datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
visual-novels
Visual Novel Dataset
This dataset contains parsed Visual Novel scripts for training language models. The dataset consists of approximately 60 million tokens of parsed scripts.
Dataset Structure
The dataset follows a general structure for visual novel scripts:
Dialogue lines: Dialogue lines are formatted with the speaker's name followed by a colon, and the dialogue itself enclosed in quotes. For example:
John: "Hello, how are you?"
Actions and narration: Actions and… See the full description on the dataset page: https://huggingface.co/datasets/alpindale/visual-novels.visual-novels-v1.1
Visual Novels v1.1 (WIP)
This dataset contains parsed Visual Novel scripts.
Dataset Structure
We provide 2 variants of the same dataset:
Extracted (Not Yet released)
Contains the raw unformatted version directly extracted from the game's files.
jsonl
Parsed versions of the raw extracted versions. A sample is provided below:
{
"meta": {
"game": "<Game Title>.jsonl",
"scene_key": "<Scene Key>"
},
"namedconversation": [
{… See the full description on the dataset page: https://huggingface.co/datasets/KaraKaraWitch/visual-novels-v1.1.