| dataset_info: | |
| features: | |
| - name: text | |
| sequence: string | |
| splits: | |
| - name: train | |
| num_bytes: 327344603 | |
| num_examples: 668582 | |
| - name: validation | |
| num_bytes: 8406146 | |
| num_examples: 17144 | |
| download_size: 189165954 | |
| dataset_size: 335750749 | |
| configs: | |
| - config_name: default | |
| data_files: | |
| - split: train | |
| path: data/train-* | |
| - split: validation | |
| path: data/validation-* | |
| # Tiny Conversations | |
| ## Overview | |
| This dataset consists of dialogue samples sourced from two main resources: the **Cornell Movie Dialogs** and the **Taiga TV Series Subtitles**. The dataset primarily contains Russian language dialogues and is designed for various natural language processing tasks such as language modeling, and dialogue systems. | |
| ### Sources | |
| 1. **Cornell Movie Dialogs**: | |
| - **Source**: [Cornell Movie Dialogs](https://github.com/Koziev/NLP_Datasets) | |
| - **License**: CC0-1.0 | |
| - **Description**: This dataset includes cleaned subtitles from a collection of movie dialogues. Notably, many dialogues are sampled from the middle of conversations. | |
| 2. **Taiga TV Series Subtitles**: | |
| - **Source**: [Russian Subtitles Dataset](https://github.com/dbklim/Russian_subtitles_dataset) | |
| - **License**: Apache-2.0 | |
| - **Description**: The dataset is based on the Taiga corpus, specifically from a collection of subtitles across 347 TV series in multiple languages. For this dataset, only the Russian language subtitles were retained. |