README.md

---
dataset_info:
- config_name: C
  features:
  - name: blob_id
    dtype: large_string
  - name: language
    dtype: large_string
  - name: repo_name
    dtype: large_string
  - name: path
    dtype: large_string
  - name: src_encoding
    dtype: large_string
  - name: length_bytes
    dtype: int64
  - name: score
    dtype: float64
  - name: int_score
    dtype: int64
  - name: detected_licenses
    large_list: large_string
  - name: license_type
    dtype: large_string
  splits:
  - name: train
    num_bytes: 1100442974
    num_examples: 5848375
  download_size: 571816053
  dataset_size: 1100442974
- config_name: CSharp
  features:
  - name: blob_id
    dtype: string
  - name: language
    dtype: string
  - name: repo_name
    dtype: string
  - name: path
    dtype: string
  - name: src_encoding
    dtype: string
  - name: length_bytes
    dtype: int64
  - name: score
    dtype: float64
  - name: int_score
    dtype: int64
  splits:
  - name: train
    num_bytes: 1868459403.9283383
    num_examples: 11425016
  download_size: 1217425523
  dataset_size: 1868459403.9283383
- config_name: Cpp
  features:
  - name: blob_id
    dtype: large_string
  - name: language
    dtype: large_string
  - name: repo_name
    dtype: large_string
  - name: path
    dtype: large_string
  - name: src_encoding
    dtype: large_string
  - name: length_bytes
    dtype: int64
  - name: score
    dtype: float64
  - name: int_score
    dtype: int64
  - name: detected_licenses
    large_list: large_string
  - name: license_type
    dtype: large_string
  splits:
  - name: train
    num_bytes: 3167426435
    num_examples: 16246746
  download_size: 1632803797
  dataset_size: 3167426435
- config_name: Go
  features:
  - name: blob_id
    dtype: string
  - name: language
    dtype: string
  - name: repo_name
    dtype: string
  - name: path
    dtype: string
  - name: src_encoding
    dtype: string
  - name: length_bytes
    dtype: int64
  - name: score
    dtype: float64
  - name: int_score
    dtype: int64
  splits:
  - name: train
    num_bytes: 264121853.30269703
    num_examples: 1917163
  download_size: 172795994
  dataset_size: 264121853.30269703
- config_name: Java
  features:
  - name: blob_id
    dtype: string
  - name: language
    dtype: string
  - name: repo_name
    dtype: string
  - name: path
    dtype: string
  - name: src_encoding
    dtype: string
  - name: length_bytes
    dtype: int64
  - name: score
    dtype: float64
  - name: int_score
    dtype: int64
  splits:
  - name: train
    num_bytes: 8052241351.786225
    num_examples: 44990158
  download_size: 5225632804
  dataset_size: 8052241351.786225
- config_name: JavaScript
  features:
  - name: blob_id
    dtype: string
  - name: language
    dtype: string
  - name: repo_name
    dtype: string
  - name: path
    dtype: string
  - name: src_encoding
    dtype: string
  - name: length_bytes
    dtype: int64
  - name: score
    dtype: float64
  - name: int_score
    dtype: int64
  splits:
  - name: train
    num_bytes: 2047070211.6887486
    num_examples: 13253431
  download_size: 1270586356
  dataset_size: 2047070211.6887486
- config_name: Markdown
  features:
  - name: blob_id
    dtype: string
  - name: language
    dtype: string
  - name: repo_name
    dtype: string
  - name: path
    dtype: string
  - name: src_encoding
    dtype: string
  - name: length_bytes
    dtype: int64
  - name: score
    dtype: float64
  - name: int_score
    dtype: int64
  splits:
  - name: train
    num_bytes: 3084028046.5768914
    num_examples: 20687077
  download_size: 2023721165
  dataset_size: 3084028046.5768914
- config_name: PHP
  features:
  - name: blob_id
    dtype: string
  - name: language
    dtype: string
  - name: repo_name
    dtype: string
  - name: path
    dtype: string
  - name: src_encoding
    dtype: string
  - name: length_bytes
    dtype: int64
  - name: score
    dtype: float64
  - name: int_score
    dtype: int64
  splits:
  - name: train
    num_bytes: 1489550005.454035
    num_examples: 9914497
  download_size: 968613058
  dataset_size: 1489550005.454035
- config_name: Python
  features:
  - name: blob_id
    dtype: string
  - name: language
    dtype: string
  - name: repo_name
    dtype: string
  - name: path
    dtype: string
  - name: src_encoding
