Add the Latest Features For Basics Autograd Tutorial

beginner_source/basics/autogradqs_tutorial.py

@@ -32,7 +32,7 @@
 y = torch.zeros(3)  # expected output
 w = torch.randn(5, 3, requires_grad=True)
 b = torch.randn(3, requires_grad=True)
-z = torch.matmul(x, w)+b
+z = torch.matmul(x, w) + b
 loss = torch.nn.functional.binary_cross_entropy_with_logits(z, y)
 
 
@@ -133,7 +133,8 @@
 #   - To mark some parameters in your neural network as **frozen parameters**.
 #   - To **speed up computations** when you are only doing forward pass, because computations on tensors that do
 #     not track gradients would be more efficient.
-
+#   For additional reference, you can view the autograd mechanics
+#   documentation:https://docs.pytorch.org/docs/stable/notes/autograd.html#locally-disabling-gradient-computation
 
 ######################################################################
 
@@ -160,6 +161,39 @@
 # - accumulates them in the respective tensor’s ``.grad`` attribute
 # - using the chain rule, propagates all the way to the leaf tensors.
 #
+#   We can also visualize the computational graph by the following 2 methods:
+#
+#   1. TORCH_LOGS="+autograd"
+#    By setting the TORCH_LOGS="+autograd" environment variable, we can enable runtime autograd logs for debugging.
+#
+#    We can perform the logging in the following manner:
+#    TORCH_LOGS="+autograd"  python test.py
+#
+# 2. Torchviz
+# Torchviz is a package to render the computational graph visually.
+#
+# We can generate an image for the computational graph in the example given below:
+#
+#   import torch
+#   from torch import nn
+#   from torchviz import make_dot
+#
+#   model = nn.Sequential(
+#       nn.Linear(8, 16),
+#       nn.ReLU(),
+#       nn.Linear(16, 1)
+#   )
+
+#   x = torch.randn(1, 8, requires_grad=True)
+#   y = model(x).mean()
+
+#   log the internal operations using torchviz
+#   import os
+#   os.environ['TORCH_LOGS'] = "+autograd"
+
+#   dot = make_dot(y, params=dict(model.named_parameters()), show_attrs=True, show_saved=True)
+#   dot.render('simple_graph', format='png')
+#
 # .. note::
 #   **DAGs are dynamic in PyTorch**
 #   An important thing to note is that the graph is recreated from scratch; after each