YoungSeng

RuntimeError: one of the variables needed for gradient computation has been modified by an inplace operation

1

Hello, thank you for your geart work! Here is a question. When I set skeleton_aware=0, ``` 0%| | 0/15000 [00:01

音频维度问题

1

关于DiffuseStyleGesture+提取出的特征的维度，提取音频特征的维度为什么要这样设置：40+64+2+2+1024+1 为什么MFCC是40，log-mel是64，韵律特征是4等等，这样设置有什么特别的用意吗，为什么要这样取特征的维度

2个疑问

1

1）你放出的模型是只在ZEGGS上训练的吗，step为450000时候的模型，我按照指示重新训了似乎结果不及你公开的那个模型效果 2）关于diffusion实时加速师弟你有好建议吗？（我试了什么都不动，PLMS 和DDIM 都还是慢）

Transition between generated gesture

2

The segments we trained are all 4s long, and it is difficult to generalize to arbitrary length gestures by positional encoding alone. [MDM](https://github.com/GuyTevet/motion-diffusion-model)-based models that require time-awareness (arbitrarily long inference)...