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

cannot import name 'deform_conv_cuda' from partially initialized module 'thirdparty.deform_conv' (most likely due to a circular import)

DWCTOD opened this issue · comments

大佬您好,我想运行试一下效果,运行 video.py 的时候报了这个错误,不知道要怎么解决,谢谢啦

cannot import name 'deform_conv_cuda' from partially initialized module 'thirdparty.deform_conv' (most likely due to a circular import)

报错位置:
File "xxx / DCPose/thirdparty/deform_conv/functions/deform_conv.py", line 5, in
from .. import deform_conv_cuda

Hi DWCTOD:
Checking whether the following commands have been executed.

cd thirdparty/deform_conv
python setup.py develop

If the above command is executed successfully, the files below will be available in the thirdparty/deform_conv directory.

deform_conv_cuda.cpython-36m-x86_64-linux-gnu.so
deform_pool_cuda.cpython-36m-x86_64-linux-gnu.so

谢谢大佬的解答,上面的问题解决了
使用 https://drive.google.com/drive/folders/1VPcwo9jVhnJpf5GVwR2-za1PFE6bCOXE 的权重文件好像不能用

image

DCPose/tools/inference.py", line 63, in
model = get_inference_model()
in get_inference_model
model_state_dict = {k.replace('module.', ''): v for k, v in checkpoint_dict['state_dict'].items()}
KeyError: 'state_dict'

谢谢大佬的解答,上面的问题解决了
使用 https://drive.google.com/drive/folders/1VPcwo9jVhnJpf5GVwR2-za1PFE6bCOXE 的权重文件好像不能用

image

DCPose/tools/inference.py", line 63, in
model = get_inference_model()
in get_inference_model
model_state_dict = {k.replace('module.', ''): v for k, v in checkpoint_dict['state_dict'].items()}
KeyError: 'state_dict'

我们更新项目,现在可以直接从 https://drive.google.com/drive/folders/1VPcwo9jVhnJpf5GVwR2-za1PFE6bCOXE 获取DCPose的权重。😁

大佬您好,请问如果要对多人进行姿态估计的话,要修改哪些地方

commented

大佬您好,请问如果要对多人进行姿态估计的话,要修改哪些地方

不需要修改,可以直接进行多人姿态估计