Error loading finetuned llama2 model while running inference

Could you share how your model.tar.gz looks?