## This model is a merge of LLAMA-13b and SuperCOT LoRA

[huggyllama/llama-13b](https://huggingface.co/huggyllama/llama-13b) + [kaiokendev/SuperCOT-LoRA/13b/gpu/cutoff-2048](https://huggingface.co/kaiokendev/SuperCOT-LoRA)

CUDA_VISIBLE_DEVICES=0 python llama.py c4 --wbits 4 --true-sequential --act-order --groupsize 128


In ooba make sure to use --groupsize 128 --wbits 4