Perform summary, question-answer, and classification tasks on medical data
- Develop transformers adapted to the needs of the medical sector
Can it be enriched or improved?
Yes, by adding annotations specific to medical sub-fields or by integrating multilingual data to extend the reach. Adaptation to other semi-structured document formats is also possible.
🔎 In summary
🧠 Recommended for
- Medical NLP researchers
- Health AI developers
- Data scientists
🔧 Compatible tools
- Hugging Face Transformers
- LayoutLM
- PyTorch
- TensorFlow
💡 Tip
Take advantage of the semi-structured format to combine textual information and visual layout in your models.
Frequently Asked Questions
What is LayoutLM and why is it suitable for medical documents?
LayoutLM is a transformer template that integrates document layout to better understand semi-structured documents, very useful for medical forms.
Does this dataset include annotations for named entity recognition (NER)?
Yes, it includes annotations for token labeling, including NER, suitable for medical applications.
Can this dataset be used to train general LLMs?
It is primarily designed for medical tasks and semi-structured documents, but some elements can be transferred to larger models.
