PyGRANSO icon indicating copy to clipboard operation
PyGRANSO copied to clipboard
verified publisher icon sun-umn

TODO in future release

Open Buyun-Liang opened this issue 4 years ago • 2 comments

  1. For avoiding unnecessary AD gradients, the GRANSO line search could be modified so that the AD gradient is only computed on demand, i.e., only compute the gradient when sufficient decrease holds

  2. Adding the MPI parallel full BFGS Python code Tim has

Buyun-Liang avatar Dec 24 '21 23:12 Buyun-Liang

  1. gradually switch to the vector_norm and matrix_norm functions

Buyun-Liang avatar Jul 27 '22 01:07 Buyun-Liang

https://pytorch.org/docs/stable/generated/torch.norm.html

Buyun-Liang avatar Jul 27 '22 01:07 Buyun-Liang