liaowang0125
liaowang0125
我看见process.py里面使用的是json后缀文件 ``` with open(root_path+'annotations/'+'vrd-train-anno.json') as f: train_image_list = json.load(f) ``` 但是官网下载下来的的csv格式文件,想问下json和csv遍历的时候图片的顺序是一致的么,因为涉及到图片重命名。我这个问题的原因是我是从官网下载的数据,不是从[Rongjie Li](https://github.com/SHTUPLUS/PySGG/blob/main/DATASET.md) 这一步下载的
I made the zip with reference to this website and running the check script ensures that the format is fine: https://github.com/pixelite1201/agora_evaluation/blob/master/docs/prediction_format.md  But when I submit it on the website,it...
I run the agora val evaluation scripts, but there are many differences between the results and the paper, evaluation results in the screenshot below, only the MVE and MPJPE are...