README.md

---
license: mit
task_categories:
  - question-answering
language:
  - en
  - zh
configs:
  - config_name: futoshiki
    data_files:
      - split: dev
        path: futoshiki/dev.parquet
      - split: test
        path: futoshiki/test.parquet
      - split: train
        path: futoshiki/train.parquet
  - config_name: kukurasu
    data_files:
      - split: dev
        path: kukurasu/dev.parquet
      - split: test
        path: kukurasu/test.parquet
      - split: train
        path: kukurasu/train.parquet
  - config_name: minesweeper
    data_files:
      - split: dev
        path: minesweeper/dev.parquet
      - split: test
        path: minesweeper/test.parquet
      - split: train
        path: minesweeper/train.parquet
  - config_name: sudoku
    data_files:
      - split: dev
        path: sudoku/dev.parquet
      - split: test
        path: sudoku/test.parquet
      - split: train
        path: sudoku/train.parquet
---

# game_data

Multi-task reasoning dataset. Each task type is a separate config with train/test/dev splits.

## Usage

```python
from datasets import load_dataset

# Load a specific task and split
ds = load_dataset("CL-From-Nothing/game_data", "futoshiki", split="train")
ds = load_dataset("CL-From-Nothing/game_data", "minesweeper", split="test")
```

## Available Tasks

`futoshiki`, `kukurasu`, `minesweeper`, `sudoku`

## Available Splits

`train`, `test`, `dev`