Machine Learning in Linux: GPT4All – local AI chat application

GPT4All Chat is a locally-running AI chat application powered by the GPT4All-J Apache 2 Licensed chatbot. The software lets you communicate with a large language model (LLM) to get helpful answers, insights, and suggestions.

The model runs on a local computer’s CPU and doesn’t require a net connection. No chat data is sent to external services.

This is free and open source cross-platform software.

Installation

We will take you through the simple steps of installing the Graphical User Interface (GUI) and the Command-Line Interface (CLI) versions.

GUI

Download the Linux installer with the command:

$ wget https://gpt4all.io/installers/gpt4all-installer-linux.run

Make the file executable:

$ chmod u+x gpt4all-installer-linux.run

Run the installer with the command:

$ ./gpt4all-installer-linux.run

By default GPT4ALL is installed to /opt/gpt4all but you can change the installation folder. You’ll see the following dialog box once you’ve specified the installation folder.

The GPT4All Chat installer decompresses a 3GB LLM model during the installation process which may take time on a slower machine.

On a fresh installation of Ubuntu 23.04 we also needed to install the libxcb-cursor package.

$ sudo apt install libxcb-cursor0

The GUI can then be started from Activities.

CLI

On a fresh installation of Ubuntu 23.04, we first need to install git.

$ sudo apt install git

Next clone the project’s GitHub directory:

$ git clone https://github.com/nomic-ai/gpt4all

Download the 4GB LLM model to the gpt4all/chat directory:

$ cd gpt4all/chat
$ wget https://the-eye.eu/public/AI/models/nomic-ai/gpt4all/gpt4all-lora-quantized.bin

Start the CLI with the command:

$ ./gpt4all-lora-quantized-linux-x86

Next page: Page 2 – In Operation and Summary

Pages in this article:
Page 1 – Introduction and Installation
Page 2 – In Operation and Summary