Datasets:
add paper
#15
by
loubnabnl
HF staff
- opened
README.md
CHANGED
|
@@ -413,6 +413,8 @@ configs:
|
|
| 413 |
|
| 414 |
> 1.3 trillion tokens of the finest educational data the 🌐 web has to offer
|
| 415 |
|
|
|
|
|
|
|
| 416 |
## What is it?
|
| 417 |
|
| 418 |
📚 FineWeb-Edu dataset consists of **1.3T tokens** and **5.4T tokens** ([FineWeb-Edu-score-2](https://huggingface.co/datasets/HuggingFaceFW/fineweb-edu-score-2)) of educational web pages filtered from 🍷 FineWeb dataset. This is the 1.3 trillion version.
|
|
@@ -542,7 +544,7 @@ The dataset is released under the **Open Data Commons Attribution License (ODC-B
|
|
| 542 |
We plan to work on better educational classifier to improve the quality of FineWeb-Edu.
|
| 543 |
|
| 544 |
### Citation Information
|
| 545 |
-
|
| 546 |
```
|
| 547 |
@software{lozhkov2024fineweb-edu,
|
| 548 |
author = {Lozhkov, Anton and Ben Allal, Loubna and von Werra, Leandro and Wolf, Thomas},
|
|
|
|
| 413 |
|
| 414 |
> 1.3 trillion tokens of the finest educational data the 🌐 web has to offer
|
| 415 |
|
| 416 |
+
**Paper:** https://arxiv.org/abs/2406.17557
|
| 417 |
+
|
| 418 |
## What is it?
|
| 419 |
|
| 420 |
📚 FineWeb-Edu dataset consists of **1.3T tokens** and **5.4T tokens** ([FineWeb-Edu-score-2](https://huggingface.co/datasets/HuggingFaceFW/fineweb-edu-score-2)) of educational web pages filtered from 🍷 FineWeb dataset. This is the 1.3 trillion version.
|
|
|
|
| 544 |
We plan to work on better educational classifier to improve the quality of FineWeb-Edu.
|
| 545 |
|
| 546 |
### Citation Information
|
| 547 |
+
You can cite our paper https://arxiv.org/abs/2406.17557 or this dataset:
|
| 548 |
```
|
| 549 |
@software{lozhkov2024fineweb-edu,
|
| 550 |
author = {Lozhkov, Anton and Ben Allal, Loubna and von Werra, Leandro and Wolf, Thomas},
|