GGUF
#1
by
SilverFan
- opened
Thanks for your work. Your models have brought much joy.
I have Q5_K_M & Q6_K here: https://huggingface.co/SilverFan/WizardIceLemonTeaRP-32k-GGUF
I'm sure someone will do the full set with imatrix soon
I have mix feelings with this merge result.
Seems okay to me. But i'm just using 8k context and default samplers with 0.2 smoothing factor
It's hard to describe, it's giving better for RP responses but less smart, if char need to think-deceive, result gives mix feelings. I will use it as merge model.