Ao Gao

> I meet the same problem and I check the predicted sdf, with values all 0, which means that all vertices are outside mesh, thus the threshold 0.5 fails to...

> @Yuhuoo Yes, you may comment the try/catch to check where the error exactly occurs Thank you for your reply, but I still don't kown how to do. Can you...

> egl community I also want to know the answer fixing the bug

> Hi, > > You can use meshlab to convert all the meshes to watertight meshes. For those noises around the surface, maybe you can also try to use the...

You can change the version of scikit-image to 0.18.3.

多卡训练COCO数据集是可以的，单卡训练pascal voc也是可以的，多卡训练pascal voc就出错了，没有改动代码

> 看不出错误来源，你是用的官方pascal voc吗？能提供最小复现错误代码吗？ 是用的官方的pascal voc数据集，读取数据的时候会有数据缺失跳过的log，不知道和这个错误会不会有关系。 最小复现错误代码是指完整的错误log吗？

I find there are may some problems in the preprocess scripts. I use the scripts preprocessing the THuman2.0 dataset, but the preprocessed data cannot aligned to the smplx mesh. ...

> Hi @Zhangjzh and @Yuhuoo > > I have corrected some bugs and updated the scripts for training data generation [dataset.md](https://github.com/YuliangXiu/ICON/blob/master/docs/dataset.md). > > Please re-download the `SMPL-X.zip`, and re-run the...