What is the convergence or finish loop condition?

[anicej]

In train_util in run _loop function there is a loop:

while ( not self.lr_anneal_steps or self.step + self.resume_step < self.lr_anneal_steps ): This code appears to be similar to a while (True) loop because the variable lr_anneal_steps is always set to 0 and does not change within the loop. Consequently, the loop will continue indefinitely without any variation. If my understanding is incorrect, please inform me once the loop has finished executing.

[Mally-cj]

I also think it never stop