Merge branch 'main' into fix/3855-deprecated-inplace-ops

Author: svekars

beginner_source/blitz/cifar10_tutorial.py

@@ -57,7 +57,7 @@
 """
 import torch
 import torchvision
-import torchvision.transforms as transforms
+from torchvision.transforms import v2
 
 ########################################################################
 # The output of torchvision datasets are PILImage images of range [0, 1].
@@ -69,9 +69,10 @@
 #     BrokenPipeError or RuntimeError related to multiprocessing, try setting
 #     the num_worker of torch.utils.data.DataLoader() to 0.
 
-transform = transforms.Compose(
-    [transforms.ToTensor(),
-     transforms.Normalize((0.5, 0.5, 0.5), (0.5, 0.5, 0.5))])
+transform = v2.Compose([
+    v2.ToImage(),
+    v2.ToDtype(torch.float32, scale=True),
+    v2.Normalize((0.5, 0.5, 0.5), (0.5, 0.5, 0.5))])
 
 batch_size = 4
 
@@ -191,7 +192,7 @@ def forward(self, x):
 ########################################################################
 # Let's quickly save our trained model:
 
-PATH = './cifar_net.pth'
+PATH = './cifar_net.pt'
 torch.save(net.state_dict(), PATH)
 
 ########################################################################
@@ -302,7 +303,7 @@ def forward(self, x):
 # Let's first define our device as the first visible cuda device if we have
 # CUDA available:
 
-device = torch.device('cuda:0' if torch.cuda.is_available() else 'cpu')
+device = torch.device(torch.accelerator.current_accelerator().type if torch.accelerator.is_available() else 'cpu')
 
 # Assuming that we are on a CUDA machine, this should print a CUDA device:
 
@@ -355,9 +356,9 @@ def forward(self, x):
 # -  `Discuss PyTorch on the Forums`_
 # -  `Chat with other users on Slack`_
 #
-# .. _Train a state-of-the-art ResNet network on imagenet: https://github.com/pytorch/examples/tree/master/imagenet
-# .. _Train a face generator using Generative Adversarial Networks: https://github.com/pytorch/examples/tree/master/dcgan
-# .. _Train a word-level language model using Recurrent LSTM networks: https://github.com/pytorch/examples/tree/master/word_language_model
+# .. _Train a state-of-the-art ResNet network on imagenet: https://github.com/pytorch/examples/tree/main/imagenet
+# .. _Train a face generator using Generative Adversarial Networks: https://github.com/pytorch/examples/tree/main/dcgan
+# .. _Train a word-level language model using Recurrent LSTM networks: https://github.com/pytorch/examples/tree/main/word_language_model
 # .. _More examples: https://github.com/pytorch/examples
 # .. _More tutorials: https://github.com/pytorch/tutorials
 # .. _Discuss PyTorch on the Forums: https://discuss.pytorch.org/

beginner_source/examples_tensor/polynomial_tensor.py

@@ -19,15 +19,13 @@
 """
 
 import torch
-import math
-
 
 dtype = torch.float
 device = torch.device("cpu")
 # device = torch.device("cuda:0") # Uncomment this to run on GPU
 
 # Create random input and output data
-x = torch.linspace(-math.pi, math.pi, 2000, device=device, dtype=dtype)
+x = torch.linspace(-torch.pi, torch.pi, 2000, device=device, dtype=dtype)
 y = torch.sin(x)
 
 # Randomly initialize weights

beginner_source/introyt/captumyt.py

@@ -106,16 +106,7 @@
 -  Matplotlib version 3.3.4, since Captum currently uses a Matplotlib
    function whose arguments have been renamed in later versions
 
-To install Captum in an Anaconda or pip virtual environment, use the
-appropriate command for your environment below:
-
-With ``conda``:
-
-.. code-block:: sh
-
-    conda install pytorch torchvision captum flask-compress matplotlib=3.3.4 -c pytorch
-
-With ``pip``:
+To install Captum in a virtual environment, use:
 
 .. code-block:: sh
 

beginner_source/introyt/tensorboardyt_tutorial.py

@@ -24,13 +24,6 @@
 To run this tutorial, you’ll need to install PyTorch, TorchVision,
 Matplotlib, and TensorBoard.
 
