📚 The doc issue

The GroupNorm doc says

... The standard-deviation is calculated via the biased estimator, equivalent to torch.var(input, unbiased=False).

It should be variance, rather than standard-deviation

Suggest a potential alternative/fix

The standard-deviation variance Var[x] is calculated via the biased estimator, equivalent to torch.var(input, unbiased=False).

cc @svekars @brycebortree @sekyondaMeta @AlannaBurke @albanD @mruberry @jbschlosser @walterddr @mikaylagawarecki