qwp4w3hyb

/

Meta-Llama-3.1-70B-Instruct-iMat-GGUF

like 7

Tested the Q4_K_L with https://github.com/Nexesenex/kobold.cpp/releases/tag/v1.71013_b3455%2B9 and ~24k context. The first time Llama 3.1 70B doesn't shit the bed immediately.

Looks good at first glance.