Zhengyi Wang

Thank you for your interest! The training code for the CRM is quite extensive, and it's difficult to organize it all in a short period. I will start by gradually...

> > Thank you for your interest! The training code for the CRM is quite extensive, and it's difficult to organize it all in a short period. I will start...

We use Zero SNR trick, which means that at maximum timestep the noise to signal ratio will be 1.0/0.0. However, in general the maximum step will not be used. How...

我测试的是torch==1.13.0+cu117。可以换成torch1.x试试?

请问有什么不清晰的地方呢？

My testing environment is torch==1.13.0+cu117. Maybe you can try torch 1.x? I am not sure whether will torch 2.x work.

That sounds interesting:) I think it is likely to work.

如果不报bug的话可以忽视这个，应该是之前整理代码的时候把一些没有用到的文件删了

第一阶段默认是训练250epoch，如果想早点停止可以修改--iter参数

优化的是3D模型的参数