from transformers import AutoModelForCausalLM, AutoProcessor, BitsAndBytesConfig
from peft import PeftModel
import torch

base = "Qwen/Qwen3.8-27B"
adapter = "shikunpunk/Qwen3.8-27B-GuCheng-KTO"

quant = BitsAndBytesConfig(load_in_4bit=True, bnb_4bit_compute_dtype=torch.bfloat16,
                           bnb_4bit_quant_type="nf4", bnb_4bit_use_double_quant=True)
model = AutoModelForCausalLM.from_pretrained(base, trust_remote_code=True,
                                             torch_dtype=torch.bfloat16, device_map="auto",
                                             quantization_config=quant)
model = PeftModel.from_pretrained(model, adapter, is_trainable=False)
processor = AutoProcessor.from_pretrained(base, trust_remote_code=True)
model.eval()

说明

该 adapter 是风格保真 KTO 后训练产物，配合 SFT 风格 adapter 使用（SFT 先学风格，KTO 强化偏好对齐）。
相关代码与数据见项目 GitHub 仓库。