Update

[ghstack-poisoned]

Author: swolchok

.github/workflows/doc-build.yml

@@ -84,8 +84,8 @@ jobs:
     needs: build
     if: github.repository == 'pytorch/executorch' && github.event_name == 'push' && (github.ref == 'refs/heads/main' || startsWith(github.ref, 'refs/tags/v'))
     permissions:
+      id-token: write
       contents: write
-      contents: read
     uses: pytorch/test-infra/.github/workflows/linux_job_v2.yml@main
     with:
       repository: pytorch/executorch

README.md

@@ -1,9 +1,37 @@
-# ExecuTorch
-
-**ExecuTorch** is an end-to-end solution for enabling on-device inference
-capabilities across mobile and edge devices including wearables, embedded
-devices and microcontrollers. It is part of the PyTorch Edge ecosystem and
-enables efficient deployment of PyTorch models to edge devices.
+<div align="center">
+  <img src="./docs/source/_static/img/et-logo.png" alt="Logo" width="200">
+  <h1 align="center">ExecuTorch: A powerful on-device AI Framework</h1>
+</div>
+
+
+<div align="center">
+  <a href="https://github.com/pytorch/executorch/graphs/contributors"><img src="https://img.shields.io/github/contributors/pytorch/executorch?style=for-the-badge&color=blue" alt="Contributors"></a>
+  <a href="https://github.com/pytorch/executorch/stargazers"><img src="https://img.shields.io/github/stars/pytorch/executorch?style=for-the-badge&color=blue" alt="Stargazers"></a>
+  <a href="https://discord.gg/MeacgB7A"><img src="https://img.shields.io/badge/Discord-Join%20Us-purple?logo=discord&logoColor=white&style=for-the-badge" alt="Join our Discord community"></a>
+  <a href="https://pytorch.org/executorch/stable/index.html"><img src="https://img.shields.io/badge/Documentation-000?logo=googledocs&logoColor=FFE165&style=for-the-badge" alt="Check out the documentation"></a>
+  <hr>
+</div>
+
+**ExecuTorch** is an end-to-end solution for on-device inference and training. It powers much of Meta's on-device AI experiences across Facebook, Instagram, Meta Quest, Ray-Ban Meta Smart Glasses, WhatsApp, and more.
+
+It supports a wide range of models including LLMs (Large Language Models), CV (Computer Vision), ASR (Automatic Speech Recognition), and TTS (Text to Speech).
+
+Platform Support:
+- Operating Systems:
+  - iOS
+  - Mac
+  - Android
+  - Linux
+  - Microcontrollers
+
+- Hardware Acceleration:
+  - Apple
+  - Arm
+  - Cadence
+  - MediaTek
+  - Qualcomm
+  - Vulkan
+  - XNNPACK
 
 Key value propositions of ExecuTorch are:
 
@@ -17,35 +45,21 @@ Key value propositions of ExecuTorch are:
   experience due to a lightweight runtime and utilizing full hardware
   capabilities such as CPUs, NPUs, and DSPs.
 
-For a comprehensive technical overview of ExecuTorch and step-by-step tutorials,
-please visit our documentation website [for the latest release](https://pytorch.org/executorch/stable/index.html) (or the [main branch](https://pytorch.org/executorch/main/index.html)).
-
-Check out the [Getting Started](https://pytorch.org/executorch/stable/getting-started-setup.html#quick-setup-colab-jupyter-notebook-prototype) page for a quick spin.
-
-Check out the examples of [Llama](./examples/models/llama/README.md), [Llava](./examples/models/llava/README.md) and [other models](./examples/README.md) running on edge devices using ExecuTorch.
+## Getting Started
+To get started you can:
 
+- Visit the [Step by Step Tutorial](https://pytorch.org/executorch/main/index.html) on getting things running locally and deploy a model to a device
+- Use this [Colab Notebook](https://pytorch.org/executorch/stable/getting-started-setup.html#quick-setup-colab-jupyter-notebook-prototype) to start playing around right away
+- Jump straight into LLMs use cases by following specific instructions for [Llama](./examples/models/llama/README.md) and [Llava](./examples/models/llava/README.md)
 
