get_grad_norm_direct: fix a case of empty norm group (#5148)
fix for [#5145 ](https://github.com/microsoft/DeepSpeed/issues/5145)
empty norm group create a norm tensor with shape=[1], while other norms
will be shapeless. torch.stack does not support such case. Fixing empty
group norm to be shapless as well, instead of shape=[1].
---------
Co-authored-by: Lev Kurilenko <113481193+lekurile@users.noreply.github.com>
Co-authored-by: Lev Kurilenko <lekurile@microsoft.com>
