Docs: build with Sphinx 5 (#70309)

Fixes #60979. Also see #61045 and https://github.com/sphinx-doc/sphinx/issues/9395 for discussion.

I _believe_ the reason that we were previously pinning to Sphinx 3 was because of issues with pytorch_sphinx_theme and Sphinx 4 support, but these seem to have been resolved now. See https://torchgeo.readthedocs.io/ for an example of docs built with pytorch_sphinx_theme and Sphinx 4.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/70309
Approved by: https://github.com/albanD

docs/requirements.txt

@@ -1,6 +1,4 @@
-sphinx==3.5.4
-Jinja2==3.0.*
-docutils==0.16
+sphinx==5.0.0
 -e git+https://github.com/pytorch/pytorch_sphinx_theme.git#egg=pytorch_sphinx_theme
 sphinxcontrib.katex
 matplotlib

docs/source/autograd.rst

@@ -72,16 +72,8 @@ Locally disabling gradient computation
 
 See :ref:`locally-disable-grad-doc` for more information on the differences
 between no-grad and inference mode as well as other related mechanisms that
-may be confused with the two.
-
-.. autosummary::
-    :toctree: generated
-    :nosignatures:
-
-    no_grad
-    enable_grad
-    set_grad_enabled
-    inference_mode
+may be confused with the two. Also see :ref:`torch-rst-local-disable-grad`
+for a list of functions that can be used to locally disable gradients.
 
 .. _default-grad-layouts:
 

docs/source/conf.py

@@ -363,7 +363,7 @@ if RELEASE:
 #
 # This is also used if you do content translation via gettext catalogs.
 # Usually you set "language" from the command line for these cases.
-language = None
+language = "en"
 
 # List of patterns, relative to source directory, that match files and
 # directories to ignore when looking for source files.

docs/source/torch.rst

@@ -224,6 +224,8 @@ Parallelism
     get_num_interop_threads
     set_num_interop_threads
 
+.. _torch-rst-local-disable-grad:
+
 Locally disabling gradient computation
 --------------------------------------
 The context managers :func:`torch.no_grad`, :func:`torch.enable_grad`, and