Create README.md

README.md

@@ -0,0 +1,119 @@
+---
+task_categories:
+- text-retrieval
+- sentence-similarity
+language:
+- en
+tags:
+- embeddings
+- vector-database
+- benchmark
+---
+# GAS ANN Centroids
+
+## Dataset Description
+
+This dataset contains pre-computed centroids of each embedding models our GAS (Geometry-Aware Selection) algorithm, designed for performing ANN (Approximate Nearest Neighbor) benchmark.
+
+### Purpose
+
+Benchmark centroids dataset for evaluating vector database performance, specifically designed for use with [VectorDBBench](https://github.com/zilliztech/VectorDBBench).
+
+### Dataset Summary
+
+- **Supported Embedding Model**
+  - [google/embeddinggemma-300m](https://huggingface.co/google/embeddinggemma-300m)
+
+## Dataset Structure
+
+Each embedding model directory has two data:
+
+| Data | Description |
+|-------|-------------|
+| `centroids.npy` | centroids as followed IVF |
+| `tree_info.pkl` | tree metadata for GAS centroids |
+
+## Data Fields
+
+## Dataset Creation
+
+### Source Data
+
+Source dataset is a large open-domain, Wikipedia dataset: [mixedbread-ai/wikipedia-data-en-2023-11](https://huggingface.co/datasets/mixedbread-ai/wikipedia-data-en-2023-11).
+
+### Preprocessing
+
+1. Create Centroids by GAS approach:
+
+   Description TBD
+   
+3. Normalization: All embeddings are L2-normalized
+
+### Embedding Generation
+
+- Model: google/embeddinggemma-300m
+- Dimension: 768
+- Max Token Length: 2048
+- Normalization: L2-normalized
+
+## Usage
+
+```python
+import wget
+
+def download_centroids(embedding_model: str, dataset_dir: str) -> None:
+    """Download pre-computed centroids and tree info for GAS."""
+    
+    dataset_link = "https://huggingface.co/datasets/cryptolab-playground/gas-centroids/resolve/main/embeddinggemma-300m"
+    
+    wget.download(f"{dataset_link}/centroids.npy", out="centroids.npy")
+    wget.download(f"{dataset_link}/tree_info.pkl", out="tree_info.pkl")
+```
+
+
+## License
+
+Apache 2.0
+
+## Citation
+
+If you use this dataset, please cite:
+
+```bibtex
+@dataset{gas-centroids,
+  author = {CryptoLab, Inc.},
+  title = {GAS Centroids},
+  year = {2025},
+  publisher = {Hugging Face},
+  url = {https://huggingface.co/datasets/cryptolab-playground/gas-centroids}
+}
+```
+
+### Source Dataset Citation
+
+```bibtex
+@dataset{wikipedia_data_en_2023_11,
+  author = {mixedbread-ai},
+  title = {Wikipedia Data EN 2023 11},
+  year = {2023},
+  publisher = {Hugging Face},
+  url = {https://huggingface.co/datasets/mixedbread-ai/wikipedia-data-en-2023-11}
+}
+```
+
+### Embedding Model Citation
+
+```bibtex
+@misc{embeddinggemma,
+  title={Embedding Gemma},
+  author={Google},
+  year={2024},
+  url={https://huggingface.co/google/embeddinggemma-300m}
+}
+```
+
+### Acknowledgments
+
+- Original dataset: mixedbread-ai/wikipedia-data-en-2023-11
+- Embedding model: google/embeddinggemma-300m
+- Benchmark framework: VectorDBBench