How to use Incomple/Llama-3.1-8B-Instruct_dpo_sg_values with PEFT:
from peft import PeftModel from transformers import AutoModelForCausalLM base_model = AutoModelForCausalLM.from_pretrained("meta-llama/Llama-3.1-8B-Instruct") model = PeftModel.from_pretrained(base_model, "Incomple/Llama-3.1-8B-Instruct_dpo_sg_values")
Incomple commited on 