Explain why crawl coverage starts at CC-MAIN-2024-22, link FineWeb-2 dataset
Browse files
README.md
CHANGED
|
@@ -23,6 +23,8 @@ A FineWeb-style Greek-language text dataset extracted from Common Crawl, followi
|
|
| 23 |
|
| 24 |
Pipeline source: [github.com/alexliap/greek-cc](https://github.com/alexliap/greek-cc).
|
| 25 |
|
|
|
|
|
|
|
| 26 |
## Dataset structure
|
| 27 |
|
| 28 |
Each processed crawl (e.g. `CC-MAIN-2024-22`) contributes two kinds of files:
|
|
@@ -72,6 +74,7 @@ Licensed under [ODC-By 1.0](https://opendatacommons.org/licenses/by/1-0/), match
|
|
| 72 |
|
| 73 |
## References
|
| 74 |
|
|
|
|
| 75 |
- Penedo et al., [The FineWeb Datasets: Decanting the Web for the Finest Text Data at Scale](https://arxiv.org/abs/2406.17557) (arXiv 2406.17557)
|
| 76 |
- Penedo et al., [FineWeb2: One Pipeline to Scale Them All](https://arxiv.org/abs/2506.20920) (arXiv 2506.20920)
|
| 77 |
- [datatrove](https://github.com/huggingface/datatrove), the processing library both FineWeb papers and this pipeline are built on
|
|
|
|
| 23 |
|
| 24 |
Pipeline source: [github.com/alexliap/greek-cc](https://github.com/alexliap/greek-cc).
|
| 25 |
|
| 26 |
+
Crawl coverage starts at `CC-MAIN-2024-22` rather than Common Crawl's earliest snapshots: this project picks up right where the [FineWeb-2](https://huggingface.co/datasets/HuggingFaceFW/fineweb-2) dataset's own Greek (`ell_Grek`) subset leaves off (2013 through April 2024), so it extends FineWeb-2's Greek coverage forward instead of re-extracting and re-deduping ground FineWeb-2 already covers.
|
| 27 |
+
|
| 28 |
## Dataset structure
|
| 29 |
|
| 30 |
Each processed crawl (e.g. `CC-MAIN-2024-22`) contributes two kinds of files:
|
|
|
|
| 74 |
|
| 75 |
## References
|
| 76 |
|
| 77 |
+
- [FineWeb-2 dataset](https://huggingface.co/datasets/HuggingFaceFW/fineweb-2), whose Greek (`ell_Grek`) subset this project picks up after
|
| 78 |
- Penedo et al., [The FineWeb Datasets: Decanting the Web for the Finest Text Data at Scale](https://arxiv.org/abs/2406.17557) (arXiv 2406.17557)
|
| 79 |
- Penedo et al., [FineWeb2: One Pipeline to Scale Them All](https://arxiv.org/abs/2506.20920) (arXiv 2506.20920)
|
| 80 |
- [datatrove](https://github.com/huggingface/datatrove), the processing library both FineWeb papers and this pipeline are built on
|