-**[UPDATE - 10/24]** We have added support for running [Llama 3.2 Quantized 1B/3B](./examples/models/llama/README.md) models via ExecuTorch.
-
-## Feedback
+## Feedback and Engagement
 
 We welcome any feedback, suggestions, and bug reports from the community to help
-us improve our technology. Please use the [PyTorch
-Forums](https://discuss.pytorch.org/c/executorch) for discussion and feedback
-about ExecuTorch using the **ExecuTorch** category, and our [GitHub
-repository](https://github.com/pytorch/executorch/issues) for bug reporting.
-
-We recommend using the latest release tag from the
-[Releases](https://github.com/pytorch/executorch/releases) page when developing.
+us improve our technology. Check out the [Discussion Board](https://github.com/pytorch/executorch/discussions) or chat real time with us on [Discord](https://discord.gg/MeacgB7A)
 
 ## Contributing
 
-See [CONTRIBUTING.md](CONTRIBUTING.md) for details about issues, PRs, code
-style, CI jobs, and other development topics.
-
-To connect with us and other community members, we invite you to join PyTorch Slack community by filling out this [form](https://docs.google.com/forms/d/e/1FAIpQLSeADnUNW36fjKjYzyHDOzEB_abKQE9b6gqqW9NXse6O0MWh0A/viewform). Once you've joined, you can:
-* Head to the `#executorch-general` channel for general questions, discussion, and community support.
-* Join the `#executorch-contributors` channel if you're interested in contributing directly to project development.
+We welcome contributions. To get started review the [guidelines](CONTRIBUTING.md) and chat with us on [Discord](https://discord.gg/MeacgB7A)
 
 
 ## Directory Structure

backends/cadence/fusion_g3/operators/op_mean.cpp

@@ -60,7 +60,7 @@ int prepare_data(
   return num_axis_dims;
 }
 
-Tensor& mean_dim_out(
+Tensor& mean_out(
     KernelRuntimeContext& ctx,
     const Tensor& in,
     optional<ArrayRef<int64_t>> dim_list,
@@ -169,29 +169,32 @@ Tensor& mean_dim_out(
         InvalidArgument,
         out);
 
-    ET_SWITCH_REALHB_TYPES(in.scalar_type(), ctx, "mean.out", CTYPE_IN, [&] {
-      ET_SWITCH_FLOATH_TYPES(
-          out.scalar_type(), ctx, "mean.out", CTYPE_OUT, [&] {
-            CTYPE_OUT* out_data = out.mutable_data_ptr<CTYPE_OUT>();
-            const size_t num =
-                torch::executor::get_reduced_dim_product(in, dim_list);
-            for (size_t out_ix = 0; out_ix < out.numel(); ++out_ix) {
-              CTYPE_OUT sum = 0;
-              if (in.numel() > 0) {
-                sum = torch::executor::
-                    map_reduce_over_dim_list<CTYPE_IN, CTYPE_OUT>(
-                        [](CTYPE_IN v) { return static_cast<CTYPE_OUT>(v); },
-                        [](CTYPE_OUT outv, CTYPE_OUT acc) {
-                          return acc + outv;
-                        },
-                        in,
-                        dim_list,
-                        out_ix);
-              }
-              out_data[out_ix] = sum / static_cast<float>(num);
-            }
-          });
-    });
+    ET_SWITCH_REALHBBF16_TYPES(
+        in.scalar_type(), ctx, "mean.out", CTYPE_IN, [&] {
+          ET_SWITCH_FLOATHBF16_TYPES(
+              out.scalar_type(), ctx, "mean.out", CTYPE_OUT, [&] {
+                CTYPE_OUT* out_data = out.mutable_data_ptr<CTYPE_OUT>();
+                const size_t num =
+                    torch::executor::get_reduced_dim_product(in, dim_list);
+                for (size_t out_ix = 0; out_ix < out.numel(); ++out_ix) {
+                  CTYPE_OUT sum = 0;
+                  if (in.numel() > 0) {
+                    sum = torch::executor::
+                        map_reduce_over_dim_list<CTYPE_IN, CTYPE_OUT>(
+                            [](CTYPE_IN v) {
+                              return static_cast<CTYPE_OUT>(v);
+                            },
+                            [](CTYPE_OUT outv, CTYPE_OUT acc) {
+                              return acc + outv;
+                            },
+                            in,
+                            dim_list,
+                            out_ix);
+                  }
+                  out_data[out_ix] = sum / static_cast<float>(num);
+                }
+              });
+        });
   }
 
   return out;

backends/qualcomm/tests/test_qnn_delegate.py

@@ -37,8 +37,9 @@
     skip_annotation,
     update_spill_fill_size,
 )
+from executorch.examples.models.llama.llama_transformer import MOEFeedForward
 
