ViTAE-Transformer / ViTPose

The official repo for [NeurIPS'22] "ViTPose: Simple Vision Transformer Baselines for Human Pose Estimation" and [TPAMI'23] "ViTPose++: Vision Transformer for Generic Body Pose Estimation"
Apache License 2.0
1.34k stars 181 forks source link

Train on custom dataset - number of keypoints #113

Open PallottaEnrico opened 1 year ago

PallottaEnrico commented 1 year ago

Hi, is there a quick way to train this model on a dataset with a different number of keypoints?

jojolee123 commented 8 months ago

I'm trying to solve the same problem, can we talk about it,vx: WAN-YIL

Srinikstudent commented 3 months ago

are you guys able to train on a custom dataset?