在 docker 容器中运行时加载 torch 和 Sentence_transformers 的预训练模型失败

Question

在加载 torch 的预训练模型以及sentence_transformers("distilbert-base-nli-stsb-mean-tokens")尝试在 docker 容器中运行时，我遇到了以下错误。

Error: Invalid value for '-A' / '--app': 
 Unable to load celery application.
 While trying to load the module app.celery the following error occurred:
 Traceback (most recent call last):
   File "/usr/local/lib/python3.8/site-packages/celery/bin/celery.py", line 53, in convert
     return find_app(value)
   File "/usr/local/lib/python3.8/site-packages/celery/app/utils.py", line 384, in find_app
     sym = symbol_by_name(app, imp=imp)
   File "/usr/local/lib/python3.8/site-packages/kombu/utils/imports.py", line 56, in symbol_by_name
     module = imp(module_name, package=package, **kwargs)
   File "/usr/local/lib/python3.8/site-packages/celery/utils/imports.py", line 100, in import_from_cwd
     return imp(module, package=package)
   File "/usr/local/lib/python3.8/importlib/__init__.py", line 127, in import_module
     return _bootstrap._gcd_import(name[level:], package, level)
   File "<frozen importlib._bootstrap>", line 1014, in _gcd_import
   File "<frozen importlib._bootstrap>", line 991, in _find_and_load
   File "<frozen importlib._bootstrap>", line 975, in _find_and_load_unlocked
   File "<frozen importlib._bootstrap>", line 671, in _load_unlocked
   File "<frozen importlib._bootstrap_external>", line 783, in exec_module
   File "<frozen importlib._bootstrap>", line 219, in _call_with_frames_removed
   File "/code/app.py", line 997, in <module>
     load_model()
   File "/code/app.py", line 255, in load_model
     embedder = SentenceTransformer('distilbert-base-nli-stsb-mean-tokens')
   File "/usr/local/lib/python3.8/site-packages/sentence_transformers/SentenceTransformer.py", line 48, in __init__
     os.makedirs(model_path, exist_ok=True)
   File "/usr/local/lib/python3.8/os.py", line 213, in makedirs
     makedirs(head, exist_ok=exist_ok)
   File "/usr/local/lib/python3.8/os.py", line 213, in makedirs
     makedirs(head, exist_ok=exist_ok)
   File "/usr/local/lib/python3.8/os.py", line 213, in makedirs
     makedirs(head, exist_ok=exist_ok)
   [Previous line repeated 1 more time]
   File "/usr/local/lib/python3.8/os.py", line 223, in makedirs
     mkdir(name, mode)
 PermissionError: [Errno 13] Permission denied: '/nonexistent'

这里说创建文件夹时出现权限被拒绝错误。但我尝试USER root在Dockerfile. 被这个问题困扰了好久。请任何人在这里帮助我。

更新： 我的 Dockerfile：

FROM python:3.8.5-slim

WORKDIR /code

ENV ENVIRONMENT='LOCAL'
ENV FLASK_APP=app.py
ENV FLASK_RUN_HOST=0.0.0.0
ENV PYTHONDONTWRITEBYTECODE 1
ENV PYTHONUNBUFFERED 1

RUN apt-get update && apt-get install -y sudo netcat apt-utils
RUN apt-get install -y python3-dev  build-essential python3-pip

COPY ./requirements_local.txt /code/requirements_local.txt
RUN pip install -r /code/requirements_local.txt

EXPOSE 8000
COPY . /code/

CMD [ "gunicorn", "app:app", "-b", "0.0.0.0:8000","--timeout","7200"]

Docker-compose：

services:
  web:
    build: 
      context: .
      dockerfile: ./Dockerfile.prod
    hostname: flaskapp
    env_file:
      - ./.env.prod
    links:
      - redis
      - celery
    depends_on:
      - redis
    volumes:
      - data:/code
      - type: bind
        source: /home/ubuntu/models
        target: /mnt/models

Answer 1

Sentence-transformers 下载模型并将其存储在 ~/.cache 目录中（或者 cache_folder 评估的任何内容 - https://github.com/UKPLab/sentence-transformers/blob/a13a4ec98b8fdda83855aca7992ea793444a207f/sentence_transformers/SentenceTransformer.py#L63）。对于您来说，这看起来像 /nonexistant 目录。权限被拒绝错误表明您没有访问该目录（创建缓存文件夹）的权限。

您可以修改 Dockerfile 以创建此目录并使其可供任何需要访问此目录的用户访问：-

RUN mkdir ~/.cache
RUN chmod -R 777 ~/.cache # don't do this in production - modify command to give permission to users who require it.

或者您可以尝试在 Dockerfile 本身中下载模型 -

RUN python -c 'from sentence-transformers import SentenceTransformer; embedder = SentenceTransformer("distilbert-base-nli-stsb-mean-tokens")'