庞

How can I get the key points in the inference phase?

1

Thanks for your great job first! In my task, I need to extract the key points in the video, so which files should I read? Or does any utils in...

Hi, thanks for your nice paper. I met a question that in your paper you say the numbers of frames of the videos are variable. So how do you fuse...

Question of the parameter 'train_seg_conf_th'

1

Thank you for your work! I found that the parameter 'train_seg_conf_th' effects the model performance strongly on the ShanghaiTech dataset. After I reading the code, I found that this parameter...

Some question about the experimental results

2

**Thanks for your nice work first! I run the code without any change, but I cannot get the similar result on ShanghaiTech as the paper post, which is lower than...

你好，感谢你们的开源工作。在主页中你们说使用了中英文数据约1000万个样本，请问你们能再说明一下其中中文数据样本的数量吗？谢谢。

Thanks for your work first! According to my own experiments on ShanghaiTech, the variation of AUC results are big with different initializations, so I want to know how to get...

Some trajectories missed in the frames and the question in table 1.

4

Hello, thanks for your inspiring work first. I have a question that after I visualized the keypoint trajectories, I found that some trajectories of the targets may be missed, especially...

Can I find the paper?

1

Thanks for your nice work! I am curious that whether you have published the paper because I want to get more detail destricription. Thanks again!

Some tools are missing

1

Thanks for your great work. After reading the code, i found that some tools are not implemented in the code, for example, the Evaluate_Action_Feasibility tool and Justify_Decision_Logic tool. So where...

Thank you for your work. I conducted tests using Qwen3-235B-A20B, and the hallucination rate is as high as 92%. The main issue is that a large number of segments are...