|  | 
| 1 |  | -# PyTorch Model Delegation to Neutron Backend | 
|  | 1 | +# ExecuTorch Neutron Backend examples | 
|  | 2 | +This directory contains examples demonstrating the use of ExecuTorch AoT flow to convert a PyTorch model to ExecuTorch | 
|  | 3 | +format and delegate the model computation to eIQ Neutron NPU using the eIQ Neutron Backend. | 
| 2 | 4 | 
 | 
| 3 |  | -In this guide we will show how to use the ExecuTorch AoT flow to convert a PyTorch model to ExecuTorch format and delegate the model computation to eIQ Neutron NPU using the eIQ Neutron Backend. | 
|  | 5 | +## Layout | 
|  | 6 | +* `experimental/` - contains CifarNet model example. | 
|  | 7 | +* `models` - various example models. | 
|  | 8 | +* `aot_neutron_compile.py` - script with end-to-end ExecuTorch AoT Neutron Backend workflow. | 
|  | 9 | +* `README.md` - this file. | 
|  | 10 | +* `run_aot_example.sh` - utility script for aot_neutron_compile.py. | 
|  | 11 | +* `setup.sh` - setup script for Neutron Converter installation. | 
| 4 | 12 | 
 | 
| 5 |  | -First we will start with an example script converting the model. This example show the CifarNet model preparation. It is the same model which is part of the `example_cifarnet` | 
|  | 13 | +## Setup | 
|  | 14 | +Please finish tutorial [Setting up ExecuTorch](https://pytorch.org/executorch/main/getting-started-setup). | 
| 6 | 15 | 
 | 
| 7 |  | -The steps are expected to be executed from the executorch root folder. | 
| 8 |  | -1. Run the setup.sh script to install the neutron-converter: | 
|  | 16 | +Run the setup.sh script to install the neutron-converter: | 
| 9 | 17 | ```commandline | 
| 10 |  | -$ examples/nxp/setup.sh | 
|  | 18 | +$ ./examples/nxp/setup.sh | 
| 11 | 19 | ``` | 
| 12 | 20 | 
 | 
| 13 |  | -2. Now run the `aot_neutron_compile.py` example with the `cifar10` model  | 
| 14 |  | -```commandline | 
| 15 |  | -$ python -m examples.nxp.aot_neutron_compile --quantize \ | 
| 16 |  | -    --delegate --neutron_converter_flavor SDK_25_09 -m cifar10  | 
| 17 |  | -``` | 
|  | 21 | +## Supported models | 
|  | 22 | +* CifarNet | 
|  | 23 | +* MobileNetV2 | 
|  | 24 | + | 
|  | 25 | +## PyTorch Model Delegation to Neutron Backend | 
|  | 26 | +First we will start with an example script converting the model. This example show the CifarNet model preparation.  | 
|  | 27 | +It is the same model which is part of the `example_cifarnet` in  | 
|  | 28 | +[MCUXpresso SDK](https://www.nxp.com/design/design-center/software/development-software/mcuxpresso-software-and-tools-/mcuxpresso-software-development-kit-sdk:MCUXpresso-SDK). | 
|  | 29 | + | 
|  | 30 | +The NXP MCUXpresso software and tools offer comprehensive development solutions designed to help accelerate embedded  | 
|  | 31 | +system development of applications based on MCUs from NXP. The MCUXpresso SDK includes a flexible set of peripheral  | 
|  | 32 | +drivers designed to speed up and simplify development of embedded applications. | 
|  | 33 | + | 
|  | 34 | +The steps are expected to be executed from the `executorch` root folder. | 
|  | 35 | + | 
|  | 36 | +1. Run the `aot_neutron_compile.py` example with the `cifar10` model  | 
|  | 37 | +    ```commandline | 
|  | 38 | +    $ python -m examples.nxp.aot_neutron_compile --quantize \ | 
|  | 39 | +        --delegate --neutron_converter_flavor SDK_25_09 -m cifar10  | 
|  | 40 | +    ``` | 
| 18 | 41 | 
 | 
| 19 |  | -3. It will generate you `cifar10_nxp_delegate.pte` file which can be used with the MXUXpresso SDK `cifarnet_example` project, presented [here](https://mcuxpresso.nxp.com/mcuxsdk/latest/html/middleware/eiq/executorch/docs/nxp/topics/example_applications.html#how-to-build-and-run-executorch-cifarnet-example). | 
| 20 |  | -To get the MCUXpresso SDK follow this [guide](https://mcuxpresso.nxp.com/mcuxsdk/latest/html/middleware/eiq/executorch/docs/nxp/topics/getting_mcuxpresso.html), use the MCUXpresso SDK v25.03.00.  | 
|  | 42 | +2. It will generate you `cifar10_nxp_delegate.pte` file which can be used with the MCUXpresso SDK `cifarnet_example`  | 
|  | 43 | +project, presented [here](https://mcuxpresso.nxp.com/mcuxsdk/latest/html/middleware/eiq/executorch/docs/nxp/topics/example_applications.html#how-to-build-and-run-executorch-cifarnet-example). | 
|  | 44 | +This project will guide you through the process of deploying your PTE model to the device. | 
|  | 45 | +To get the MCUXpresso SDK follow this [guide](https://mcuxpresso.nxp.com/mcuxsdk/latest/html/middleware/eiq/executorch/docs/nxp/topics/getting_mcuxpresso.html), | 
|  | 46 | +use the MCUXpresso SDK v25.09.00.  | 
0 commit comments