-With ``conda``:
-
-.. code-block:: sh
-
-    conda install pytorch torchvision -c pytorch
-    conda install matplotlib tensorboard
-
 With ``pip``:
 
 .. code-block:: sh

beginner_source/understanding_leaf_vs_nonleaf_tutorial.py

@@ -87,11 +87,26 @@
 #    \cdots \cdot
 #    \frac{\partial \mathbf{f}_1}{\partial \mathbf{x}}
 #
-# .. figure:: /_static/img/understanding_leaf_vs_nonleaf/comp-graph-1.png
-#    :alt: Computational graph after forward pass
-#
-#    Computational graph after forward pass
-#
+# .. mermaid::
+#
+#    graph TD
+#
+#        x["x<br/>is_leaf=True<br/>requires_grad=False<br/>retains_grad=False<br/>grad=None"]
+#        W["W<br/>is_leaf=True<br/>requires_grad=True<br/>retains_grad=False<br/>grad=None"]
+#        b["b<br/>is_leaf=True<br/>requires_grad=True<br/>retains_grad=False<br/>grad=None"]
+#        matmul["x @ W"]
+#        z["z = x @ W + b<br/>is_leaf=False<br/>requires_grad=True<br/>retains_grad=False<br/>grad=None"]
+#        relu["y_pred = relu(z)<br/>is_leaf=False<br/>requires_grad=True<br/>retains_grad=False<br/>grad=None"]
+#        y["y<br/>is_leaf=True<br/>requires_grad=False<br/>retains_grad=False<br/>grad=None"]
+#        loss["loss = mse(y_pred, y)<br/>is_leaf=False<br/>requires_grad=True<br/>retains_grad=False<br/>grad=None"]
+#
+#        x --> matmul
+#        W --> matmul
+#        matmul --> z
+#        b --> z
+#        z --> relu
+#        relu --> loss
+#        y --> loss
 # PyTorch considers a node to be a *leaf* if it is not the result of a
 # tensor operation with at least one input having ``requires_grad=True``
 # (e.g. ``x``, ``W``, ``b``, and ``y``), and everything else to be
@@ -260,11 +275,26 @@
 # convention, this attribute will print ``False`` for any leaf node, even
 # if it requires its gradient.
 #
-# .. figure:: /_static/img/understanding_leaf_vs_nonleaf/comp-graph-2.png
-#    :alt: Computational graph after backward pass
-#
-#    Computational graph after backward pass
-#
+# .. mermaid::
+#
+#    graph TD
+#
+#         x["x<br/>is_leaf=True<br/>requires_grad=False<br/>retains_grad=False<br/>grad=None"]
+#         W["W<br/>is_leaf=True<br/>requires_grad=True<br/>retains_grad=False<br/>grad=torch.Tensor"]
+#         b["b<br/>is_leaf=True<br/>requires_grad=True<br/>retains_grad=False<br/>grad=torch.Tensor"]
+#         matmul["x @ W"]
+#         z["z = x @ W + b<br/>is_leaf=False<br/>requires_grad=True<br/>retains_grad=True<br/>grad=torch.Tensor"]
+#         relu["y_pred = relu(z)<br/>is_leaf=False<br/>requires_grad=True<br/>retains_grad=True<br/>grad=torch.Tensor"]
+#         y["y<br/>is_leaf=True<br/>requires_grad=True<br/>retains_grad=False<br/>grad=None"]
+#         loss["loss = mse(y_pred, y)<br/>is_leaf=False<br/>requires_grad=True<br/>retains_grad=True<br/>grad=torch.Tensor"]
+#
+#         x --> matmul
+#         W --> matmul
+#         matmul --> z
+#         b --> z
+#         z --> relu
+#         relu --> loss
+#         y --> loss
 # If you call ``retain_grad()`` on a leaf tensor, it results in a no-op
 # since leaf tensors already retain their gradients by default (when
 # ``requires_grad=True``).

index.rst

@@ -103,6 +103,13 @@ Welcome to PyTorch Tutorials
    :link: intermediate/pinmem_nonblock.html
    :tags: Getting-Started
 
+.. customcarditem::
+   :header: Data Loading Optimization in PyTorch
+   :card_description: Optimize DataLoader configuration with num_workers, pin_memory, persistent_workers for maximum training throughput.
+   :image: _static/img/thumbnails/cropped/generic-pytorch-logo.png
+   :link: intermediate/intermediate_data_loading_tutorial.html
+   :tags: Getting-Started,Best-Practice
+
 .. customcarditem::
    :header: Understanding requires_grad, retain_grad, Leaf, and Non-leaf Tensors
    :card_description: Learn the subtleties of requires_grad, retain_grad, leaf, and non-leaf tensors