This overload of add_ is deprecated.
add_(Number alpha, Tensor other)
Consider using one of the following signatures instead:
add_(Tensor other, *, Number alpha) (Triggered internally at /pytorch/torch/csrc/utils/python_arg_parser.cpp:882.)
buf.mul_(momentum).add_(actual_lr, d_p + weight_decay * p.data)
Line 109 in lars.py has a UserWarning which says
I'm using PyTorch 1.7.0 on CUDA 11.0.