Parse and extract text from scholarly documents
Search information in uploaded PDFs
Gemma-3 OCR App
Search documents and retrieve relevant chunks
Extract named entities from text
Extract text from multilingual invoices
Extract key entities from text queries
Find relevant text chunks from documents based on a query
Find relevant text chunks from documents based on queries
Extract text and summarize from documents
OCR that extract text from image of hindi and english
Analyze legal PDFs and answer questions
Search for similar text in documents
Grobid End to End Evaluation is a tool designed to assess and validate the performance of Grobid, a machine learning model used for parsing and extracting text from scholarly documents. It evaluates the accuracy and reliability of Grobid's output by comparing it against ground truth data, ensuring the extracted text meets high standards of quality and usability.
What file formats does Grobid End to End Evaluation support?
Grobid End to End Evaluation supports PDF, XML, and plain text formats for both input documents and ground truth files.
Can I customize the evaluation metrics?
Yes, Grobid End to End Evaluation allows users to define custom metrics and weighting to suit specific requirements.
How do I handle documents with complex layouts?
Grobid is specifically designed to handle complex layouts, including multi-column text, tables, and figures. Ensure your ground truth accurately reflects these elements for proper evaluation.