我下载了tinyperson数据集，发现里面的文件夹层次比较复杂，json文件较多，有点分不清哪些是哪些。
我将mini_annotations/下的三个json文件进行统计，发现他们加起来只有69371个标注，而并非论文中写的72651个。
想请问我是否漏统计了一些json文件，究竟哪些json文件合在一起才是完整的数据集呢？
感谢解答！
I downloaded Tinyperson dataset and found that the folder hierarchy was complicated, with many json files, which was a bit confusing.
I counted the three json files under mini_annotations/ and found that they add up to only 69,371 annotations, rather than the 72,651 annotations written in the paper.
May I ask whether I missed some json files? Which json files together are the complete data set?
Thanks for your answer!
我下载了tinyperson数据集，发现里面的文件夹层次比较复杂，json文件较多，有点分不清哪些是哪些。
我将mini_annotations/下的三个json文件进行统计，发现他们加起来只有69371个标注，而并非论文中写的72651个。
想请问我是否漏统计了一些json文件，究竟哪些json文件合在一起才是完整的数据集呢？
感谢解答！
I downloaded Tinyperson dataset and found that the folder hierarchy was complicated, with many json files, which was a bit confusing.
I counted the three json files under mini_annotations/ and found that they add up to only 69,371 annotations, rather than the 72,651 annotations written in the paper.
May I ask whether I missed some json files? Which json files together are the complete data set?
Thanks for your answer!