Installation

Installing

Download locas_installer.py, and let magic happens.

Window	Unix
python locas_installer.py	python3 locas_installer.py

Let's go through all questions!

Question 1: Choose path

Please choose the path for LocalAssistant. [...]:

Now, path is where you want LocalAssistant to be.

Your path
│
├───.venv <dir - a python virtual environment>
│
├───documents <dir - to store docs used for `locas docs ...`>
│
├───models <dir - where installed models is stored>
│
├───users <dir - where users' data is stored>
│
└───locas.cmd <file - script to use `locas.cmd` anywhere>

Choose the folder, copy its path (Ctrl+Shift+C for Window), and paste it in. If leaving it empty, the path will be that inside the bracket [...].

Notice: It may ask for comfirmation when folder is already existed.

Question 2: Choose Pytorch compute

Choose your PyTorch compute from these:
  - ...
Select one [cpu]:

Notice: MacOS will be skipped. (Only cpu is supported)

What shown below is your device compute. Copy that compute name to the answer and you are done. Leave it empty for using cpu. Remember to choose the RIGHT compute, Cuda 12.4 is not with Cuda 12.6.

Question 3: Choose version

Available version: ...
  Pre-release version: ...
  Latest version: ...
Which version to install [...]:

All version that got published will be shown. Choose the one that fill your need. Leave empty for the latest version.

Updating

What you need is installing LocalAssistant again on the path where needed to be upgraded (Question 1). Your existed assets won't be touched during this process.