News for hackers. Stuff that matters.

A comment on MiniGPT-4. Parent comment

Re:MiniGPT-4 by HopenHeyHi on Monday April 17, 2023 @07:23AM (#35599209)
Thanks for a useful comment.

Do you reckon the 4bit quantized Vicuna just won't do here? https://huggingface.co/anon8231489123/vicuna-13b-GPTQ-4bit-1...

I think with this everything OpenAI demonstrated ~5 weeks ago has been recreated by actually-open AI. Even if it runs much much slower on prosumer hardware and with less good results at least it is de-magicked.

[ Parent | on HN ]
Re:MiniGPT-4 by p1esk on Monday April 17, 2023 @07:35AM (#35599272)
The magic is in the quality of GPT-4 output. That hasn’t been recreated yet.
[ Parent | on HN ]
Re:MiniGPT-4 by 152334H on Monday April 17, 2023 @07:45AM (#35599349)
It'll work! I just haven't touched any of the 4bit stuff myself, so I don't personally know how to add it. Great low-hanging fruit for anyone else to take on.
[ Parent | on HN ]
Re:MiniGPT-4 by kolinko on Monday April 17, 2023 @12:17PM (#35602910)
Open AI still hasn't exactly reached the level of gpt3.5. GPT-4 is way ahead of anything.
[ Parent | on HN ]