[Doc][NVTX] Add documentation for nvtx.range (pytorch#121699)

The context manager `torch.cuda.nvtx.range` has been around for about 4 years (see pytorch#42925). Unfortunately, it was never documented and as a consequence users are just unaware of it (see pytorch#121663).

Pull Request resolved: pytorch#121699
Approved by: https://github.com/janeyx99

docs/source/cuda.rst

@@ -133,6 +133,7 @@ NVIDIA Tools Extension (NVTX)
     nvtx.mark
     nvtx.range_push
     nvtx.range_pop
+    nvtx.range
 
 Jiterator (beta)
 -----------------------------
@@ -167,4 +168,4 @@ See the :doc:`documentation <cuda._sanitizer>` for information on how to use it.
 .. py:module:: torch.cuda.profiler
 .. py:module:: torch.cuda.random
 .. py:module:: torch.cuda.sparse
-.. py:module:: torch.cuda.streams
+.. py:module:: torch.cuda.streams

torch/jit/supported_ops.py

@@ -276,7 +276,9 @@ def _get_global_builtins():
         if len(schemas) > 0:
             schematized_ops.append("")
         else:
-            table_row = f'":any:`{fn}`", "{schemaless_op_explanations[fn]}"'
+            table_row = (
+                f'":external+python:py:obj:`{fn}`", "{schemaless_op_explanations[fn]}"'
+            )
             schemaless_ops.append(table_row)
 
     schematized_ops_str = "\n".join(schematized_ops)