| language: | |
| - en | |
| pretty_name: "quickb-kb" | |
| tags: | |
| - quickb | |
| - text-chunking | |
| - n<1K | |
| task_categories: | |
| - text-generation | |
| - text-retrieval | |
| task_ids: | |
| - document-retrieval | |
| library_name: quickb | |
| # quickb-kb | |
| Generated using [QuicKB](https://github.com/AdamLucek/quickb), a tool developed by [Adam Lucek](https://huggingface.co/AdamLucek). | |
| QuicKB optimizes document retrieval by creating fine-tuned knowledge bases through an end-to-end pipeline that handles document chunking, training data generation, and embedding model optimization. | |
| ### Chunking Configuration | |
| - **Chunker**: RecursiveTokenChunker | |
| - **Parameters**: | |
| - **chunk_size**: `400` | |
| - **chunk_overlap**: `0` | |
| - **length_type**: `'character'` | |
| - **separators**: `['\n\n', '\n', '.', '?', '!', ' ', '']` | |
| - **keep_separator**: `True` | |
| - **is_separator_regex**: `False` | |
| ### Dataset Statistics | |
| - Total chunks: 526 | |
| - Average chunk size: 57.8 words | |
| - Source files: 1 | |
| ### Dataset Structure | |
| This dataset contains the following fields: | |
| - `text`: The content of each text chunk | |
| - `source`: The source file path for the chunk | |
| - `id`: Unique identifier for each chunk |