ruoxi
ruoxi
HI，I'm interested in your great work, Could we also have extracted features of THUMOS 14 and release codes on it
same as you. Do you have any methods to solve it?Thank you
it does not work for me either. sad
what gpu nums and batch size should be set to reproduce the result in paper @Gengzigang
Have you solved the problem? I need to compile DCN v2 for RTX 3090
姿态估计都很慢
训练的话要快两天好像，不过我没有复现出结果，才81点多，不过没认真复现。而且因为下巴的坐标是网络直接预测出来的，不像是posewarper利用肩膀的坐标给出下巴的坐标（虽然我也不理解为什么这么做），所以其实直接跟posewarper的结果略有些不公平。如果测试的话提供了检测结果，总体时间比起关节点跟踪的话好多了-_-。 @Caicaizi-cdy
> > Hello, thanks for your interest! For the experiments on GTEA and BEOID, we utilized the features provided by SF-Net and you can find them [here](https://github.com/Flowerfan/SF-Net). On the other...
the result I reproduce is about 40.83mAP on GTEA, there is still some gap.