| license: cc-by-sa-4.0 | |
| language: | |
| - hu | |
| # hogyankell | |
| Hungarian "how-to" dataset crawled from [hogyankell.hu](https://hogyankell.hu/) wiki. | |
| ## Token count | |
| Tokenizer: [`magyar-nlp-szine-java/exotic_modernbert_128k_tokenizer_modified`](https://huggingface.co/magyar-nlp-szine-java/exotic_modernbert_128k_tokenizer_modified) | |
| | Metric | Value | | |
| |---|---| | |
| | Total rows | 2,715 | | |
| | Total tokens | 2,104,244 | | |
| | Avg tokens/row | 775.0 | | |
| | Min tokens | 8 | | |
| | Max tokens | 18,772 | | |
| ## Columns | |
| - `pageid` — Wikipedia-style page ID | |
| - `title` — article title | |
| - `url` — source URL | |
| - `categories` — list of categories | |
| - `wikitext` — raw wikitext markup | |
| - `plain_text` — cleaned plain text | |
| - `token_count` — number of tokens (using the tokenizer above) | |