It has come a long way in decade. Can run VMs or LXC with SDN vlans and full network stacks. I run a full home cyber range network on it. Can split up vGPUs or passthrough hardware. Not so much desktop but I use mostly cli.

Qubes is great for isolation of malicious code for threat hunting. Most people get hung up on the networking/usb and copy/paste between VMs. I use Proxmox similar to Qubes.

Don't distract Primal need relay choosing first.

Hope someone who met him IRL can check-in on him. 🫂

Just to calrify. In order to do YouTube videos I need to buy the biggest microphone and boom arm I can find?

🎙️✔️1 2

After all. If you help train OpenAI you are indirectly helping Worldcoin 😉

Oops links got jumbled.

## Web Interfaces and API Interfaces

[Text Generation WebUI](https://github.com/oobabooga/text-generation-webui)

[h2oGFPT](https://github.com/h2oai/h2ogpt)

[localai.io](https://localai.io)

[LLaMa2-chatbot](https://github.com/a16z-infra/llama2-chatbot)

[Petals Distributed Models](https://petals.dev/)

I just posted a note about this with some good starting points. 🤙

# Self-Host AI Models for Customizing and Privacy

It is very east now to run and fine-tune your own LLM (Large Language Model) like ChatGPT at home. The benefit is learning how it all works and the privacy to fine tune a model to your needs. Instead of training OpenAI's for them.

Here are some projects to get you started with self-hosted LLMs on a basic NVIDIA GPU or even with just CPU.

## Web Interfaces and API Interfaces

[Text Generation WebUI](https://github.com/oobabooga/text-generation-webui)

[h2oGFPT](https://github.com/h2oai/h2ogpt)

[localai.io](https://localai.io)

[LLaMa2-chatbot](https://github.com/a16z-infra/llama2-chatbot)

[Petals Distributed Models](https://petals.dev/)

## Voice Assistant

[Willow](https://github.com/toverainc/willow)

[HomeAssistant](https://www.home-assistant.io/)

## Models

[HuggingFace](https://huggingface.co/models)

## DataSets for Training

[LLMDataHub](https://github.com/Zjh-819/LLMDataHub)

[HuggingFace DataSets](https://huggingface.co/datasets)

## Testing

[TermGPT](https://github.com/Sentdex/TermGPT)

#selfhost #self-host #llm #ai #llama2 #gpt #linux #diy

That time of year in Southern California 🔥

Pretty sure cd and ls are top of most people. 🤙

## What are your top ten cli commands used?

```

history | awk '{cmd[$2]++} END {for(elem in cmd) {print cmd[elem] " " elem}}' | sort -n -r | head -10

```

Then add aliases and functions for them in your .bashrc 😉

He was awesome and a good actor in things he did post-peewee.

*freebased 🤣

People see them like the police. Wanted when needed, but otherwise feel better when they are not around.

I have similar but is RUN GCC. 🤙

Waiting for ability to add/edit relays. Then will be sexy as in freedom sexy. 🤙

Elon Musk is even defacing the night sky. Is nothing sacred? 👾

One thing can do is multiple user profiles with one having Google Services running.

The problem with the world is that the intelligent people are full of self-doubt, while the stupid ones are full of confidence."

- Charles Bukowski