Feature/SK-1356 | Added example of keyword detection using client api

[carl-andersson]

This pull request to examples/pytorch-keyworddetection-api as a new example for keyword detection using PyTorch. The changes involve adding new files and new documentation.

New Example Script:

• Added a new client script for keyword detection using Fedn. This script includes data loading (data.py), model training (defined in model.py), validation, and prediction functionalities.

Documentation Updates:

• Updated the README file with instructions on setting up and running the keyword detection example, including prerequisites, environment setup, and usage instructions.

Data Handling:

• Added a new data handling script that includes classes for background noise and keyword datasets, as well as functions for data loading and preprocessing.

Configuration Files:

• Created a new environment configuration file for setting up the required dependencies using Conda.
• Updated the .gitignore file to exclude data files and specific file types generated during the example execution.

API Client Setup:

• Added a script to initialize the seed model and start a session using the FEDn API client.## Description

[Wrede]

requires conda?

[carl-andersson]

No, but you can't install SOX via pip. So in that case you will need to run
brew install sox

[Wrede]

maybe add link to dataset explaination: https://huggingface.co/datasets/google/speech_commands