# fineweb-2 by FineData: AI Training Dataset
Source: https://savrn.com/datasets/fineweb-2
Markdown alternate of the page above; the site index is https://savrn.com/llms.txt

---

## Dataset Card

By FineData, published under odc-by, revision af9c13333eb9.

A sparkling update with 1000s of languages

### Table of Contents

- FineWeb2
- What is it?
- Languages and available subsets
  - How many tokens?
- Changelog
- How to download and use FineWeb2
  - Usingdatatrove
  - Using huggingface_hub
  - Using datasets
- Dataset processing steps
  - Language Identification
  - Deduplication
  - Data Filtering
  - PII Anonymization and fixes
- Dataset performance evaluation and ablations
  - Hyper-parameters for ablation models
  - Score normalization
  - Global scores across languages
  - Comparison with other datasets
- Dataset card for FineWeb2
- Dataset Description
  - Dataset Summary
- Dataset Structure
  - Data Instances
  - Data Fields
  - Data Splits
- Dataset Creation
  - Curation Rationale
  - Source Data
  - Data processing steps
  - Annotations
  - Personal and Sensitive Information and opt-out
- Considerations for Using the Data
  - Social Impact of Dataset
  - Discussion of Biases
  - Other Known Limitations
- Additional Information
  - Licensing Information
  - Future work and community initiatives
- Citation Information

### What is it?

This is the second iteration of the popular[FineWeb](https://savrn.com/datasets/fineweb)dataset, bringing high quality pretraining data to over 1000 languages.

[Read the full dataset card (4,166 words)](https://savrn.com/datasets/fineweb-2/card)

## Structure

### aai_Latn 623 rows

| Split | Rows | Size |
| --- | --- | --- |
| test | 8 | 28.3 KB |
| train | 615 | 2.9 MB |

textstringidstringdumpstringurlstringdatestringfile_pathstringlanguagestringlanguage_scorefloat64language_scriptstringminhash_cluster_sizeint64top_langsstring

### aak_Latn 1,568 rows

| Split | Rows | Size |
| --- | --- | --- |
| test | 8 | 34.7 KB |
| train | 1,560 | 4.7 MB |

textstringidstringdumpstringurlstringdatestringfile_pathstringlanguagestringlanguage_scorefloat64language_scriptstringminhash_cluster_sizeint64top_langsstring

### aau_Latn 2,015 rows

| Split | Rows | Size |
| --- | --- | --- |
| test | 23 | 35.5 KB |
| train | 1,992 | 2.9 MB |

textstringidstringdumpstringurlstringdatestringfile_pathstringlanguagestringlanguage_scorefloat64language_scriptstringminhash_cluster_sizeint64top_langsstring

### aaz_Latn 2,584 rows

| Split | Rows | Size |
| --- | --- | --- |
| test | 25 | 51.2 KB |
| train | 2,559 | 3.5 MB |

textstringidstringdumpstringurlstringdatestringfile_pathstringlanguagestringlanguage_scorefloat64language_scriptstringminhash_cluster_sizeint64top_langsstring

### aba_Latn 199 rows

| Split | Rows | Size |
| --- | --- | --- |
| test | 1 | 1.6 KB |
| train | 198 | 759.1 KB |

textstringidstringdumpstringurlstringdatestringfile_pathstringlanguagestringlanguage_scorefloat64language_scriptstringminhash_cluster_sizeint64top_langsstring

### abk_Cyrl 2,696 rows

| Split | Rows | Size |
| --- | --- | --- |
| test | 26 | 231.7 KB |
| train | 2,670 | 23.7 MB |

textstringidstringdumpstringurlstringdatestringfile_pathstringlanguagestringlanguage_scorefloat64language_scriptstringminhash_cluster_sizeint64top_langsstring

### abq_Cyrl 980 rows

| Split | Rows | Size |
| --- | --- | --- |
| test | 12 | 106.6 KB |
| train | 968 | 8.0 MB |

textstringidstringdumpstringurlstringdatestringfile_pathstringlanguagestringlanguage_scorefloat64language_scriptstringminhash_cluster_sizeint64top_langsstring

### abs_Latn 213 rows

| Split | Rows | Size |
| --- | --- | --- |
| test | 3 | 9.2 KB |
| train | 210 | 586.7 KB |

textstringidstringdumpstringurlstringdatestringfile_pathstringlanguagestringlanguage_scorefloat64language_scriptstringminhash_cluster_sizeint64top_langsstring

### abt_Latn 4,131 rows

| Split | Rows | Size |
| --- | --- | --- |
| test | 33 | 54.8 KB |
| train | 4,098 | 7.6 MB |

textstringidstringdumpstringurlstringdatestringfile_pathstringlanguagestringlanguage_scorefloat64language_scriptstringminhash_cluster_sizeint64top_langsstring

### abx_Latn 918 rows

| Split | Rows | Size |
| --- | --- | --- |
| test | 9 | 9.6 KB |
| train | 909 | 2.2 MB |

textstringidstringdumpstringurlstringdatestringfile_pathstringlanguagestringlanguage_scorefloat64language_scriptstringminhash_cluster_sizeint64top_langsstring

### aby_Latn 1,085 rows

| Split | Rows | Size |
| --- | --- | --- |
| test | 5 | 4.3 KB |
| train | 1,080 | 2.2 MB |

textstringidstringdumpstringurlstringdatestringfile_pathstringlanguagestringlanguage_scorefloat64language_scriptstringminhash_cluster_sizeint64top_langsstring

### abz_Latn 252 rows

| Split | Rows | Size |
| --- | --- | --- |
| test | 1 | 3.3 KB |
| train | 251 | 535.2 KB |

textstringidstringdumpstringurlstringdatestringfile_pathstringlanguagestringlanguage_scorefloat64language_scriptstringminhash_cluster_sizeint64top_langsstring

## Details

Repository

HuggingFaceFW/fineweb-2

Publisher

FineData

Task category

Text generation

Tags

Not stated by the source

Size category

n>1T

Languages

aai, aak, aau, aaz, aba, abi, abk, abn

Revision

af9c13333eb981300149d5ca60a8e9d659b276b9

Last updated

2025-10-27

## License and Download

License

odc-by

Access

No access gate

[Download from FineData](https://huggingface.co/datasets/HuggingFaceFW/fineweb-2)

Released by FineData through its official repository on Hugging Face.

## Models Trained on This Dataset

- Trained on (disclosed)[zh-pii-tagger-35m](https://savrn.com/models/zh-pii-tagger-35m)
- Trained on (disclosed)[prompt-injection-guard-base](https://savrn.com/models/prompt-injection-guard-base)
- Trained on (disclosed)[prompt-injection-guard-small](https://savrn.com/models/prompt-injection-guard-small)
- Trained on (disclosed)[Sanskrit-322M-Base](https://savrn.com/models/sanskrit-322m-base)
- Trained on (disclosed)[futolm-cs-base-gguf](https://savrn.com/models/futolm-cs-base-gguf)
- Trained on (disclosed)[viya-base](https://savrn.com/models/viya-base)

## FineData

[All models and datasets](https://savrn.com/model-publishers/huggingfacefw)

## Explore More

- [The dataset directory](https://savrn.com/datasets)
- [The model directory](https://savrn.com/models)
- [Research papers](https://savrn.com/papers)

## Source

- Repository metadata, read 2026-10-19.
- [Hugging Face record](https://huggingface.co/datasets/HuggingFaceFW/fineweb-2)
- [How the hub is built](https://savrn.com/model-hub/methodology)