README.md

metadata

tags:
  - giskard
  - knowledge-base
  - information-retrieval
task_categories:
  - text-generation
  - text2text-generation
  - question-answering
  - text-retrieval

Dataset Card for GTimothee/my-knowledge-base

This repository was created using the giskard library, an open-source Python framework designed to evaluate and test AI systems.

This dataset comprises a giskard's KnowledgeBase containing 310 documents. If embeddings were generated before the saving process, they are included and will be automatically loaded into a vector store when required.

Usage

You can load this knowledge base using the following code:

from giskard.rag import KnowledgeBase
kb = KnowledgeBase.load_from_hf_hub("GTimothee/my-knowledge-base")

Configuration

The configuration details for this Knowledge Base (can also be found in the config.json file):

{
    "columns": null,
    "chunk_size": 2048,
    "min_topic_size": 8,
    "language": "en",
    "seed": null,
    "embedding_model": null
}

---

Built with

Giskard helps identify performance, bias, and security issues in AI applications, supporting both LLM-based systems like RAG agents and traditional machine learning models for tabular data.