NLTK "averaged_perceptron_tagger_eng" Loading Error When Using DirectoryLoader

[rdyan0053]

I'm following the documentation at [Milvus RAG with vLLM](https://milvus.io/docs/zh/milvus_rag_with_vllm.md) to build a RAG system using Milvus, vLLM, and Llama 3.1. When I execute the following code:

from langchain_community.document_loaders import DirectoryLoader
import pprint

path = "../../RAG/rtdocs_new/"
global_pattern = '*.html'
loader = DirectoryLoader(path=path, glob=global_pattern)
docs = loader.load()

I encounter the error:

[nltk_data] Error loading averaged_perceptron_tagger_eng: <urlopen error [Errno 111] Connection refused>

It appears that NLTK is attempting to download the averaged_perceptron_tagger_eng model and failing due to connection issues. How can I resolve this error so that the DirectoryLoader works as expected in my environment with Chinese mainland?
Any help or guidance would be greatly appreciated. Thank you!

[yhmo]

I have no idea. Some links that might be helpful:
langchain-ai/langchain#27701 (comment)
https://blog.csdn.net/qq_34529292/article/details/145714915

[xiaofan-luan]

It seems NLTK is trying to download the averaged_perceptron_tagger_eng model but fails due to connection restrictions

maybe you can try with VPN or predownload it in a environemnt