Yu Sun
Yu Sun
Yes, but you can also directly train from our pre-trained model with v1_hrnet_3dpw_ft.yml to have protocol 3, which is just train with 3DPW.
\* if you choose to use more datasets during training than [EFT](https://arxiv.org/pdf/2004.03686.pdf)
In this case, it fails in detection. I am solving this problem by fine-tuning ROMP on detection datasets with small subjects. For now, you can crop the image into multiple...
Yes, it is all about the scale of people in images. You may get the 3D mesh results of people with enough scale. It is complicate to get the rendering...
@ayedaemon Sorry for accidently closing this issue. I am working on this problem too. I can't promise the exact time of releasing the related code. But the problem does allievated...
No, we use the center map for sampling, sorry about that bug. Please set `match_preds_to_gts_for_supervision: True` during evaluation to make use of predicted center.
I see. https://github.com/yanch2116/CharacterDriven-BlenderAddon 我同事的这个是你需要的。
Thanks for your interests! I have just commit a version that support CPU. You just have to change the `GPU: 0` to `GPU: -1` in config files (like src/configs/basic_test.yml or...
There are several things need to be noticed. 1. I have just upload the commit. So you have to 'git pull' to drag the update. Or please simply reload the...
Sorry for the bug. I have just try the Colab. It turns out that the pyrender need GPU for visualization. So current code could run on CPU and save the...