Error loading finetuned llama2 model while running inference

try with ml.g5.12xlarge according to AWS guide.