2024-05-07 nightly release (4142d3f)

Author: pytorchbot

docsrc/conf.py

@@ -184,6 +184,8 @@
     "ConverterImplSignature": "ConverterImplSignature",
 }
 
+nbsphinx_execute = "never"
+
 # -- A patch that prevents Sphinx from cross-referencing ivar tags -------
 # See http://stackoverflow.com/a/41184353/3343043
 

docsrc/index.rst

@@ -111,6 +111,7 @@ Tutorials
    tutorials/_rendered_examples/dynamo/torch_compile_transformers_example
    tutorials/_rendered_examples/dynamo/torch_compile_advanced_usage
    tutorials/_rendered_examples/dynamo/torch_compile_stable_diffusion
+   tutorials/_rendered_examples/dynamo/custom_kernel_plugins
 
 Python API Documenation
 ------------------------

docsrc/requirements.txt

@@ -1,8 +1,8 @@
-sphinx==4.5.0
+sphinx==5.0.1
 sphinx-gallery==0.13.0
-breathe==4.33.1
-exhale==0.3.1
+breathe==4.34.0
+exhale==0.3.7
 -e git+https://github.com/pytorch/pytorch_sphinx_theme.git#egg=pytorch_sphinx_theme
-nbsphinx==0.8.8
+nbsphinx==0.9.3
 docutils==0.17.1
 pillow

docsrc/sg_execution_times.rst

@@ -0,0 +1,46 @@
+
+:orphan:
+
+.. _sphx_glr_sg_execution_times:
+
+
+Computation times
+=================
+**00:00.000** total execution time for 4 files **from all galleries**:
+
+.. container::
+
+  .. raw:: html
+
+    <style scoped>
+    <link href="https://cdnjs.cloudflare.com/ajax/libs/twitter-bootstrap/5.3.0/css/bootstrap.min.css" rel="stylesheet" />
+    <link href="https://cdn.datatables.net/1.13.6/css/dataTables.bootstrap5.min.css" rel="stylesheet" />
+    </style>
+    <script src="https://code.jquery.com/jquery-3.7.0.js"></script>
+    <script src="https://cdn.datatables.net/1.13.6/js/jquery.dataTables.min.js"></script>
+    <script src="https://cdn.datatables.net/1.13.6/js/dataTables.bootstrap5.min.js"></script>
+    <script type="text/javascript" class="init">
+    $(document).ready( function () {
+        $('table.sg-datatable').DataTable({order: [[1, 'desc']]});
+    } );
+    </script>
+
+  .. list-table::
+   :header-rows: 1
+   :class: table table-striped sg-datatable
+
+   * - Example
+     - Time
+     - Mem (MB)
+   * - :ref:`sphx_glr_tutorials__rendered_examples_dynamo_torch_compile_advanced_usage.py` (``../examples/dynamo/torch_compile_advanced_usage.py``)
+     - 00:00.000
+     - 0.0
+   * - :ref:`sphx_glr_tutorials__rendered_examples_dynamo_torch_compile_resnet_example.py` (``../examples/dynamo/torch_compile_resnet_example.py``)
+     - 00:00.000
+     - 0.0
+   * - :ref:`sphx_glr_tutorials__rendered_examples_dynamo_torch_compile_stable_diffusion.py` (``../examples/dynamo/torch_compile_stable_diffusion.py``)
+     - 00:00.000
+     - 0.0
+   * - :ref:`sphx_glr_tutorials__rendered_examples_dynamo_torch_compile_transformers_example.py` (``../examples/dynamo/torch_compile_transformers_example.py``)
+     - 00:00.000
+     - 0.0

docsrc/tutorials/images/circ_pad_example.png

@@ -35,7 +35,7 @@ Here's an example usage
     inputs = [torch.randn((1, 3, 224, 224)).cuda()]
     # trt_ep is a torch.fx.GraphModule object
     trt_gm = torch_tensorrt.compile(model, ir="dynamo", inputs) 
-    torchtrt.save(trt_gm, "trt.ep", inputs=inputs)
+    torch_tensorrt.save(trt_gm, "trt.ep", inputs=inputs)
 
     # Later, you can load it and run inference
     model = torch.export.load("trt.ep").module()
@@ -97,4 +97,4 @@ Here's an example usage
     # file_path can be trt.ep or trt.ts file obtained via saving the model (refer to the above section)
     inputs = [torch.randn((1, 3, 224, 224)).cuda()]
     model = torch_tensorrt.load(<file_path>).module()
-    model(*inputs)
+    model(*inputs)

docsrc/user_guide/saving_models.rst

@@ -10,3 +10,4 @@ a number of ways you can leverage this backend to accelerate inference.
 * :ref:`torch_compile_transformer`: Compiling a Transformer model using ``torch.compile``
 * :ref:`torch_compile_advanced_usage`: Advanced usage including making a custom backend to use directly with the ``torch.compile`` API
 * :ref:`torch_compile_stable_diffusion`: Compiling a Stable Diffusion model using ``torch.compile``
+* :ref:`custom_kernel_plugins`: Creating a plugin to use a custom kernel inside TensorRT engines