From 0f51ae244400eea693184ab5be27beba091e57d8 Mon Sep 17 00:00:00 2001 From: Nicolas Hug Date: Fri, 8 Oct 2021 10:50:46 +0100 Subject: [PATCH 1/3] Added docs for reference scripts --- docs/source/index.rst | 1 + docs/source/models.rst | 2 ++ 2 files changed, 3 insertions(+) diff --git a/docs/source/index.rst b/docs/source/index.rst index 3e02cd34ad4..2079285bfd4 100644 --- a/docs/source/index.rst +++ b/docs/source/index.rst @@ -38,6 +38,7 @@ architectures, and common image transformations for computer vision. ops transforms utils + training_references .. toctree:: :maxdepth: 1 diff --git a/docs/source/models.rst b/docs/source/models.rst index 4ac582a33a2..9d05b509899 100644 --- a/docs/source/models.rst +++ b/docs/source/models.rst @@ -1,3 +1,5 @@ +.. _models: + torchvision.models ################## From e89ff7dcddc8ecf63b94c103d2021ca7a1c20bc1 Mon Sep 17 00:00:00 2001 From: Nicolas Hug Date: Fri, 8 Oct 2021 10:51:12 +0100 Subject: [PATCH 2/3] better if the file is included too --- docs/source/training_references.rst | 23 +++++++++++++++++++++++ 1 file changed, 23 insertions(+) create mode 100644 docs/source/training_references.rst diff --git a/docs/source/training_references.rst b/docs/source/training_references.rst new file mode 100644 index 00000000000..a847c372518 --- /dev/null +++ b/docs/source/training_references.rst @@ -0,0 +1,23 @@ +Training references +=================== + +On top of the many models, datasets, and image transforms, Torchvision also +provides training reference scripts. These are the scripts that we use to train +the :ref:`models ` which are then available with pre-trained weights. + +These scripts are not part of the core package and are instead available `on +GitHub `_. + +In general, these scripts rely on the latest (not yet released) pytorch version +or the latest torchvision version. This means that to use them, **you might need +to install the latest pytorch and torchvision versions**, with e.g.:: + + conda install pytorch torchvision -c pytorch-nightly + +If you need to rely on an older stable version of pytorch or torchvision, e.g. +torchvision 0.10, then it's safer to use the scripts from that corresponding +release on GitHub, namely +https://github.com/pytorch/vision/tree/v0.10.0/references. + +Please also note that while these scripts are largely stable, they do not have +backward compatibility guarantees. From d2203c0308166d5f12d205775d36c923487a9972 Mon Sep 17 00:00:00 2001 From: Nicolas Hug Date: Fri, 8 Oct 2021 11:10:35 +0100 Subject: [PATCH 3/3] Addressed comments --- docs/source/index.rst | 4 ++-- docs/source/training_references.rst | 14 ++++++++++---- 2 files changed, 12 insertions(+), 6 deletions(-) diff --git a/docs/source/index.rst b/docs/source/index.rst index 2079285bfd4..d96086704c3 100644 --- a/docs/source/index.rst +++ b/docs/source/index.rst @@ -38,13 +38,13 @@ architectures, and common image transformations for computer vision. ops transforms utils - training_references .. toctree:: :maxdepth: 1 - :caption: Examples + :caption: Examples and training references auto_examples/index + training_references .. automodule:: torchvision :members: diff --git a/docs/source/training_references.rst b/docs/source/training_references.rst index a847c372518..fc22ac5eba6 100644 --- a/docs/source/training_references.rst +++ b/docs/source/training_references.rst @@ -6,7 +6,16 @@ provides training reference scripts. These are the scripts that we use to train the :ref:`models ` which are then available with pre-trained weights. These scripts are not part of the core package and are instead available `on -GitHub `_. +GitHub `_. We currently +provide references for +`classification `_, +`detection `_, +`segmentation `_, +`similarity learning `_, +and `video classification `_. + +While these scripts are largely stable, they do not offer backward compatibility +guarantees. In general, these scripts rely on the latest (not yet released) pytorch version or the latest torchvision version. This means that to use them, **you might need @@ -18,6 +27,3 @@ If you need to rely on an older stable version of pytorch or torchvision, e.g. torchvision 0.10, then it's safer to use the scripts from that corresponding release on GitHub, namely https://github.com/pytorch/vision/tree/v0.10.0/references. - -Please also note that while these scripts are largely stable, they do not have -backward compatibility guarantees.