    dtype: string
  - name: length_bytes
    dtype: int64
  - name: score
    dtype: float64
  - name: int_score
    dtype: int64
  splits:
  - name: train
    num_bytes: 3720485148.4425817
    num_examples: 25286019
  download_size: 2468949715
  dataset_size: 3720485148.4425817
- config_name: Ruby
  features:
  - name: blob_id
    dtype: string
  - name: language
    dtype: string
  - name: repo_name
    dtype: string
  - name: path
    dtype: string
  - name: src_encoding
    dtype: string
  - name: length_bytes
    dtype: int64
  - name: score
    dtype: float64
  - name: int_score
    dtype: int64
  splits:
  - name: train
    num_bytes: 442935449.42350453
    num_examples: 2976874
  download_size: 280009998
  dataset_size: 442935449.42350453
- config_name: Rust
  features:
  - name: blob_id
    dtype: string
  - name: language
    dtype: string
  - name: repo_name
    dtype: string
  - name: path
    dtype: string
  - name: src_encoding
    dtype: string
  - name: length_bytes
    dtype: int64
  - name: score
    dtype: float64
  - name: int_score
    dtype: int64
  splits:
  - name: train
    num_bytes: 158802257.43148246
    num_examples: 1135379
  download_size: 100605171
  dataset_size: 158802257.43148246
- config_name: SQL
  features:
  - name: blob_id
    dtype: string
  - name: language
    dtype: string
  - name: repo_name
    dtype: string
  - name: path
    dtype: string
  - name: src_encoding
    dtype: string
  - name: length_bytes
    dtype: int64
  - name: score
    dtype: float64
  - name: int_score
    dtype: int64
  splits:
  - name: train
    num_bytes: 384998995.138193
    num_examples: 2504412
  download_size: 256836639
  dataset_size: 384998995.138193
- config_name: Shell
  features:
  - name: blob_id
    dtype: string
  - name: language
    dtype: string
  - name: repo_name
    dtype: string
  - name: path
    dtype: string
  - name: src_encoding
    dtype: string
  - name: length_bytes
    dtype: int64
  - name: score
    dtype: float64
  - name: int_score
    dtype: int64
  splits:
  - name: train
    num_bytes: 586236950.3127995
    num_examples: 4133547
  download_size: 386805586
  dataset_size: 586236950.3127995
- config_name: Swift
  features:
  - name: blob_id
    dtype: string
  - name: language
    dtype: string
  - name: repo_name
    dtype: string
  - name: path
    dtype: string
  - name: src_encoding
    dtype: string
  - name: length_bytes
    dtype: int64
  - name: score
    dtype: float64
  - name: int_score
    dtype: int64
  splits:
  - name: train
    num_bytes: 408330793.98799247
    num_examples: 2454309
  download_size: 254708210
  dataset_size: 408330793.98799247
- config_name: TypeScript
  features:
  - name: blob_id
    dtype: string
  - name: language
    dtype: string
  - name: repo_name
    dtype: string
  - name: path
    dtype: string
  - name: src_encoding
    dtype: string
  - name: length_bytes
    dtype: int64
  - name: score
    dtype: float64
  - name: int_score
    dtype: int64
  splits:
  - name: train
    num_bytes: 709628890.4418851
    num_examples: 4290356
  download_size: 419501643
  dataset_size: 709628890.4418851
configs:
- config_name: C
  data_files:
  - split: train
    path: C/train-*
- config_name: CSharp
  data_files:
  - split: train
    path: CSharp/train-*
- config_name: Cpp
  data_files:
  - split: train
    path: Cpp/train-*
- config_name: Go
  data_files:
  - split: train
    path: Go/train-*
- config_name: Java
  data_files:
  - split: train
    path: Java/train-*
- config_name: JavaScript
  data_files:
  - split: train
    path: JavaScript/train-*
- config_name: Markdown
  data_files:
  - split: train
    path: Markdown/train-*
- config_name: PHP
  data_files:
  - split: train
    path: PHP/train-*
- config_name: Python
  data_files:
  - split: train
    path: Python/train-*
- config_name: Ruby
  data_files:
  - split: train
    path: Ruby/train-*
- config_name: Rust
  data_files:
  - split: train
    path: Rust/train-*
- config_name: SQL
  data_files:
  - split: train
    path: SQL/train-*
- config_name: Shell
  data_files:
  - split: train
    path: Shell/train-*
- config_name: Swift
  data_files:
  - split: train
    path: Swift/train-*
- config_name: TypeScript
  data_files:
  - split: train
    path: TypeScript/train-*
---
# 💻 Stack-Edu

