Probably the new best "base" model to use to merge

#1
by paulml - opened

The other ones are unusable when they are quantized: https://huggingface.co/CultriX/NeuralTrix-7B-dpo/discussions/1

Did you manage to make it work with Monarch-7B?

Did you manage to make it work with Monarch-7B?

I tried the Q4_K_M here: https://huggingface.co/seyf1elislam/Monarch-7B-GGUF

It worked great so this is probably the best quantized 7B model!

Oh that's funny, I changed the name of this model to Beagle4 (https://huggingface.co/mlabonne/Beagle4). This is not the same model as this new Monarch.

Sign up or log in to comment