Live data from Hacker News

Synthia-7B-v1.3 hits 64.85 on 4-evals; LLaMA-2-70B-Chat at 66.8

Post reply on HN

Print view View original

1–4 of 4 posts

Synthia-7B-v1.3 hits 64.85 on 4-evals; LLaMA-2-70B-Chat at 66.8

#1

Post by JackRumford » Fri, Sep 29, 2023, 11:02 AM UTC

Synthia-7B-v1.3 hits 64.85 on 4-evals; LLaMA-2-70B-Chat at 66.8
twitter.com

Re: Synthia-7B-v1.3 hits 64.85 on 4-evals; LLaMA-2-70B-Chat at 66.8

#2

Post by JackRumford » Fri, Sep 29, 2023, 11:04 AM UTC

HuggingFace model: https://huggingface.co/migtissera/Synthia-7B-v1.3

Re: Synthia-7B-v1.3 hits 64.85 on 4-evals; LLaMA-2-70B-Chat at 66.8

#3

Post by sgolem » Fri, Sep 29, 2023, 12:03 PM UTC

Can this be run with llama.cpp?

Re: Synthia-7B-v1.3 hits 64.85 on 4-evals; LLaMA-2-70B-Chat at 66.8

#4

Post by JackRumford » Fri, Sep 29, 2023, 2:22 PM UTC

sgolem wrote: post #3

Can this be run with llama.cpp?

yep. this guy keeps up to date with quantised models: https://huggingface.co/TheBloke

Post reply on HN