search

amathislab / BUCTD

[ICCV 2023] "Rethinking pose estimation in crowds: overcoming the detection information-bottleneck and ambiguity"
https://amathislab.github.io/BUCTD
Apache License 2.0
84 stars 8 forks source link

Pretrained model keypoint number mismatch #5

Closed Laudwika closed 10 months ago

Laudwika commented 11 months ago

Hi, I'm trying to test the code, and while the coco config file and weights work well, there is a mismatch between the ochuman and crowdpose weights.

The OcHuman weights says their missing weights with 17 keypoints on the coco yaml

The CrowdPose weights say they are missing weights with 14 keypoints on the crowdpose yaml

Also, will there be plans to release an inference code to perform single image inference?

Thank you, I appreciate your work!

zhoumu53 commented 10 months ago

Hi @Laudwika , thank you for your message, and sorry for the confusion.

Please find out the weights for OCHuman here.

For CrowdPose, we will upload the correct checkpoint asap.

AlexEMG commented 10 months ago

Addressed in #6

SHUOFEN2019416 commented 10 months ago

How to implement visualization? Thank you!

AlexEMG commented 10 months ago

@Laudwika, we updated the PR. You can also find all weights here: https://zenodo.org/records/10039883

@SHUOFEN2019416, please open a new PR for a new question. But yes, we can share code!