Adding the quantized namespace to torch.nn and importing it from torch (pytorch#21600)

Summary:
Stack:
    :black_circle:  **pytorch#21600 Adding the quantized namespace to torch**  [:yellow_heart:](https://our.intern.facebook.com/intern/diff/D15742149/)

Add nn.quantized name space to torch
Pull Request resolved: pytorch#21600

Differential Revision: D15742149

Pulled By: zafartahirov

fbshipit-source-id: 60dede12c81861f369d208b06f5b68e9384312f6

Author: zaf

torch/__init__.py

@@ -295,6 +295,7 @@ def manager_path():
 import torch.autograd
 from torch.autograd import no_grad, enable_grad, set_grad_enabled  # noqa: F401
 import torch.nn
+import torch.nn.quantized
 import torch.optim
 import torch.multiprocessing
 import torch.sparse

torch/nn/quantized/__init__.py

@@ -1 +1,12 @@
-from . import functional  # noqa: F401
+r"""
+The quantized module contains the data structures and ops for implementation of
+the quantized models.
+"""
+
+from torch.nn.quantized.functional import add_relu
+from torch.nn.quantized.functional import relu
+
+__all__ = [
+    'add_relu',
+    'relu'
+]