使用lora和glm2微调合并后，报错。

[hOVOzw]

使用lora和glm2微调合并后，报错。
使用llama factory对模型微调后导出模型，已经在model_config.py修改了配置。

**预期的结果 /正常打开webui

**实际结果 /
2024-02-18 01:39:12 | ERROR | stderr | Traceback (most recent call last):
2024-02-18 01:39:12 | ERROR | stderr | File "D:\ANACONDA\envs\chatchat\lib\multiprocessing\process.py", line 315, in _bootstrap
2024-02-18 01:39:12 | ERROR | stderr | self.run()
2024-02-18 01:39:12 | ERROR | stderr | File "D:\ANACONDA\envs\chatchat\lib\multiprocessing\process.py", line 108, in run
2024-02-18 01:39:12 | ERROR | stderr | self._target(*self._args, **self._kwargs)
2024-02-18 01:39:12 | ERROR | stderr | File "D:\langchain_Chatchat\startup.py", line 389, in run_model_worker
2024-02-18 01:39:12 | ERROR | stderr | app = create_model_worker_app(log_level=log_level, **kwargs)
2024-02-18 01:39:12 | ERROR | stderr | File "D:\langchain_Chatchat\startup.py", line 217, in create_model_worker_app
2024-02-18 01:39:12 | ERROR | stderr | worker = ModelWorker(
2024-02-18 01:39:12 | ERROR | stderr | File "D:\ANACONDA\envs\chatchat\lib\site-packages\fastchat\serve\model_worker.py", line 77, in init
2024-02-18 01:39:12 | ERROR | stderr | self.model, self.tokenizer = load_model(
2024-02-18 01:39:12 | ERROR | stderr | File "D:\ANACONDA\envs\chatchat\lib\site-packages\fastchat\model\model_adapter.py", line 348, in load_model
2024-02-18 01:39:12 | ERROR | stderr | model, tokenizer = adapter.load_model(model_path, kwargs)
2024-02-18 01:39:12 | ERROR | stderr | File "D:\ANACONDA\envs\chatchat\lib\site-packages\fastchat\model\model_adapter.py", line 823, in load_model
2024-02-18 01:39:12 | ERROR | stderr | tokenizer = AutoTokenizer.from_pretrained(
2024-02-18 01:39:12 | ERROR | stderr | File "D:\ANACONDA\envs\chatchat\lib\site-packages\transformers\models\auto\tokenization_auto.py", line 801, in from_pretrained
2024-02-18 01:39:12 | ERROR | stderr | return tokenizer_class.from_pretrained(pretrained_model_name_or_path, *inputs, **kwargs)
2024-02-18 01:39:12 | ERROR | stderr | File "D:\ANACONDA\envs\chatchat\lib\site-packages\transformers\tokenization_utils_base.py", line 2029, in from_pretrained
2024-02-18 01:39:12 | ERROR | stderr | return cls._from_pretrained(
2024-02-18 01:39:12 | ERROR | stderr | File "D:\ANACONDA\envs\chatchat\lib\site-packages\transformers\tokenization_utils_base.py", line 2261, in _from_pretrained
2024-02-18 01:39:12 | ERROR | stderr | tokenizer = cls(*init_inputs, **init_kwargs)
2024-02-18 01:39:12 | ERROR | stderr | File "C:\Users\Administrator.cache\huggingface\modules\transformers_modules\chatglm2-lora\tokenization_chatglm.py", line 78, in init
2024-02-18 01:39:12 | ERROR | stderr | super().init(padding_side=padding_side, clean_up_tokenization_spaces=clean_up_tokenization_spaces, **kwargs)
2024-02-18 01:39:12 | ERROR | stderr | File "D:\ANACONDA\envs\chatchat\lib\site-packages\transformers\tokenization_utils.py", line 363, in init
2024-02-18 01:39:12 | ERROR | stderr | super().init(**kwargs)
2024-02-18 01:39:12 | ERROR | stderr | File "D:\ANACONDA\envs\chatchat\lib\site-packages\transformers\tokenization_utils_base.py", line 1603, in init
2024-02-18 01:39:12 | ERROR | stderr | super().init(**kwargs)
2024-02-18 01:39:12 | ERROR | stderr | File "D:\ANACONDA\envs\chatchat\lib\site-packages\transformers\tokenization_utils_base.py", line 861, in init
2024-02-18 01:39:12 | ERROR | stderr | setattr(self, key, value)
2024-02-18 01:39:12 | ERROR | stderr | AttributeError: can't set attribute
Loading checkpoint shards: 0%| | 0/7 [00:00<?, ?it/s]
Loading checkpoint shards: 14%|████████▏ | 1/7 [00:00<00:05, 1.00it/s]
Loading checkpoint shards: 29%|████████████████▎ | 2/7 [00:01<00:04, 1.12it/s]
Loading checkpoint shards: 43%|████████████████████████▍ | 3/7 [00:02<00:03, 1.30it/s]
Loading checkpoint shards: 57%|████████████████████████████████▌ | 4/7 [00:03<00:02, 1.45it/s]
Loading checkpoint shards: 71%|████████████████████████████████████████▋ | 5/7 [00:03<00:01, 1.38it/s]
Loading checkpoint shards: 86%|████████████████████████████████████████████████▊ | 6/7 [00:04<00:00, 1.28it/s]
Loading checkpoint shards: 100%|█████████████████████████████████████████████████████████| 7/7 [00:05<00:00, 1.43it/s]
Loading checkpoint shards: 100%|█████████████████████████████████████████████████████████| 7/7 [00:05<00:00, 1.34it/s]
2024-02-18 01:39:17 | ERROR | stderr |
2024-02-18 01:39:20 | INFO | model_worker | Register to controller

