v-urushkin/NaturalGPT2-small
Text Generation
•
Updated
•
49
•
1
Error code: ClientConnectionError
The dataset contains 2 million natural language sentences. The following data sources were used: three news sub-corpuses Taiga (Lenta.ru, Interfax, N+1) and Russian Wikipedia texts
Data source | Number of sentences taken |
---|---|
Sub-corpuse Lenta.ru from "Taiga" | 124 681 |
Sub-corpuse Interfax from "Taiga" | 230 246 |
Sub-corpuse N+1 from "Taiga" | 33 268 |
Texts of Russian Wikipedia | 1 611 805 |
The texts do not contain any characters other than Cyrillic alphabet.
Sample sentences: