From 40e80336127fdd52133fec4815529f6395af0d5e Mon Sep 17 00:00:00 2001 From: Mike Ruberry Date: Wed, 3 Feb 2021 06:43:33 +0000 Subject: [PATCH] 1.8 linalg doc updates --- docs/source/linalg.rst | 4 ++++ torch/linalg/__init__.py | 5 +++-- 2 files changed, 7 insertions(+), 2 deletions(-) diff --git a/docs/source/linalg.rst b/docs/source/linalg.rst index f592eac72aea..668632db76a9 100644 --- a/docs/source/linalg.rst +++ b/docs/source/linalg.rst @@ -6,6 +6,10 @@ torch.linalg Common linear algebra operations. +This module is in BETA. New functions are still being added, and some +functions may change in future PyTorch releases. See the documentation of each +function for details. + .. automodule:: torch.linalg .. currentmodule:: torch.linalg diff --git a/torch/linalg/__init__.py b/torch/linalg/__init__.py index 44139c574a7d..e54c1d765767 100644 --- a/torch/linalg/__init__.py +++ b/torch/linalg/__init__.py @@ -523,7 +523,8 @@ always be real-valued, even if :attr:`input` is complex. .. note:: Unlike NumPy's ``linalg.svd``, this always returns a namedtuple of - three tensors, even when :attr:`compute_uv=False`. + three tensors, even when :attr:`compute_uv=False`. This behavior may + change in a future PyTorch release. .. note:: The singular values are returned in descending order. If :attr:`input` is a batch of matrices, then the singular values of each matrix in the batch is returned in descending order. @@ -914,7 +915,7 @@ * unlike ``numpy.linalg.qr``, this function always returns a tuple of two tensors. When ``mode='r'``, the `Q` tensor is an - empty tensor. + empty tensor. This behavior may change in a future PyTorch release. .. note:: Backpropagation is not supported for ``mode='r'``. Use ``mode='reduced'`` instead.