Merge branch 'main' into fix/profiler-recipe-format

Author: AlannaBurke

.ci/docker/requirements.txt

@@ -14,7 +14,7 @@ tqdm==4.66.1
 numpy==1.24.4
 matplotlib
 librosa
-torch==2.6
+torch==2.7
 torchvision
 torchdata
 networkx
@@ -36,7 +36,7 @@ datasets
 transformers
 torchmultimodal-nightly # needs to be updated to stable as soon as it's avaialable
 onnx
-onnxscript
+onnxscript>=0.2.2
 onnxruntime
 evaluate
 accelerate>=0.20.1
@@ -67,7 +67,7 @@ iopath
 pygame==2.6.0
 pycocotools
 semilearn==0.3.2
-torchao==0.5.0
+torchao==0.10.0
 segment_anything==1.0
-torchrec==1.0.0; platform_system == "Linux"
-fbgemm-gpu==1.1.0; platform_system == "Linux"
+torchrec==1.1.0; platform_system == "Linux"
+fbgemm-gpu==1.2.0; platform_system == "Linux"

.github/scripts/docathon-label-sync.py

@@ -25,11 +25,11 @@ def main():
     issue_number = int(re.findall(r'#(\d{1,5})', pull_request_body)[0])
     issue = repo.get_issue(issue_number)
     issue_labels = issue.labels
-    docathon_label_present = any(label.name == 'docathon-h1-2024' for label in issue_labels)
+    docathon_label_present = any(label.name == 'docathon-h1-2025' for label in issue_labels)
 
     # if the issue has a docathon label, add all labels from the issue to the PR.
     if not docathon_label_present:
-        print("The 'docathon-h1-2024' label is not present in the issue.")
+        print("The 'docathon-h1-2025' label is not present in the issue.")
         return    
     pull_request_labels = pull_request.get_labels()
     issue_label_names = [label.name for label in issue_labels]

.github/workflows/docathon-assign.yml

@@ -28,14 +28,14 @@ jobs:
                   repo: context.repo.repo,
                   issue_number: issueNumber
                 });
-              const hasLabel = issue.labels.some(label => label.name === 'docathon-h1-2024');
+              const hasLabel = issue.labels.some(label => label.name === 'docathon-h1-2025');
               if (hasLabel) {
                 if (issue.assignee !== null) {
                   await github.rest.issues.createComment({
                     owner: context.repo.owner,
                     repo: context.repo.repo,
                     issue_number: issueNumber,
-                    body: "The issue is already assigned. Please pick an opened and unnasigned issue with the [docathon-h1-2024 label](https://github.com/pytorch/pytorch/issues?q=is%3Aopen+is%3Aissue+label%3Adocathon-h1-2024)."
+                    body: "The issue is already assigned. Please pick an opened and unnasigned issue with the [docathon-h1-2025 label](https://github.com/pytorch/pytorch/issues?q=is%3Aopen+is%3Aissue+label%3Adocathon-h1-2025)."
                   });
                 } else {
                   await github.rest.issues.addAssignees({
@@ -46,7 +46,7 @@ jobs:
                   });
                 }
               } else {
-                const commmentMessage = "This issue does not have the correct label. Please pick an opened and unnasigned issue with the [docathon-h1-2024 label](https://github.com/pytorch/pytorch/issues?q=is%3Aopen+is%3Aissue+label%3Adocathon-h1-2024)."
+                const commmentMessage = "This issue does not have the correct label. Please pick an opened and unnasigned issue with the [docathon-h1-2025 label](https://github.com/pytorch/pytorch/issues?q=is%3Aopen+is%3Aissue+label%3Adocathon-h1-2025)."
                 await github.rest.issues.createComment({
                   owner: context.repo.owner,
                   repo: context.repo.repo,

.jenkins/build.sh

@@ -22,11 +22,10 @@ sudo apt-get install -y pandoc
 #Install PyTorch Nightly for test.
 # Nightly - pip install --pre torch torchvision torchaudio -f https://download.pytorch.org/whl/nightly/cu102/torch_nightly.html
 # Install 2.5 to merge all 2.4 PRs - uncomment to install nightly binaries (update the version as needed).
-# sudo pip uninstall -y torch torchvision torchaudio torchtext torchdata
-# sudo pip3 install torch==2.6.0 torchvision --no-cache-dir --index-url https://download.pytorch.org/whl/test/cu124
 # sudo pip uninstall -y fbgemm-gpu torchrec
+# sudo pip uninstall -y torch torchvision torchaudio torchtext torchdata torchrl tensordict
 # sudo pip3 install fbgemm-gpu==1.1.0 torchrec==1.0.0 --no-cache-dir --index-url https://download.pytorch.org/whl/test/cu124
-
+# pip3 install torch==2.7.0 torchvision torchaudio --no-cache-dir --index-url https://download.pytorch.org/whl/test/cu126
 # Install two language tokenizers for Translation with TorchText tutorial
 python -m spacy download en_core_web_sm
 python -m spacy download de_core_news_sm

