DecaYale

Hi, maybe not. We referred to PVNet's code base for this visualization. We may add this visualization part later.

Hi, you could refer to this link for solutions. https://github.com/DecaYale/RNNPose/issues/5 I will modify the readme file later.

Yes, your current configuration might be incorrect. The pretrained model should have been loaded with the parameter --pretrained_path. But with the default configuration file we provided, the training will start...

We always set batch_size =1 in our paper. If you want to change the batch size > 1, there are 2 possible solutions: 1. Use DDP (https://pytorch.org/tutorials/intermediate/ddp_tutorial.html) to distribute the...

Sure, you can use pvnet to estimate the initial poses.

Have you solved this? We tested this part, no issue had occurred. Our code can also run with the torch's distributed data parallel without apex. Maybe you just need to...

This might be caused by the update of apex repo. I suggest comment this step and try to install apex manually later. Or just use torch's distributed data parallel to...

Does our provided docker work on your workstation? Of course, manually adapting the code should also be feasible.

Did you work in the provided docker environment? You can bypass the torch.multiprocessing by modifying our training code. Or just directly cal the function eval() could also works. https://github.com/DecaYale/RNNPose/blob/main/tools/eval.py#:~:text=def%20eval-,(,-config_path%2C I...

Thank you for your interest in our work. I am sorry to hear about the issue you encountered. We tested our training at the time we released the code, and...