-from executorch.examples.models.llama.llama_transformer import ModelArgs, MOEFeedForward
+from executorch.examples.models.llama.model_args import ModelArgs
 
 from executorch.examples.qualcomm.utils import setup_common_args_and_variables
 

docs/source/_static/img/et-logo.png

@@ -22,7 +22,16 @@ def apply_tensor_contraints(op_name: str, tensor_constraints: list[object]) -> N
             tensor_constraints.extend(
                 [
                     cp.Dtype.In(lambda deps: [torch.float]),
-                    cp.Rank.Le(lambda deps: 2**3),
+                    cp.Rank.Le(lambda deps: 2**2),
+                    cp.Value.Ge(lambda deps, dtype, struct: -2),
+                    cp.Value.Le(lambda deps, dtype, struct: 2),
+                ]
+            )
+        case "mean.dim":
+            tensor_constraints.extend(
+                [
+                    cp.Dtype.In(lambda deps: [torch.float]),
+                    cp.Rank.Le(lambda deps: 2**2),
                 ]
             )
         case "exp.default":
@@ -86,8 +95,27 @@ def facto_testcase_gen(op_name: str) -> List[Tuple[List[str], OrderedDict[str, s
                         cp.Value.Le(lambda deps, dtype: 2),
                     ]
                 )
+        elif in_spec.type.is_scalar_type():
+            spec.inspec[index].constraints.extend(
+                [
+                    cp.Dtype.In(lambda deps: apply_scalar_contraints(op_name)),
+                ]
+            )
         elif in_spec.type.is_tensor():
             spec.inspec[index].constraints.extend(tensor_constraints)
+        elif in_spec.type.is_dim_list():
+            spec.inspec[index].constraints.extend(
+                [
+                    cp.Length.Ge(lambda deps: 1),
+                    cp.Optional.Eq(lambda deps: False),
+                ]
+            )
+        elif in_spec.type.is_bool():
+            spec.inspec[index].constraints.extend(
+                [
+                    cp.Dtype.In(lambda deps: [torch.bool]),
+                ]
+            )
 
     return [
         (posargs, inkwargs)

examples/cadence/operators/facto_util.py

@@ -259,6 +259,35 @@ def test_g3__softmax_out(
 
         self.run_and_verify(model, (inputs,))
 
+    # pyre-ignore[16]: Module `parameterized.parameterized` has no attribute `expand`.
+    @parameterized.expand([*facto_util.facto_testcase_gen("mean.dim")])
+    def test_g3_mean_dim_out(
+        self,
+        posargs: List[int],
+        inkwargs: OrderedDict[str, str],
+    ) -> None:
+        class Meandim(nn.Module):
+            def forward(
+                self,
+                x: torch.Tensor,
+                dim_list: Tuple[int],
+                keepdim: bool,
+                dtype: torch.dtype = torch.float32,
+            ) -> torch.Tensor:
+                return torch.ops.aten.mean.dim(
+                    x,
+                    dim_list,
+                    keepdim,
+                    dtype=dtype,
+                )
+
+        model = Meandim()
+
+        self.run_and_verify(
+            model,
+            inputs=tuple(posargs),
+        )
+
 
 if __name__ == "__main__":
     unittest.main()

examples/cadence/operators/test_g3_ops.py

@@ -14,6 +14,8 @@ runtime.python_library(
     srcs = [
         "llama_transformer.py",
         "rope.py",
+        "attention.py",
+        "model_args.py",
     ],
     _is_external_target = True,
     base_module = "executorch.examples.models.llama",