Install Windows Subsystem for Linux (WSL) to use Local AI Models

Purpose

For local services on desktops and laptops, the VHToolkit uses a local endpoint wrapped around an AI model, which are often developed on Linux with Python. These can run on Windows with the Windows Subsystem for Linux (WSL).

Installation and Setup

To install WSL, follow instructions on https://learn.microsoft.com/en-us/windows/wsl/install.

Conda enables developers to set up encapsulated Linux / Python development environments, keeping all dependencies isolated per feature (e.g., local sensing, local LLM). To install Conda / MiniForge, download and run the Windows installer from https://conda-forge.org/download. Alternatively, detailed information and instructions to install from the WSL command line are at https://github.com/conda-forge/miniforge.

Usage

Overview

To launch WSL, first open a command line with Windows key + R, and type 'cmd'. In the new command line window, type:

wsl ~

This opens the prompt in your Linux home folder.

Conda is used to create separate environments for each locally running service in order to isolate dependencies. We use the following naming convention: <type>_<tech-name>_env, where type is nlp, asr, tts, etc. For example: asr_whisper_env, nlp_rasa_env, sen_deepface_env. To create a Conda environment, type:

conda create -n <env-name> python=3.xx
conda init  
conda activate <env-name>

Examples

See the following examples for detailed instructions on how to set up a local AI service:

• DeepFace (Sensing)
• Rasa (NLP)