✨[Converter] Implement aten::layer_norm, aten::_softmax

[apbose]

Torch op:
layer_norm(Tensor input, SymInt[] normalized_shape, Tensor? weight=None, Tensor? bias=None, float eps=1e-05, bool cudnn_enable=True) -> Tensor
softmax.int(Tensor self, int dim, ScalarType? dtype=None) -> Tensor
Aten op : torch.ops.aten.layer_norm.default
torch.ops.aten._softmax.default

[github-actions]

This issue has not seen activity for 90 days, Remove stale label or comment or this will be closed in 10 days