| --- |
| language: |
| - th |
| license: other |
| task_categories: |
| - text-generation |
| arxiv: 2512.18834 |
| configs: |
| - config_name: minhash_deduped |
| data_files: |
| - split: train |
| path: minhash_deduped/**/*.parquet |
| - config_name: quality_filtered |
| data_files: |
| - split: train |
| path: quality_filtered/**/*.parquet |
| - config_name: matched |
| data_files: |
| - split: train |
| path: consensus/*.parquet |
| default: minhash_deduped |
| --- |
| |
| <p align="center"> |
| <a href="https://huggingface.co/collections/AdaMLLab/mixminmatch"> |
| <img src="https://img.shields.io/badge/🤗_Collection-MixMinMatch-blue" alt="MixMinMatch Collection"> |
| </a> |
| </p> |
| |
| ThaMix ([https://arxiv.org/abs/2512.18834](https://arxiv.org/abs/2512.18834)) is a Thai pretraining corpus built by combining seven publicly available Thai datasets, applying Thai-specific quality filtering, and performing cross-dataset deduplication. |
|
|
| ## Subsets |
|
|
| | Subset | Description | |
| |--------|-------------| |
| | `quality_filtered` | Quality-filtered data before deduplication | |
| | `minhash_deduped` | Document-level MinHash deduplication | |
| | `matched` | Documents appearing in 2+ source datasets | |
|
|
| The matched subset uses cross-dataset agreement as a signal for quality. |
|
|
| ## Usage |
|
|
| ```python |
| from datasets import load_dataset |
| |
| ds = load_dataset("AdaMLLab/ThaMix", "minhash_deduped") |
| ds = load_dataset("AdaMLLab/ThaMix", "quality_filtered") |
| ds = load_dataset("AdaMLLab/ThaMix", "matched") |
| ``` |
|
|
| ## Sources |
|
|
| - FineWeb-2 (`HuggingFaceFW/fineweb-2`, `tha_Thai`) |
| - HPLT 2.0 (`HPLT/HPLT2.0_cleaned`, `tha_Thai`) |
| - CulturaX (`uonlp/CulturaX`, `th`) |
| - C4 (`allenai/c4`, `th`) |
| - SEA CommonCrawl (`sailor2/sea-commoncrawl`, `thai`) |
| - FinePDFs (`HuggingFaceFW/finepdfs`, `tha_Thai`) |
| - SEA-LION Pile v2 (`aisingapore/SEA-PILE-v2`, `th`) |
|
|
| ## Pipeline |
|
|
| 1. Quality filtering with Thai-specific thresholds (Thai script ratio, repetition patterns, line quality) |
| 2. Document-level MinHash deduplication (5-gram shingles, 14 bands, 8 hashes per band, similarity threshold 0.8) |
| 3. Cross-source matching to identify documents appearing in 2+ independent sources |
|
|
| ## Citation |
|
|
| ```bib |
| @misc{alrashed2025mixminmatch, |
| title={Mix, MinHash, and Match: Cross-Source Agreement for Multilingual Pretraining Datasets}, |
| author={Sultan Alrashed and Francesco Orabona}, |
| year={2025}, |
| eprint={2512.18834v2}, |
| archivePrefix={arXiv}, |
| primaryClass={cs.CL}, |
| url={https://arxiv.org/abs/2512.18834v2}, |
| } |
| ``` |
|
|
| ## License |
|
|
| See individual source dataset licenses. |
|
|
