yassinetb/COGITAO

COGITAO数据集包含两种类型的数据：泛化和组合性。每种类型都有不同的实验设置和实验，数据以.parquet文件形式存储在嵌套的子文件夹中。每个文件包含输入网格、输出网格、转换套件和任务标识符等键。在组合性研究中，还提供了演示输入和演示输出键。该数据集可以使用Python的datasets库加载。

The COGITAO dataset consists of two types of data: generalization and compositionality. Each type has different experiment settings and experiments, with data stored in .parquet files within nested subfolders. Each file contains keys for input grid, output grid, transformation suite, and task key. In the compositionality study, additional keys for demonstration input and output are provided. The dataset can be loaded using the datasets library in Python.