chatglm2报错：ValueError: weight is on the meta device, we need a `value` to put in on 0

[angel1288]

Traceback (most recent call last): File "/data3/push_recall/LLM-Tuning/chatglm2_lora_tuning.py", line 153, in main() File "/data3/push_recall/LLM-Tuning/chatglm2_lora_tuning.py", line 93, in main model = AutoModel.from_pretrained( File "/data3/env/miniconda3/envs/baichuan/lib/python3.9/site-packages/transformers/models/auto/auto_factory.py", line 462, in from_pretrained return model_class.from_pretrained( File "/data3/env/miniconda3/envs/baichuan/lib/python3.9/site-packages/transformers/modeling_utils.py", line 2828, in from_pretrained dispatch_model(model, device_map=device_map, offload_dir=offload_folder, offload_index=offload_index) File "/data3/env/miniconda3/envs/baichuan/lib/python3.9/site-packages/accelerate/big_modeling.py", line 370, in dispatch_model attach_align_device_hook_on_blocks( File "/data3/env/miniconda3/envs/baichuan/lib/python3.9/site-packages/accelerate/hooks.py", line 478, in attach_align_device_hook_on_blocks add_hook_to_module(module, hook) File "/data3/env/miniconda3/envs/baichuan/lib/python3.9/site-packages/accelerate/hooks.py", line 155, in add_hook_to_module module = hook.init_hook(module) File "/data3/env/miniconda3/envs/baichuan/lib/python3.9/site-packages/accelerate/hooks.py", line 251, in init_hook set_module_tensor_to_device(module, name, self.execution_device) File "/data3/env/miniconda3/envs/baichuan/lib/python3.9/site-packages/accelerate/utils/modeling.py", line 140, in set_module_tensor_to_device raise ValueError(f"{tensor_name} is on the meta device, we need a value to put in on {device}.") ValueError: weight is on the meta device, we need a value to put in on 0. 请问这个什么导致的？谢谢哈~

[beyondguo]

这个好像是个accelerate包的一些问题，你可以先自行查阅一下，比如https://discuss.huggingface.co/t/meta-device-error-while-instantiating-model/33402

另外你使用的各种包的版本跟我的一致吗？你的device是什么？

[zhohn0]

同遇到这个问题，单卡

[Mike-ihr]

同遇到这个问题，单卡3090，请问该如何解决