| --- |
| license: cc-by-sa-4.0 |
| task_categories: |
| - text-generation |
| language: |
| - ja |
| size_categories: |
| - 100K<n<1M |
| --- |
| |
| # Description |
| This is a templated version of data from ~40 Japanese open source downstream task datasets. |
| This dataset was built with the intention of instruction tuning LLMs with high quality data that was not machine translated to and from Japanese. |
| The raw data version can be found [here](https://huggingface.co/datasets/Ego/jpflan-raw). |
|
|
| # Dataset format |
| This dataset version contains at most 20000 samples for each dataset, half of which are 0-shot examples, and the other half are few-shot examples (max 3). |
| To create your own version of the dataset, please refer to [here](https://github.com/Egojr/jpflan). |
|
|
| ```json |
| { |
| "dataset_name": 'Original dataset name' |
| "input": 'Input text' |
| "output": 'Output text', |
| "shots": 'Number of shots in input', |
| } |
| ``` |
