量化报错的问题，移动量化逻辑前后代码都尝试了还是

[moonlight123154]

Traceback (most recent call last):
File "D:\big_model\chaglm_main\lora3.py", line 66, in
model2 = PeftModel.from_pretrained(model, peft_model_id2)
File "D:\anaconda3\envs\zhipu\lib\site-packages\peft\peft_model.py", line 352, in from_pretrained
model = MODEL_TYPE_TO_PEFT_MODEL_MAPPING[config.task_type](model, config, adapter_name)
File "D:\anaconda3\envs\zhipu\lib\site-packages\peft\peft_model.py", line 1059, in init
super().init(model, peft_config, adapter_name)
File "D:\anaconda3\envs\zhipu\lib\site-packages\peft\peft_model.py", line 126, in init
self.base_model = cls(model, {adapter_name: peft_config}, adapter_name)
File "D:\anaconda3\envs\zhipu\lib\site-packages\peft\tuners\lora\model.py", line 111, in init
super().init(model, config, adapter_name)
File "D:\anaconda3\envs\zhipu\lib\site-packages\peft\tuners\tuners_utils.py", line 147, in init
self.inject_adapter(self.model, adapter_name)
File "D:\anaconda3\envs\zhipu\lib\site-packages\peft\tuners\tuners_utils.py", line 302, in inject_adapter
self._create_and_replace(peft_config, adapter_name, target, target_name, parent, current_key=key)
File "D:\anaconda3\envs\zhipu\lib\site-packages\peft\tuners\lora\model.py", line 182, in _create_and_replace
new_module = self._create_new_module(lora_config, adapter_name, target, **kwargs)
File "D:\anaconda3\envs\zhipu\lib\site-packages\peft\tuners\lora\model.py", line 263, in _create_new_module
raise ValueError(
ValueError: Target module QuantizedLinear() is not supported. Currently, only the following modules are supported: torch.nn.Linear, torch.nn.Embedding, torch.nn.Conv2d, transformers.pytorch_utils.Conv1D.

请问这个怎么解决
还想请教一下微调之后的模型可以再量化吗

[zRzRzRzRzRzRzR]

已经修复，可以查看最新的代码