[dosubot]

🤖

你好，@hOVOzw！很高兴再次见到你。希望你近期一切都好。

根据您提供的错误日志，问题似乎出在加载模型的过程中。错误信息显示在尝试设置属性时出现了问题，这可能是由于模型或分词器的版本不兼容，或者模型文件的路径不正确。

在解决这个问题时，您可以参考以下几点：

1. 确保MODEL_PATH字典中的模型路径正确。如果您使用的是本地模型，路径应指向您机器上模型的正确位置。如果您使用的是Hugging Face的模型，确保模型名称正确。

2. 检查您使用的模型是否与Langchain-Chatchat应用程序兼容。并非所有模型都可能与应用程序兼容。

3. 如果您使用GPU运行模型（EMBEDDING_DEVICE或LLM_DEVICE设置为"cuda"），确保您有足够的GPU内存。一些模型需要大量的GPU内存，如果不足可能会导致错误。

4. 检查ONLINE_LLM_MODEL中在线模型的API密钥是否正确，API是否正常工作。

5. 如果您将它们作为代理模型使用，确保您使用的模型包含在SUPPORT_AGENT_MODEL列表中。

如果错误仍然存在，提供确切的错误消息或您所做修改的更多详细信息将会有所帮助。

此外，我在Langchain-Chatchat的存储库中找到了一些类似的已解决问题，可能对您有所帮助：

1. 执行python startup.py -a 报错
2. [BUG] AttributeError: 'ChatGLMTokenizer' object has no attribute 'sp_tokenizer'
3. python startup.py -a出错

这些问题的解决方案可能对您的问题有所启示。

Sources

• 执行python startup.py -a 报错
• [BUG] AttributeError: 'ChatGLMTokenizer' object has no attribute 'sp_tokenizer'
• python startup.py -a出错
• configs/model_config.py.example

About Dosu

This response is meant to be useful and save you time. It is not meant to be a precise solution, but rather a starting point for your own research.

Help me be more useful! Please leave a 👍 if this is helpful and 👎 if it is irrelevant.

If you want to continue the conversation, start your reply with @dosu-bot.

[zRzRzRzRzRzRzR]

建议使用peft加载，分别挂载微调和原始的模型，wiki中有相关内容