.jenkins/validate_tutorials_built.py

@@ -31,6 +31,7 @@
     "prototype_source/vmap_recipe",
     "prototype_source/torchscript_freezing",
     "prototype_source/nestedtensor",
+    "prototype_source/gpu_direct_storage", # requires specific filesystem + GPUDirect Storage to be set up
     "recipes_source/recipes/saving_and_loading_models_for_inference",
     "recipes_source/recipes/saving_multiple_models_in_one_file",
     "recipes_source/recipes/tensorboard_with_pytorch",
@@ -51,7 +52,7 @@
     "intermediate_source/text_to_speech_with_torchaudio",
     "intermediate_source/tensorboard_profiler_tutorial", # reenable after 2.0 release.
     "advanced_source/semi_structured_sparse", # reenable after 3303 is fixed.
-    "recipes_source/recipes/reasoning_about_shapes"
+    "intermediate_source/torchrec_intro_tutorial", # reenable after 3302 is fixe
 ]
 
 def tutorial_source_dirs() -> List[Path]:

README.md

@@ -20,7 +20,7 @@ You can submit the following types of issues:
 
 We use sphinx-gallery's [notebook styled examples](https://sphinx-gallery.github.io/stable/tutorials/index.html) to create the tutorials. Syntax is very simple. In essence, you write a slightly well formatted Python file and it shows up as an HTML page. In addition, a Jupyter notebook is autogenerated and available to run in Google Colab.
 
-Here is how you can ceate a new tutorial (for a detailed description, see [CONTRIBUTING.md](./CONTRIBUTING.md)):
+Here is how you can create a new tutorial (for a detailed description, see [CONTRIBUTING.md](./CONTRIBUTING.md)):
 
 NOTE: Before submitting a new tutorial, read [PyTorch Tutorial Submission Policy](./tutorial_submission_policy.md).
 

_static/img/distributed/fsdp_implicit.png

@@ -211,14 +211,5 @@
 
 <img height="1" width="1" style="border-style:none;" alt="" src="https://www.googleadservices.com/pagead/conversion/795629140/?label=txkmCPmdtosBENSssfsC&amp;guid=ON&amp;script=0"/>
 
-<script>
-    //temporarily add a link to survey
-    var survey = '<div class="survey-banner"><p><i class="fas fa-poll" aria-hidden="true">&nbsp </i> Take the <a href="https://forms.gle/KZ4xGL65VRMYNbbG6">PyTorch Docs/Tutorials survey</a>.</p></div>'
-    if ($(".pytorch-call-to-action-links").length) {
-        $(".pytorch-call-to-action-links").before(survey);
-    } else {
-        $("#pytorch-article").prepend(survey);
-    }
-</script>
 
 {% endblock %}

_static/img/install_msvc.png

@@ -134,10 +134,18 @@ def tokenize(self, path):
 # -----------------------------
 #
 # This is a tutorial on dynamic quantization, a quantization technique
-# that is applied after a model has been trained. Therefore, we'll simply load some
-# pretrained weights into this model architecture; these weights were obtained
-# by training for five epochs using the default settings in the word language model
-# example.
+# that is applied after a model has been trained. Therefore, we'll simply
+# load some pretrained weights into this model architecture; these
+# weights were obtained by training for five epochs using the default
+# settings in the word language model example.
+#
+# Before running this tutorial, download the required pre-trained model:
+#
+# .. code-block:: bash
+#
+#     wget https://s3.amazonaws.com/pytorch-tutorial-assets/word_language_model_quantize.pth
+#
+# Place the downloaded file in the data directory or update the model_data_filepath accordingly.
 
 ntokens = len(corpus.dictionary)
 

_static/img/pt2e_quant_xpu_inductor.png

@@ -13,7 +13,7 @@ Learn the Basics
     Tensors
     https://pytorch.org/tutorials/beginner/basics/tensor_tutorial.html
 
-4. dataquickstart_tutorial.py
+4. data_tutorial.py
     Datasets & DataLoaders
     https://pytorch.org/tutorials/beginner/basics/data_tutorial.html
 

