dan-tabsdata/tabsdata_benchmarking_source_data
收藏Hugging Face2026-04-23 更新2026-05-03 收录
下载链接：
https://hf-mirror.com/datasets/dan-tabsdata/tabsdata_benchmarking_source_data
下载链接
链接失效反馈官方服务：
资源简介：
---
configs:
- config_name: small
data_files:
- split: train
path: nyc_data_small.parquet
- config_name: medium
data_files:
- split: train
path: nyc_data_medium.parquet
- config_name: large
data_files:
- split: train
path: nyc_data_large.parquet
---
# TabsData Benchmarking Source Data
NYC-based benchmarking datasets in three sizes for performance testing and benchmarking workflows.
## Configurations
### `small`
A small-scale subset (~20 MB) suitable for quick tests and development iteration.
```python
from datasets import load_dataset
ds = load_dataset("dan-tabsdata/tabsdata_benchmarking_source_data", "small")
```
### `medium`
A medium-scale subset (~192 MB) for moderate benchmarking workloads.
```python
from datasets import load_dataset
ds = load_dataset("dan-tabsdata/tabsdata_benchmarking_source_data", "medium")
```
### `large`
A large-scale dataset (~2 GB) for full-scale performance benchmarking.
```python
from datasets import load_dataset
ds = load_dataset("dan-tabsdata/tabsdata_benchmarking_source_data", "large")
```
提供机构：
dan-tabsdata
