重装TensorFlow后仍无法导入，Jupyter/VSCode内核崩溃求助

问题概述 #

在Jupyter Notebook和VSCode中导入TensorFlow时，内核频繁崩溃，报错：

Canceled future for execute_request message before replies were done
The Kernel crashed while executing code in the the current cell or a previous cell. Please review the code in the cell(s) to identify a possible cause of the failure. Click here for more info. View Jupyter log for further details.

已尝试通过pip install tensorflow和python3 -m pip install tensorflow卸载重装TensorFlow，无效；TensorFlow在Colab中运行正常，本地Anaconda环境下TensorFlow 2.12.0安装在/Users/hanyusu/opt/anaconda3/lib/python3.9/site-packages，依赖项齐全。

报错日志 #

15:55:19.800 [error] Error in waiting for cell to complete [Error: Canceled future for execute_request message before replies were done
    at t.KernelShellFutureHandler.dispose (~/.vscode/extensions/ms-toolsai.jupyter-2023.4.1011241018-darwin-arm64/out/extension.node.js:2:32419)
    at ~/.vscode/extensions/ms-toolsai.jupyter-2023.4.1011241018-darwin-arm64/out/extension.node.js:2:51471
    at Map.forEach (<anonymous>)
    at v._clearKernelState (~/.vscode/extensions/ms-toolsai.jupyter-2023.4.1011241018-darwin-arm64/out/extension.node.js:2:51456)
    at v.dispose (~/.vscode/extensions/ms-toolsai.jupyter-2023.4.1011241018-darwin-arm64/out/extension.node.js:2:44938)
    at ~/.vscode/extensions/ms-toolsai.jupyter-2023.4.1011241018-darwin-arm64/out/extension.node.js:24:105531
    at te (~/.vscode/extensions/ms-toolsai.jupyter-2023.4.1011241018-darwin-arm64/out/extension.node.js:2:1587099)
    at Zg.dispose (~/.vscode/extensions/ms-toolsai.jupyter-2023.4.1011241018-darwin-arm64/out/extension.node.js:24:105507)
    at nv.dispose (~/.vscode/extensions/ms-toolsai.jupyter-2023.4.1011241018-darwin-arm64/out/extension.node.js:24:112790)
    at process.processTicksAndRejections (node:internal/process/task_queues:96:5)]
15:55:19.801 [warn] Cell completed with errors {
  message: 'Canceled future for execute_request message before replies were done'
}
15:55:19.801 [warn] Cancel all remaining cells due to cancellation or failure in execution

解决方案 #

1. 用Anaconda重新安装TensorFlow，避免依赖冲突 #

卸载现有TensorFlow：

pip uninstall -y tensorflow

通过conda安装匹配版本的TensorFlow，conda会自动处理依赖兼容性：

conda install tensorflow=2.12.0

安装完成后，先在终端验证环境是否正常：

python -c "import tensorflow as tf; print(tf.__version__)"

如果终端能正常输出版本号，说明环境本身无问题，问题出在Jupyter/VSCode的内核配置。

2. 重置并关联正确的Jupyter内核 #

确保Anaconda环境的内核被正确注册到Jupyter：

python -m ipykernel install --user --name=base

（如果你的Anaconda环境不是base，替换为对应环境名）

重启VSCode/Jupyter，在右下角切换内核到你的Anaconda环境（路径为/Users/hanyusu/opt/anaconda3/bin/python），然后执行Kernel -> Restart & Clear Output清除缓存。

3. 检查系统资源与GPU内存配置 #

如果是M系列Mac，TensorFlow启动时可能因GPU内存分配问题崩溃，导入前添加内存增长配置：

import tensorflow as tf
gpus = tf.config.list_physical_devices('GPU')
if gpus:
    try:
        tf.config.experimental.set_memory_growth(gpus[0], True)
    except RuntimeError as e:
        print(e)

如果是CPU环境，关闭其他占用大量内存的进程，确保TensorFlow有足够内存启动。

4. 排查依赖版本冲突 #

执行conda list查看环境中所有包的版本，重点检查以下TF核心依赖：

• numpy：需与TensorFlow 2.12.0兼容（推荐1.21.x-1.23.x版本）
• protobuf：需为3.20.x或更低版本（TF 2.12.0不支持过高版本的protobuf）
• keras：需与TF版本一致（TF 2.12.0对应keras 2.12.0）

如果发现版本不匹配，手动调整：

conda install numpy=1.23.5 protobuf=3.20.3 keras=2.12.0

内容的提问来源于stack exchange，提问作者Hanyu Su