---
title: arks_data
canonical_url: "https://www.modelscope.cn/datasets/xlangai/arks_data"
md_url: "https://www.modelscope.cn/datasets/xlangai/arks_data.md"
repository: xlangai/arks_data
last_updated: 2025-08-19
license: apache-2.0
storage_size: "852 MB"
downloads: 189
stars: 0
---

# arks_data

> arks_data - xlangai 在 ModelScope 开源的数据集。Dataset Card for Dataset Name

xlangai/arks_data 是 ModelScope 魔搭社区上的数据集，存储大小 852 MB，采用 apache-2.0 许可。

- **Repository**: xlangai/arks_data
- **License**: apache-2.0
- **Storage size**: 852 MB
- **Downloads**: 189
- **Stars**: 0
- **Last updated**: 2025-08-19

Source: https://www.modelscope.cn/datasets/xlangai/arks_data

---

# Dataset Card for Dataset Name

This dataset contains 4 sub-datasets, namely Pony, Ring, ScipyM, TensorflowM. You can find more information about this dataset from our paper **"ARKS: Active Retrieval in Knowledge Soup for Code Generation"**

paper Arxiv link: https://arxiv.org/abs/2402.12317

paper website: https://arks-codegen.github.io

# How to load this dataset

load one dataset:
```
from datasets import load_dataset
data_files = {"corpus": "Pony/Pony_docs.jsonl"}
dataset = load_dataset("xlangai/arks_data", data_files=data_files)
```

load several datasets:
```
from datasets import load_dataset
data_files = {"corpus": ["Pony/Pony_docs.jsonl", "Ring/Ring_docs.jsonl"]}
dataset = load_dataset("xlangai/arks_data", data_files=data_files)
```