NealCaren/inkbench-baseline-detailed-test

该数据集包含了对InkBench数据集的机器生成转录文本。数据集结构包括每个示例的图像文件名、文档类型、真实转录文本（来自InkBench）和模型生成的转录文本等字段。原始图像未包含在内以减少数据集大小。该数据集需与原始InkBench数据集一起使用以进行评估。This dataset contains machine-generated transcriptions of the InkBench dataset. The dataset structure includes fields such as image name, document type, ground truth transcription (from InkBench), and model-generated transcription for each example. The original images are NOT included to reduce dataset size. This dataset should be used with the original InkBench dataset for evaluation.