Managing and Viewing Datasets

• QpiAI Pro provides an intuitive interface for managing and viewing your datasets.

💡Steps to access and manage datasets

• Navigate to the datasets to view all the created datasets.
• View key details such as dataset name, domain, subdomain, data type, creation date, and annotation status.
• Filter and sort datasets based on metadata for easy organization.
• Locate the dataset of interest and click on view details.
  • Once you click the “view details” you will get an intuitive dataset management interface as a side bar with following functionalities:
    • View: Visualize the dataset in an interactive data visualization window, along with functionality for embeddings-based EDA.
    • Dataset Health Check: Assess the quality and integrity of your dataset along with annotation stats and detailed insights.
    • Auto Annotate: Automatically generate annotations using advanced AI models for efficient labeling based on provided Text Prompts.
    • Few Shot Annotation: Automatically Annotate the datasets using minimal examples, leveraging pre-trained models for accelerated annotations.
    • Manual Annotation: Create or refine annotations manually to ensure accuracy & precision.
    • Save: Save the generated annotations.
    • Download the annotated dataset for external use, integration, or backup. The dataset can be exported in the following formats:
      • COCO
      • YOLO
      • PASCAL VOC
    • Dataset Views: Access and/or manage the saved views for better organization of annotation experiments and/or custom views created during EDA.

These features simplify dataset management, ensuring high-quality preparation for AI workflows. In the next chapter let us take a deep dive into the dataset visualization and EDA.

Note : For detailed instructions on downloading datasets in different formats, please refer to the Appendix. It provides a comprehensive explanation of all supported export formats and the step-by-step instructions.