Suggestion

#1
by Yobenboben - opened

There's an 8B model with a really interesting merge recipe. I wonder how well the same technique would work when applied to 70B and for creative purposes. You seem to like to experiment, so I thought you'd be interested.
Said model: https://huggingface.co/Casual-Autopsy/Llama-3-VNTL-Yollisa-8B
Thanks for your work

Owner

There's an 8B model with a really interesting merge recipe. I wonder how well the same technique would work when applied to 70B and for creative purposes. You seem to like to experiment, so I thought you'd be interested.
Said model: https://huggingface.co/Casual-Autopsy/Llama-3-VNTL-Yollisa-8B
Thanks for your work

Thanks for this! Always looking for new tricks.

Sign up or log in to comment