pytorch · ssnl · Apr 12, 2019 · Apr 12, 2019 · Apr 12, 2019 · Apr 12, 2019
diff --git a/docs/source/data.rst b/docs/source/data.rst
diff --git a/docs/source/notes/cuda.rst b/docs/source/notes/cuda.rst
@@ -262,6 +262,7 @@ also preserve :class:`torch.device` and :class:`torch.dtype` of a Tensor).
     y_cpu = torch.ones_like(x_cpu)
     y_gpu = torch.zeros_like(x_gpu)
 
+.. _cuda-memory-pinning:
 
 Use pinned memory buffers
 ^^^^^^^^^^^^^^^^^^^^^^^^^

diff --git a/docs/source/notes/multiprocessing.rst b/docs/source/notes/multiprocessing.rst
@@ -1,3 +1,5 @@
+.. _multiprocessing-best-practices:
+
 Multiprocessing best practices
 ==============================
 
@@ -20,6 +22,8 @@ memory and will only send a handle to another process.
 This allows to implement various training methods, like Hogwild, A3C, or any
 others that require asynchronous operation.
 
+.. _multiprocessing-cuda-note:
+
 CUDA in multiprocessing
 -----------------------