hongsukchoi / Pose2Mesh_RELEASE

Official Pytorch implementation of "Pose2Mesh: Graph Convolutional Network for 3D Human Pose and Mesh Recovery from a 2D Human Pose", ECCV 2020
MIT License
677 stars 69 forks source link

test issue about PW3D #40

Closed Hao-hui closed 2 years ago

Hao-hui commented 2 years ago

Hi, I trained pose2mesh on Human36m using "pose2mesh_human36J_train_human36.yml" , how can I test this model on PW3D ? I'm looking forward to hearing from you. Many thanks!

hongsukchoi commented 2 years ago

You can make your own test yaml file and write the saved checkpoint's directory like here https://github.com/hongsukchoi/Pose2Mesh_RELEASE/blob/e91bdd69077e964c5591f7c52b6f484d5180bd12/asset/yaml/pose2mesh_human36J_test_human36.yml#L37

Hao-hui commented 2 years ago

Thank you for your prompt reply! As you said, I modified it and test_list as follows, but the problem doesn't seem to be solved. image I have attached the screenshot of the problem, I guess it may be the problem of the number of joints, please help me how to modify it?
image

hongsukchoi commented 2 years ago

aha.. now I understand the problem. Human36M and 3DPW have different joint sets. So basically you would not test the Human3.6M trained model on 3DPW.

check this out: https://github.com/hongsukchoi/Pose2Mesh_RELEASE#running-pose2mesh

Hao-hui commented 2 years ago

I'm sorry for my negligence. Thank you very much for your patient answer. I will close this.