Stack-Edu is a 125B token dataset of educational code filtered from [The Stack v2](https://huggingface.co/datasets/bigcode/the-stack-v2), precisely the curated training corpus of [StarCoder2](https://arxiv.org/abs/2402.19173) models denoted StarCoder2Data. It is intended for Language Models training.

This dataset was curated using a classifier-based filtering strategy, inspired by [📚 FineWeb-Edu](https://huggingface.co/datasets/HuggingFaceFW/fineweb-edu), to retain only the highest-quality educational programming content.

## Downloading the data
This dataset only contains the SWHIDs to download the code files and not the content of the files itself. The contents can be downloaded from Software Heritage's S3 bucket to ensure data compliance. 
Please refer to [the-stack-v2](https://huggingface.co/datasets/bigcode/the-stack-v2-train-full-ids) for the data license.

When running on a 16-core AWS `us-east-1` instance, this script takes ~6 hours to download the files: 
```python
import boto3
import gzip
from datasets import load_dataset
from botocore.exceptions import ClientError

num_proc = 16
s3 = boto3.client('s3')
bucket_name = "softwareheritage"

def download_contents(blob_id):
    key = f"content/{blob_id}"
    try:
        obj = s3.get_object(Bucket=bucket_name, Key=key)
        with gzip.GzipFile(fileobj=obj['Body']) as fin:
            content = fin.read().decode("utf-8", errors="ignore")
        return {"text": content, "download_success": True}
    except ClientError as e:
        if e.response['Error']['Code'] == 'NoSuchKey':
            print(f"File not found: {key}")
            return {"text": "", "download_success": False}
        else:
            raise

# For Python
ds = load_dataset("HuggingFaceTB/stack-edu", "Python", split="train", num_proc=num_proc)
ds = ds.map(download_contents, input_columns="blob_id", num_proc=num_proc)

# Filter out failed downloads
ds = ds.filter(lambda x: x['download_success'])

# Optionally, print the first example to verify the data
print(ds[0])
```

## Details
The table below shows the number of tokens in each programming language using [SmolLM2](https://huggingface.co/HuggingFaceTB/SmolLM2-1.7B-Instruct) tokenizer.

| Language   | Stack-Edu (B tokens) |
|------------|----------------------|
| Python     | 21.8                 |
| Cpp        | 16.0                 |
| Markdown   | 14.0                 |
| C          | 11.1                 |
| JavaScript | 11.1                 |
| Java       | 42.1                 |
| SQL        | 9.62                 |
| PHP        | 9.07                 |
| C-Sharp    | 8.87                 |
| TypeScript | 3.03                 |
| Shell      | 3.13                 |
| Swift      | 1.83                 |
| Go         | 1.80                 |
| Rust       | 1.75                 |
| Ruby       | 1.61                 |


## Dataset curation and results
To build Stack-Edu, we:

- Selected 15 largest programming languages from StarCoder2Data 
- Trained 15 language-specific classifiers, using [StarEncoder](https://huggingface.co/bigcode/starencoder) model on synthetic annotations generated by Llama3-70B-Instruct. The classifiers for each language are available in this [collection](https://huggingface.co/collections/HuggingFaceTB/the-ultimate-collection-of-code-classifiers-67b5aa3eb8994a4b71453005).
- Applied a filtering threshold of 3 (out of 5) to retain highly educational content, except for Java, which performed best at threshold 2.

Stack-Edu shows consistent improvement over StarCoder2data on all the programming languages on MultiPL-E benchmark.

<img src="https://cdn-uploads.huggingface.co/production/uploads/61c141342aac764ce1654e43/GWnPgD0diMu0I8buK6mvG.png" width="600"/>

### Citation Information

```
@misc{allal2025smollm2smolgoesbig,
      title={SmolLM2: When Smol Goes Big -- Data-Centric Training of a Small Language Model}, 
      author={Loubna Ben Allal and Anton Lozhkov and Elie Bakouch and Gabriel Martín Blázquez and Guilherme Penedo and Lewis Tunstall and Andrés Marafioti and Hynek Kydlíček and Agustín Piqueres Lajarín and Vaibhav Srivastav and Joshua Lochner and Caleb Fahlgren and Xuan-Son Nguyen and Clémentine Fourrier and Ben Burtenshaw and Hugo Larcher and Haojun Zhao and Cyril Zakka and Mathieu Morlon and Colin Raffel and Leandro von Werra and Thomas Wolf},
      year={2025},
      eprint={2502.02737},
      archivePrefix={arXiv},
      primaryClass={cs.CL},
      url={https://arxiv.org/abs/2502.02737}, 
}
```