运行Gensim文本挖掘代码时VS Code Jupyter内核崩溃求助

问题详情 #

运行Gensim开展文本挖掘时，VS Code的Jupyter内核突然崩溃，返回错误码3221225477，核心报错信息：

Canceled future for execute_request message before replies were done

已尝试卸载重装pyzmq、traitlets、jupyter-client，问题依旧；网络搜索数小时未找到有效解决方案。

Conda环境配置 #

backcall                      0.2.0
backports.functools-lru-cache 1.6.4
certifi                       2022.6.15
colorama                      0.4.5
Cython                        0.29.28
debugpy                       1.6.0
decorator                     5.1.1
entrypoints                   0.4
gensim                        4.2.0
ipykernel                     5.5.0
ipython                       7.34.0
ipython-genutils              0.2.0
jedi                          0.17.2
jupyter-client                6.1.10
jupyter-core                  4.11.1
matplotlib-inline             0.1.3
nest-asyncio                  1.5.5
numpy                         1.21.6
packaging                     21.3
parso                         0.7.1
pickleshare                   0.7.5
pip                           22.1.2
prompt-toolkit                3.0.30
psutil                        5.9.1
Pygments                      2.12.0
pyparsing                     3.0.9
python-dateutil               2.8.2
pywin32                       304
pyzmq                         19.0.2
scipy                         1.7.3
setuptools                    63.4.2
six                           1.16.0
smart-open                    6.0.0
tornado                       6.2
traitlets                     4.3.3
wcwidth                       0.2.5
wheel                         0.37.1
wincertstore                  0.2

Jupyter日志信息 #

error 10:30:28.238: Disposing session as kernel process died ExitCode: 3221225477, Reason: 
info 10:30:28.239: kill daemon
error 10:30:28.240: Raw kernel process exited code: 3221225477
error 10:30:28.245: Error in waiting for cell to complete [Error: Canceled future for execute_request message before replies were done
    at t.KernelShellFutureHandler.dispose (c:\Users\xxxx\.vscode\extensions\ms-toolsai.jupyter-2022.7.1002181843\out\extension.node.js:2:32353)
    at c:\Users\xxxx\.vscode\extensions\ms-toolsai.jupyter-2022.7.1002181843\out\extension.node.js:2:51405
    at Map.forEach (<anonymous>)
    at y._clearKernelState (c:\Users\xxxx\.vscode\extensions\ms-toolsai.jupyter-2022.7.1002181843\out\extension.node.js:2:51390)
    at y.dispose (c:\Users\xxxx\.vscode\extensions\ms-toolsai.jupyter-2022.7.1002181843\out\extension.node.js:2:44872)
    at c:\Users\xxxx\.vscode\extensions\ms-toolsai.jupyter-2022.7.1002181843\out\extension.node.js:2:2320856
    at t.swallowExceptions (c:\Users\xxxx\.vscode\extensions\ms-toolsai.jupyter-2022.7.1002181843\out\extension.node.js:7:118974)
    at dispose (c:\Users\xxxx\.vscode\extensions\ms-toolsai.jupyter-2022.7.1002181843\out\extension.node.js:2:2320834)
    at t.RawSession.dispose (c:\Users\xxxx\.vscode\extensions\ms-toolsai.jupyter-2022.7.1002181843\out\extension.node.js:2:2325771)
    at processTicksAndRejections (node:internal/process/task_queues:96:5)]
warn 10:30:28.246: Cell completed with errors {
  message: 'Canceled future for execute_request message before replies were done'
}
info 10:30:28.248: Cancel all remaining cells true || Error || undefined
info 10:30:28.248: Cancel pending cells
info 10:30:28.250: Cell 22 executed with state Error

可尝试的解决方案 #

1. 升级Jupyter相关依赖版本 #

当前ipykernel、jupyter-client、pyzmq版本偏旧，存在兼容性风险，尝试升级到稳定兼容版本：

conda update ipykernel jupyter-client pyzmq

或指定版本安装：

conda install ipykernel==6.15.0 jupyter-client==7.4.2 pyzmq==23.2.0

2. 排查内存占用问题 #

错误码3221225477对应Windows的内存访问违规，文本挖掘任务（如Word2Vec、LDA训练）易占用大量内存，尝试：

• 缩小单次处理的文本数据规模
• 降低模型参数维度（如Word2Vec的vector_size参数）
• 关闭其他占用内存的程序，释放系统资源

3. 创建全新隔离环境 #

避免现有环境依赖冲突，创建新conda环境仅安装必要组件：

conda create -n gensim_env python=3.8
conda activate gensim_env
conda install gensim jupyter ipykernel
python -m ipykernel install --user --name gensim_env --display-name "Python (gensim_env)"

之后在VS Code中切换至该新内核运行代码。

4. 重装底层依赖库 #

若内存访问违规由numpy、scipy等库的二进制文件兼容问题导致，尝试重装：

conda uninstall scipy numpy
conda install scipy numpy

5. 调整VS Code扩展设置 #

关闭Jupyter扩展的「逐行运行」功能，或暂时禁用其他可能冲突的扩展，重启VS Code后测试。

内容的提问来源于stack exchange，提问作者Arthur