Skip to content

Milestones

List view

  • All issues related to improving user experience

    No due date
    3/4 issues closed
  • Based on the LightRAG framework, we will develop an intelligent retrieval-augmented generation system that constructs a multi-dimensional indexing architecture integrating semantic and keyword-based approaches. During queries, the system will employ a hybrid retrieval strategy to precisely locate relevant knowledge and collaborate with large language models to generate trustworthy responses with source citations.

    No due date
    1/1 issues closed
  • We will utilize PaddleOCR to build a high-performance document parsing engine, focusing on text information extraction from unstructured data such as scanned documents and images. The system will enable automated processing of various document types (including PDFs, images, tables, etc.), performing text detection, recognition, and layout analysis through deep learning models to output structured text data.

    No due date