_templates/layout.html

@@ -120,7 +120,7 @@
 
 import os
 import pandas as pd
-from torchvision.io import read_image
+from torchvision.io import decode_image
 
 class CustomImageDataset(Dataset):
     def __init__(self, annotations_file, img_dir, transform=None, target_transform=None):
@@ -134,7 +134,7 @@ def __len__(self):
 
     def __getitem__(self, idx):
         img_path = os.path.join(self.img_dir, self.img_labels.iloc[idx, 0])
-        image = read_image(img_path)
+        image = decode_image(img_path)
         label = self.img_labels.iloc[idx, 1]
         if self.transform:
             image = self.transform(image)
@@ -184,7 +184,7 @@ def __len__(self):
 # ^^^^^^^^^^^^^^^^^^^^
 #
 # The __getitem__ function loads and returns a sample from the dataset at the given index ``idx``.
-# Based on the index, it identifies the image's location on disk, converts that to a tensor using ``read_image``, retrieves the
+# Based on the index, it identifies the image's location on disk, converts that to a tensor using ``decode_image``, retrieves the
 # corresponding label from the csv data in ``self.img_labels``, calls the transform functions on them (if applicable), and returns the
 # tensor image and corresponding label in a tuple.
 

advanced_source/dynamic_quantization_tutorial.py

@@ -76,7 +76,7 @@ def forward(self, x):
 # (`read more <https://pytorch.org/tutorials/beginner/hyperparameter_tuning_tutorial.html>`__ about hyperparameter tuning)
 #
 # We define the following hyperparameters for training:
-#  - **Number of Epochs** - the number times to iterate over the dataset
+#  - **Number of Epochs** - the number of times to iterate over the dataset
 #  - **Batch Size** - the number of data samples propagated through the network before the parameters are updated
 #  - **Learning Rate** - how much to update models parameters at each batch/epoch. Smaller values yield slow learning speed, while large values may result in unpredictable behavior during training.
 #

beginner_source/basics/README.txt

@@ -11,7 +11,7 @@ PyTorch Version in Google Colab
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
 Wen you are running a tutorial that requires a version of PyTorch that has
-jst been released, that version might not be yet available in Google Colab.
+just been released, that version might not be yet available in Google Colab.
 To check that you have the required ``torch`` and compatible domain libraries
 installed, run ``!pip list``.
 
@@ -27,7 +27,7 @@ Using Tutorial Data from Google Drive in Colab
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
 We've added a new feature to tutorials that allows users to open the
-ntebook associated with a tutorial in Google Colab. You may need to
+notebook associated with a tutorial in Google Colab. You may need to
 copy data to your Google drive account to get the more complex tutorials
 to work.
 

beginner_source/basics/data_tutorial.py

@@ -1,4 +1,3 @@
-# -*- coding: utf-8 -*-
 """
 PyTorch: Tensors and autograd
 -------------------------------

beginner_source/basics/optimization_tutorial.py

@@ -1,4 +1,3 @@
-# -*- coding: utf-8 -*-
 """
 PyTorch: Defining New autograd Functions
 ----------------------------------------
@@ -34,8 +33,11 @@ def forward(ctx, input):
         """
         In the forward pass we receive a Tensor containing the input and return
         a Tensor containing the output. ctx is a context object that can be used
-        to stash information for backward computation. You can cache arbitrary
-        objects for use in the backward pass using the ctx.save_for_backward method.
+        to stash information for backward computation. You can cache tensors for
+        use in the backward pass using the ``ctx.save_for_backward`` method. Other
+        objects can be stored directly as attributes on the ctx object, such as
+        ``ctx.my_object = my_object``. Check out `Extending torch.autograd <https://docs.pytorch.org/docs/stable/notes/extending.html#extending-torch-autograd>`_
+        for further details.
         """
         ctx.save_for_backward(input)
         return 0.5 * (5 * input ** 3 - 3 * input)

