News for hackers. Stuff that matters.
Re:How to run Llama 13B with a 6GB graphics card by mozillas on Sunday May 14, 2023 @10:31AM (#35938293)
I ran the 7B Vicuna (ggml-vic7b-q4_0.bin) on a 2017 MacBook Air (8GB RAM) with llama.cpp.

Worked OK for me with the default context size. 2048, like you see in most examples was too slow for my taste.

Re:How to run Llama 13B with a 6GB graphics card by koheripbal on Sunday May 14, 2023 @10:35AM (#35938331)
Given the current price (mostly free) off public llms I'm not sure what the use case of running out at home are yet.

OpenAIs paid GPT4 has few restrictions and is still cheap.

... Not to mention GPT4 with browsing feature is vastly superior to any home of the models you can run at home.

Re:How to run Llama 13B with a 6GB graphics card by 2devnull on Sunday May 14, 2023 @11:02AM (#35938517)
Many would be users can’t send their data data to openAI. Think HIPPA and other laws restricting data sharing. Federation or distribution of the models for local training is the other solution to that problem.
Re:How to run Llama 13B with a 6GB graphics card by toxik on Sunday May 14, 2023 @12:26PM (#35939120)
The point for me personally is the same as why I find it so powerful to self host SMTP, IMAP, HTTP. It’s in my hands, I know where it all begins and ends. I answer to no one.

For LLMs this means I am allowed their full potential. I can generate smut, filth, illegal content of any kind for any reason. It’s for me to decide. It’s empowering, it’s the hacker mindset.

Re:How to run Llama 13B with a 6GB graphics card by sagarm on Sunday May 14, 2023 @01:31PM (#35939634)
I think it's mostly useful if you want to do your own fine tuning, or the data you are working with can't be sent to a third party for contractual, legal, or paranoid reasons.
Re:How to run Llama 13B with a 6GB graphics card by sroussey on Sunday May 14, 2023 @05:32PM (#35941513)
I’m working on an app to index your life, and having it local is a huge plus for the people I have using it.
Re:How to run Llama 13B with a 6GB graphics card by theaussiestew on Sunday May 14, 2023 @10:54PM (#35943435)
Sounds interesting, got a link?
Re:How to run Llama 13B with a 6GB graphics card by sroussey on Tuesday May 16, 2023 @12:40AM (#35957812)
Not yet…
Re:How to run Llama 13B with a 6GB graphics card by int_19h on Monday May 15, 2023 @05:56PM (#35955050)
GPT4 API is still not universally available, for starters.