Llama-2 because you can run it on your own hardware. For the big GPU on a rented instance: Falcon 70b. OpenAI and Google can have turns playing proprietary asshat jack in the box.
How expensive would it be to run on a rented server?
For the bigger ones you could do it under $.50 / hr. I run llama 2 13b-8 bit on my 3090 no problem, which can be rented for $.20/hr.
Some of the lower pricing I’ve seen.
deleted by creator
I’m a software engineer by trade, but a hobbyist when it comes to LLMs.
https://github.com/AUTOMATIC1111
This is a good place to start. Loads of YouTubers with setup videos. I like this guy https://youtube.com/@Aitrepreneur he covers LLM and Image generation too.
Hugging face is a good place to get the actual LLM.
Unless it’s changed recently, I don’t think automatic1111 deals with LLMs. It uses stable diffusion based technologies for text-to-image, image-to-image etc. But yes, it’s a good entry point for dinking around with ai/machine learning
You’re right, I’m on mobile driving home.
I meant this one: https://github.com/oobabooga/text-generation-webui
And this YouTube video specifically for setup:
