wangziren1 / pytorch_pose_proposal_networks Goto Github PK

Can you provide the reference paper?

because of the Internet problem,I cannot download your pretrained model, could you please give the baidu yun downloads?

can you tell me how to calculate the pckH and the mAP or give me the code about iit ?

thanks for your wonderful jobs~
I have some questions that I don't understand. I hope you can help me answer them.
1、src/demo.py line26-31, I don't know why I'm doing this with the input?
2、src/demo.py line72，I want to know why do I have to add 4？

thanks for your reply!

i want to know the accuracy of the pretrained models resnet18 and resnet50?
thanks

首先感谢实现。
1，文章中每个关节点都预测一个bbox，那么gt里面bbox的w和h是怎么定的？

Could you show the code about generating mpi_annotation.json? @wangziren1

When parsing the network output, why are the width and height of key points calculated as follows:
output[:, :, :, 4:96:6] = output[:, :, :, 4:96:6].pow(2) * cfg.IMG_SIZE output[:, :, :, 5:96:6] = output[:, :, :, 5:96:6].pow(2) * cfg.IMG_SIZE
why not below:
output[:, :, :, 4:96:6] = output[:, :, :, 4:96:6] * cfg.IMG_SIZE output[:, :, :, 5:96:6] = output[:, :, :, 5:96:6]* cfg.IMG_SIZE
thanks~

hi! thanks for your work ,when i train this repo i get nan loss no matter use the resnet18 or using the resnet50. can you give me some advices,thank you !

Hi, nice work. I would like to know what license this is distributed by as there is none specified.

Thank you.