Pose-Group / DCPose

This is an official implementation of our CVPR 2021 paper "Deep Dual Consecutive Network for Human Pose Estimation" (https://openaccess.thecvf.com/content/CVPR2021/papers/Liu_Deep_Dual_Consecutive_Network_for_Human_Pose_Estimation_CVPR_2021_paper.pdf)

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Questions about the performance

zhangrj91 opened this issue · comments

Thanks for your excellent work!

I run the code successfully and I can use your pretrained model to get a result of 82.8 for PoseTrack17 validation dataset, but if I use the provided code and config files to train a new model and the result is only 81.6.

So I want to know that is there any other settings about the training or the config files about the training are not the updated version?

您好,请问一下您使用预训练模型之后是跑的那个代码可以得出81.6的结果?

您好,请问一下您使用预训练模型之后是跑的那个代码可以得出81.6的结果?

是的

@zhangrj91 你好,我们推荐直接使用预训练模型即可,因为训练模型时我们未设定死随机种子,初始化比较随机;关于其他配置参数,都已写在配置文件未改动。

commented

我才复现出80.9mAP