| --- |
| configs: |
| - config_name: default |
| data_files: |
| - split: train |
| path: data/clean-*.parquet |
| --- |
| |
| # websoup |
|
|
| This dataset is a collection of websites that were scraped based off of their simplicity (size, quality, obfuscation level) and inlined into simple & large html files. |
|
|
| If you want to build a stochastic parrot that spews random web pages (including js and css in a single output) then this dataset is decent (not perfect by any means). |
|
|
| Load with: |
|
|
| ```python |
| from datasets import load_dataset |
| ds = load_dataset("lowerkey/websoup", split="train") |
| ``` |
|
|
