flan

Maybe I should train another 500k iters, Thanks for your advice, I wonder what is the sampling correctness loss when it converges in your experiment ant the value?

Another interesting thing is, if I use bilinear sampling in the sample correctness loss, the loss even cannot converge

any solutions?

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 generate...

@Yuhuoo Yes, you may comment the try/catch to check where the error exactly occurs

I have the same question. @Kai-46 Could you give some insight on this question? > Hi, I have read this paper and some questions appear. I agree with this paragraph...

我理解的论文中所讲，是每种舞蹈的动作的采样频率都不一样，不是25FPS的固定值。 所以论文中有讲到alignment的问题。 我觉得config.json中start和end都是有用的，因为需要根据总的帧数对动作进行插值？

music的length=end-start, motion的length是json中的"length"，两个length是不一样的，需要我们再Align一遍？

@YuliangXiu