about OTTA issue

[linqiuxia-lynn]

Hi, thank you for your excellent work! I noticed that there’s no available implementation for your online test-time adaptation (OTTA), and I’m curious about your experimental results in this context. Specifically, I’m wondering how OTTA achieved strong results with the test-time adaptation setting (MPJPE: 90mm vs 87mm), given that one-sample adaptation can often lead to overfitting when there is noise in the 2D data. Would it be possible for you to release the training code for OTTA?