Enable personalized FC weight_init and sparse_emb weight_init (#31707)
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/31707
Change the initialization value for FC weight init and sparse embedding lookup init.
Previous default initialization is uniform(-\sqrt(1/input_dim), \sqrt(1/input_dim)); Now pass into a flexible hyperparameter, say \alpha into it, to change into uniform(-\sqrt(\alpha/input_dim), \sqrt(\alpha/input_dim));
Reviewed By: chonglinsun
Differential Revision: D18825615
fbshipit-source-id: 4c5f2e07f2b3f5d642fd96d64dbf68892ebeb30b