LlamaIndex报错

1.  首次运行，from pandas.core import ( 卡住，中断后报错

File "/data/wangzhiwen/anaconda3/lib/python3.10/site-packages/llama_index/legacy/utils.py", line 65, in __init__
    nltk.data.find("tokenizers/punkt", paths=[self._nltk_data_dir])
  File "/data/wangzhiwen/anaconda3/lib/python3.10/site-packages/nltk/data.py", line 583, in find
    raise LookupError(resource_not_found)
LookupError: 
**********************************************************************
  Resource punkt not found.
  Please use the NLTK Downloader to obtain the resource:

  >>> import nltk
  >>> nltk.download('punkt')
  
  For more information see: https://www.nltk.org/data.html

  Attempted to load tokenizers/punkt

  Searched in:
    - '/data/wangzhiwen/anaconda3/lib/python3.10/site-packages/llama_index/legacy/_static/nltk_cache'
**********************************************************************

原因：无网络，./anaconda3/lib/python3.10/site-packages/llama_index/legacy/_static/nltk_cache/tokenizers/punkt.zip包不能下载并解压，需手动下载后在该目录解压。 

2. 

ValueError: The current `device_map` had weights offloaded to the disk. Please provide an `offload_folder` for them. Alternatively, make sure you have `safetensors` installed if the model you are using offers the weights in this format.

原因：'max_memory': {1: '10GiB', 2: '10GiB'}内存设置不够。

3.  自定义LLM报错RuntimeError: "topk_cpu" not implemented for 'Half'

原因：输出在cpu上，将输入放至GPU即可（.cuda()）