Quick start example ML application

This is a quick start guide that will show you how to run the keyword spotting example application. The aim of this guide is to illustrate the flow of running an application on the evaluation kit rather than showing the keyword spotting functionality or performance. All use cases in the evaluation kit follow the steps.

Verify you have installed the required prerequisites.

Clone the Ethos-U55 evaluation kit repository.

git clone "https://review.mlplatform.org/ml/ethos-u/ml-embedded-evaluation-kit"
cd ml-embedded-evaluation-kit

Pull all the external dependencies with the commands below:
```
git submodule update --init
```
Next, you would need to get a neural network model. For the purpose of this quick start guide, we'll use the ds_cnn_clustered_int8 keyword spotting model from the Arm public model zoo and the principle remains the same for all of the other use cases. Download the ds_cnn_large_int8.tflite model file with the curl command below:
```
curl -L https://github.com/ARM-software/ML-zoo/blob/master/models/keyword_spotting/ds_cnn_large/tflite_clustered_int8/ds_cnn_clustered_int8.tflite?raw=true --output ds_cnn_clustered_int8.tflite
```
Vela is an open-source python tool converting TensorFlow Lite for Microcontrollers neural network model into an optimized model that can run on an embedded system containing an Ethos-U55 NPU. It is worth noting that in order to take full advantage of the capabilities of the NPU, the neural network operators should be supported by Vela. In this step, you will compile the model with Vela.
For this step, you need to ensure you have correctly installed the Vela package:
```
python3 -m venv env
source ./env/bin/activate
pip install --upgrade pip
pip install ethos-u-vela
```
In the command below, we specify that we are using the Arm® Ethos™-U55 NPU with a 128 Multiply-Accumulate units (MAC units) configured for a High End Embedded use case. The building section has more detailed explanation about Vela usage.
```
vela ds_cnn_clustered_int8.tflite \
    --accelerator-config=ethos-u55-128 \
    --block-config-limit=0 \
    --config scripts/vela/vela.ini \
    --memory-mode Shared_Sram \
    --system-config Ethos_U55_High_End_Embedded
```
An optimized model file for Ethos-U55 is generated in a folder named output.
Create a build folder in the root level of the evaluation kit.
```
mkdir build && cd build
```

Build the makefiles with CMake as shown in the command below. The build process section gives an in-depth explanation about the meaning of every parameter. For the time being, note that we point the Vela optimized model from stage 5 in the -Dkws_MODEL_TFLITE_PATH parameter.

cmake \
    -DTARGET_PLATFORM=mps3 \
    -DTARGET_SUBSYSTEM=sse-300 \
    -DCMAKE_TOOLCHAIN_FILE=scripts/cmake/bare-metal-toolchain.cmake \
    -DUSE_CASE_BUILD=kws \
    -Dkws_MODEL_TFLITE_PATH=output/ds_cnn_clustered_int8_vela.tflite \
    ..

Compile the project with a make. Details about this stage can be found in the building part of the documentation.
```
make -j4
```
Launch the project as explained here. In this quick-start guide, we'll use the Fixed Virtual Platform. Point the generated bin/ethos-u-kws.axf file in stage 8 to the FVP that you have downloaded when installing the prerequisites.
```
<path_to_FVP>/FVP_Corstone_SSE-300_Ethos-U55 -a ./bin/ethos-u-kws.axf
```
A telnet window is launched through which you can interact with the application and obtain performance figures.