MEVZU N° TAG / VOL. 130
0 blog · 0 news · 1 wiki
A LoRA variant combined with quantisation that lets you fine-tune 65B models on a single consumer GPU.