PaddlePaddle / PaddleNLP

👑 Easy-to-use and powerful NLP and LLM library with 🤗 Awesome model zoo, supporting wide-range of NLP tasks from research to industrial applications, including 🗂Text Classification, 🔍 Neural Search, ❓ Question Answering, ℹ️ Information Extraction, 📄 Document Intelligence, 💌 Sentiment Analysis etc.
https://paddlenlp.readthedocs.io
Apache License 2.0
11.73k stars 2.86k forks source link

windows环境下pipelines的quesion-answering、semantic-search运行出现Segmentation fault #3398

Closed butear closed 1 year ago

butear commented 1 year ago

软件环境

-win10
-python 3.7
-CUDA 11.2  
-cudnn 8.4.0
-paddlepaddle-gpu 2.3
-paddlenlp 2.4.0

重复问题

错误描述

Traceback (most recent call last):
  File "D:\Aaron\PaddleNLP\pipelines\ui\webapp_question_answering.py", line 178, in main
    top_k_retriever=top_k_retriever)
  File "D:\Aaron\PaddleNLP\pipelines\ui\utils.py", line 94, in query
    response_raw = requests.post(url, json=req)
  File "C:\Dev\Python\lib\site-packages\requests\api.py", line 117, in post
    return request('post', url, data=data, json=json, **kwargs)
  File "C:\Dev\Python\lib\site-packages\requests\api.py", line 61, in request
    return session.request(method=method, url=url, **kwargs)
  File "C:\Dev\Python\lib\site-packages\requests\sessions.py", line 542, in request
    resp = self.send(prep, **send_kwargs)
  File "C:\Dev\Python\lib\site-packages\requests\sessions.py", line 655, in send
    r = adapter.send(request, **kwargs)
  File "C:\Dev\Python\lib\site-packages\requests\adapters.py", line 498, in send
    raise ConnectionError(err, request=request)
requests.exceptions.ConnectionError: ('Connection aborted.', ConnectionResetError(10054, '远程主了一个现有的连接。', None, 10054, None))

INFO:     Application startup complete.
09/30/2022 09:18:57 AM Application startup complete.
INFO:     Uvicorn running on http://0.0.0.0:8891 (Press CTRL+C to quit)
09/30/2022 09:18:57 AM Uvicorn running on http://0.0.0.0:8891 (Press CTRL+C to quit)
INFO:     127.0.0.1:65016 - "GET /hs_version HTTP/1.1" 200 OK
INFO:     127.0.0.1:65017 - "GET /initialized HTTP/1.1" 200 OK
Segmentation fault

稳定复现步骤 & 代码

export PIPELINE_YAML_PATH=rest_api/pipeline/dense_qa.yaml python rest_api/application.py 8891

unset http_proxy && unset https_proxy export API_ENDPOINT=http://127.0.0.1:8891 python -m streamlit run ui/webapp_question_answering.py --server.port 8502

w5688414 commented 1 year ago

看上去是cuda版本不匹配。

Please follow this windows README.md. https://github.com/PaddlePaddle/PaddleNLP/blob/develop/pipelines/examples/question-answering/Install_windows.md

and the windows video tutorial, please refer to this link. https://www.bilibili.com/video/BV1DY4y1M7HE/?zw

butear commented 1 year ago

在以下环境问题解决: -win10 -python 3.7 -CUDA 11.2
-cudnn 8.2.1 -paddlepaddle-gpu 2.2.2 -paddlenlp 2.4.0