beginner_source/colab.rst

@@ -303,22 +303,21 @@ def num_flat_features(self, x):
 #    The values passed to the transform are the means (first tuple) and the
 #    standard deviations (second tuple) of the rgb values of the images in
 #    the dataset. You can calculate these values yourself by running these
-#    few lines of code:
-#          ```
-#           from torch.utils.data import ConcatDataset
-#           transform = transforms.Compose([transforms.ToTensor()])
-#           trainset = torchvision.datasets.CIFAR10(root='./data', train=True,
+#    few lines of code::
+#
+#        from torch.utils.data import ConcatDataset
+#        transform = transforms.Compose([transforms.ToTensor()])
+#        trainset = torchvision.datasets.CIFAR10(root='./data', train=True,
 #                                        download=True, transform=transform)
 #
-#           #stack all train images together into a tensor of shape 
-#           #(50000, 3, 32, 32)
-#           x = torch.stack([sample[0] for sample in ConcatDataset([trainset])])
+#        # stack all train images together into a tensor of shape 
+#        # (50000, 3, 32, 32)
+#        x = torch.stack([sample[0] for sample in ConcatDataset([trainset])])
 #           
-#           #get the mean of each channel            
-#           mean = torch.mean(x, dim=(0,2,3)) #tensor([0.4914, 0.4822, 0.4465])
-#           std = torch.std(x, dim=(0,2,3)) #tensor([0.2470, 0.2435, 0.2616])  
-# 
-#          ```   
+#        # get the mean of each channel            
+#        mean = torch.mean(x, dim=(0,2,3)) # tensor([0.4914, 0.4822, 0.4465])
+#        std = torch.std(x, dim=(0,2,3)) # tensor([0.2470, 0.2435, 0.2616])  
+#    
 # 
 # There are many more transforms available, including cropping, centering,
 # rotation, and reflection.

beginner_source/examples_autograd/polynomial_autograd.py

@@ -33,8 +33,6 @@
 sys.path.insert(0, os.path.abspath('./.jenkins'))
 import pytorch_sphinx_theme
 import torch
-import numpy
-import gc
 import glob
 import random
 import shutil
@@ -49,6 +47,46 @@
 pio.renderers.default = 'sphinx_gallery'
 
 
+import sphinx_gallery.gen_rst
+import multiprocessing
+
+# Monkey patch sphinx gallery to run each example in an isolated process so that
+# we don't need to worry about examples changing global state.
+#
+# Alt option 1: Parallelism was added to sphinx gallery (a later version that we
+# are not using yet) using joblib, but it seems to result in errors for us, and
+# it has no effect if you set parallel = 1 (it will not put each file run into
+# its own process and run singly) so you need parallel >= 2, and there may be
+# tutorials that cannot be run in parallel.
+#
+# Alt option 2: Run sphinx gallery once per file (similar to how we shard in CI
+# but with shard sizes of 1), but running sphinx gallery for each file has a
+# ~5min overhead, resulting in the entire suite taking ~2x time
+def call_fn(func, args, kwargs, result_queue):
+    try:
+        result = func(*args, **kwargs)
+        result_queue.put((True, result))
+    except Exception as e:
+        result_queue.put((False, str(e)))
+
+def call_in_subprocess(func):
+    def wrapper(*args, **kwargs):
+        result_queue = multiprocessing.Queue()
+        p = multiprocessing.Process(
+            target=call_fn,
+            args=(func, args, kwargs, result_queue)
+        )
+        p.start()
+        p.join()
+        success, result = result_queue.get()
+        if success:
+            return result
+        else:
+            raise RuntimeError(f"Error in subprocess: {result}")
+    return wrapper
+
+sphinx_gallery.gen_rst.generate_file_rst = call_in_subprocess(sphinx_gallery.gen_rst.generate_file_rst)
+
 try:
     import torchvision
 except ImportError:
@@ -97,14 +135,6 @@
 
 # -- Sphinx-gallery configuration --------------------------------------------
 
-def reset_seeds(gallery_conf, fname):
-    torch.cuda.empty_cache()
-    torch.manual_seed(42)
-    torch.set_default_device(None)
-    random.seed(10)
-    numpy.random.seed(10)
-    gc.collect()
-
 sphinx_gallery_conf = {
     'examples_dirs': ['beginner_source', 'intermediate_source',
                       'advanced_source', 'recipes_source', 'prototype_source'],
@@ -115,7 +145,6 @@ def reset_seeds(gallery_conf, fname):
     'first_notebook_cell': ("# For tips on running notebooks in Google Colab, see\n"
                             "# https://pytorch.org/tutorials/beginner/colab\n"
                             "%matplotlib inline"),
-    'reset_modules': (reset_seeds),
     'ignore_pattern': r'_torch_export_nightly_tutorial.py',
     'pypandoc': {'extra_args': ['--mathjax', '--toc'],
                  'filters': ['.jenkins/custom_pandoc_filter.py'],

beginner_source/examples_autograd/polynomial_custom_function.py

@@ -698,3 +698,14 @@ TorchServe
 Inductor’s
 onwards
 recompilations
+BiasCorrection
+ELU
+GELU
+NNCF
+OpenVINO
+OpenVINOQuantizer
+PReLU
+Quantizer
+SmoothQuant
+quantizer
+quantizers