This repository is an official PyTorch implementation of the paper "Learnable Triangulation of Human Pose" (ICCV 2019, oral). Proposed method archives state-of-the-art results in multi-view 3D human pose estimation!
Why do you create results['indexes'] during training if you don't use them afterwards (except for dumping the results) ? It seems they would be useful when evaluating shuffled data ... or at least in my version of torch set_epoch does not seem to have an effect in the batches order when not training distributed-ly ...
Why do you create
results['indexes']
during training if you don't use them afterwards (except for dumping the results) ? It seems they would be useful when evaluating shuffled data ... or at least in my version of torch set_epoch does not seem to have an effect in the batches order when not training distributed-ly ...