如何解决含py-langchain、llama-index、gradio的Python项目中transformers包未找到错误

问题说明 #

执行pip install transformers命令后，运行代码仍触发ModuleNotFoundError: No module named 'transformers'错误，项目依赖库包括py-langchain、llama-index、gradio。

解决方案 #

• 检查Python与pip环境一致性
  确认运行代码的Python解释器和执行pip install的pip属于同一环境：

  # Linux/macOS查看路径
  which python
  which pip
  # Windows查看路径
  where python
  where pip
  

  若路径不匹配，使用对应环境的pip重新安装，例如：

  python -m pip install transformers
  # 针对Python3可使用
  pip3 install transformers
  

• 验证transformers安装状态
  在Python交互终端中直接测试导入：

  import transformers
  print(transformers.__version__)
  

  若仍报错，强制重新安装：

  pip install --force-reinstall transformers
  

• 确认虚拟环境激活状态
  若使用虚拟环境，需确保运行代码前已激活对应环境：

  # Linux/macOS激活虚拟环境
  source venv/bin/activate
  # Windows激活虚拟环境
  venv\Scripts\activate
  

  激活后重新执行pip install transformers。

• 安装完整项目依赖
  代码中虽注释了transformers的导入，但py-langchain或llama-index可能间接依赖该库，可一次性安装所有所需依赖：

  pip install py-langchain llama-index gradio transformers
  

用户提供代码 #

import os
import sys
from dotenv import load_dotenv
import gradio as gr
from langchain import OpenAI
from llama_index import SimpleDirectoryReader, GPTListIndex, GPTVectorStoreIndex, LLMPredictor, PromptHelper
##from transformers import pipeline

load_dotenv()

os.environ['OPENAI_API_KEY'] = os.getenv('OPENAI_API_KEY') 

def construct_index(directory_path):
    max_input_size = 4096
    num_outputs = 512
    max_chunk_overlap = 20
    chunk_size_limit = 600

    prompt_helper = PromptHelper(max_input_size, num_outputs, max_chunk_overlap, chunk_size_limit=chunk_size_limit)

    llm_predictor = LLMPredictor(llm=OpenAI(temperature=0.7, model_name="text-davinci-003", max_tokens=num_outputs))

    documents = SimpleDirectoryReader(directory_path).load_data()

    index = GPTVectorStoreIndex(documents, llm_predictor=llm_predictor, prompt_helper=prompt_helper)

    index.save_to_disk('index.json')

    return index

def qabot(input_text):
    index = GPTVectorStoreIndex.load_from_disk('index.json')
    response = index.query(input_text, response_mode="compact")
    return response.response

iface = gr.Interface(fn=qabot, inputs=gr.inputs.Textbox(lines=7, label='Enter you query'), outputs="text", title="Custom-trained QA Application")

index = construct_index("docs")

iface.launch(share=True)

内容的提问来源于stack exchange，提问作者Antonio Memoli