jablonkagroup/MaCBench-Ablations

该数据集是一个包含多个配置的复合数据集，每个配置都包含了图像和其他相关特征，以及对应的训练集分割信息。具体特征包括唯一标识符(uuid)、图像(image)、标识符(canary)、名称(name)、描述(description)、关键词(keywords)、首选分数(preferred_score)、指标(metrics)、示例(examples)和相对容差(relative_tolerance)。每个配置的训练集都包含了名称、字节数和示例数量等信息。

This dataset is a composite dataset containing multiple configurations, each with its own set of features and training split information. The features include a unique identifier (uuid), image, identifier (canary), name, description, keywords, preferred score, metrics, examples, and relative tolerance. Each configurations training split includes information such as the name, number of bytes, and the number of examples.