Jupyter Notebook导入TensorFlow报错：找不到tensorflow.security模块

问题场景 #

为让TensorFlow使用GPU而非CPU运行，卸载原有版本后仅安装tensorflow-gpu包，但在Jupyter Notebook中执行import tensorflow as tf时触发以下报错：

ModuleNotFoundError                       Traceback (most recent call last)
Cell In[2], line 1
----> 1 import tensorflow as tf

File ~\AppData\Roaming\Python\Python310\site-packages\tensorflow\__init__.py:38
     35 import sys as _sys
     36 import typing as _typing
---> 38 from tensorflow.python.tools import module_util as _module_util
     39 from tensorflow.python.util.lazy_loader import LazyLoader as _LazyLoader
     41 # Make sure code inside the TensorFlow codebase can use tf2.enabled() at import.

File ~\AppData\Roaming\Python\Python310\site-packages\tensorflow\python\__init__.py:45
     43 from tensorflow.python import distribute
     44 # from tensorflow.python import keras
---> 45 from tensorflow.python.feature_column import feature_column_lib as feature_column
     46 # from tensorflow.python.layers import layers
     47 from tensorflow.python.module import module

File ~\AppData\Roaming\Python\Python310\site-packages\tensorflow\python\feature_column\feature_column_lib.py:18
     15 """FeatureColumns: tools for ingesting and representing features."""
     17 # pylint: disable=unused-import,line-too-long,wildcard-import,g-bad-import-order
---> 18 from tensorflow.python.feature_column.feature_column import *
     19 from tensorflow.python.feature_column.feature_column_v2 import *
     20 from tensorflow.python.feature_column.sequence_feature_column import *

File ~\AppData\Roaming\Python\Python310\site-packages\tensorflow\python\feature_column\feature_column.py:143
    141 from tensorflow.python.framework import sparse_tensor as sparse_tensor_lib
    142 from tensorflow.python.framework import tensor_shape
--> 143 from tensorflow.python.layers import base
    144 from tensorflow.python.ops import array_ops
    145 from tensorflow.python.ops import check_ops

File ~\AppData\Roaming\Python\Python310\site-packages\tensorflow\python\layers\base.py:16
      1 # Copyright 2015 The TensorFlow Authors. All Rights Reserved.
      2 #
      3 # Licensed under the Apache License, Version 2.0 (the "License");
   (...)
     13 # limitations under the License.
     14 # =============================================================================
     15 """Contains the base Layer class, from which all layers inherit."""
---> 16 from tensorflow.python.keras.legacy_tf_layers import base
     18 InputSpec = base.InputSpec
     20 keras_style_scope = base.keras_style_scope

File ~\AppData\Roaming\Python\Python310\site-packages\tensorflow\python\keras\__init__.py:25
     22 from tensorflow.python.keras import distribute
     24 # See b/110718070#comment18 for more details about this import.
---> 25 from tensorflow.python.keras import models
     27 from tensorflow.python.keras.engine.input_layer import Input
     28 from tensorflow.python.keras.engine.sequential import Sequential

File ~\AppData\Roaming\Python\Python310\site-packages\tensorflow\python\keras\models.py:19
     16 """Code for model cloning, plus model-related API entries."""
     18 from tensorflow.python.framework import ops
---> 19 from tensorflow.python.keras import backend
     20 from tensorflow.python.keras import metrics as metrics_module
     21 from tensorflow.python.keras import optimizer_v1

File ~\AppData\Roaming\Python\Python310\site-packages\tensorflow\python\keras\backend.py:50
     48 from tensorflow.python.framework import tensor_util
     49 from tensorflow.python.keras import backend_config
---> 50 from tensorflow.python.keras.distribute import distribute_coordinator_utils as dc
     51 from tensorflow.python.keras.engine import keras_tensor
     52 from tensorflow.python.keras.utils import control_flow_util

File ~\AppData\Roaming\Python\Python310\site-packages\tensorflow\python\keras\distribute\distribute_coordinator_utils.py:33
     31 from tensorflow.python.client import session
     32 from tensorflow.python.platform import tf_logging as logging
---> 33 from tensorflow.python.training import monitored_session
     34 from tensorflow.python.training import server_lib
     36 _worker_context = threading.local()

File ~\AppData\Roaming\Python\Python310\site-packages\tensorflow\python\training\monitored_session.py:22
     19 import os
     21 from tensorflow.core.protobuf import config_pb2
---> 22 from tensorflow.python.checkpoint import checkpoint as trackable_util
     23 from tensorflow.python.checkpoint import graph_view
     24 from tensorflow.python.distribute import distribute_coordinator_context

File ~\AppData\Roaming\Python\Python310\site-packages\tensorflow\python\checkpoint\checkpoint.py:30
     28 from tensorflow.python.checkpoint import checkpoint_options
     29 from tensorflow.python.checkpoint import functional_saver
---> 30 from tensorflow.python.checkpoint import graph_view as graph_view_lib
     31 from tensorflow.python.checkpoint import restore as restore_lib
     32 from tensorflow.python.checkpoint import save_util

File ~\AppData\Roaming\Python\Python310\site-packages\tensorflow\python\checkpoint\graph_view.py:19
     16 import copy
     17 import weakref
---> 19 from tensorflow.python.checkpoint import save_util_v1
     20 from tensorflow.python.checkpoint import trackable_view
     21 from tensorflow.python.trackable import base

File ~\AppData\Roaming\Python\Python310\site-packages\tensorflow\python\checkpoint\save_util_v1.py:25
     23 from tensorflow.core.protobuf import trackable_object_graph_pb2
     24 from tensorflow.python.checkpoint import saveable_compat
---> 25 from tensorflow.python.checkpoint import util
     26 from tensorflow.python.framework import constant_op
     27 from tensorflow.python.framework import dtypes

File ~\AppData\Roaming\Python\Python310\site-packages\tensorflow\python\checkpoint\util.py:21
     19 from tensorflow.python.ops import variables
     20 from tensorflow.python.trackable import trackable_utils
---> 21 from tensorflow.python.training import optimizer as optimizer_v1
     22 from tensorflow.python.util import object_identity
     25 def serialize_slot_variables(trackable_objects, node_ids, object_names):

File ~\AppData\Roaming\Python\Python310\site-packages\tensorflow\python\training\optimizer.py:22
     19 import abc
     21 from tensorflow.python.distribute import distribute_lib
---> 22 from tensorflow.python.distribute import distribute_utils
     23 from tensorflow.python.distribute import distribution_strategy_context as distribute_ctx
     24 from tensorflow.python.distribute import reduce_util as ds_reduce_util

File ~\AppData\Roaming\Python\Python310\site-packages\tensorflow\python\distribute\distribute_utils.py:21
     18 import contextlib
     19 import threading
---> 21 from tensorflow.python.distribute import tpu_values as tpu_values_lib
     22 from tensorflow.python.distribute import values as values_lib
     23 from tensorflow.python.eager import context

File ~\AppData\Roaming\Python\Python310\site-packages\tensorflow\python\distribute\tpu_values.py:22
     15 """Various classes representing TPU distributed values.
     16 
     17 Note that the tests are in values_test.py .
     18 
     19 """
     21 from tensorflow.python.distribute import packed_distributed_variable as packed
---> 22 from tensorflow.python.distribute import tpu_replicated_variable
     23 from tensorflow.python.distribute import tpu_util
     24 from tensorflow.python.distribute import values

File ~\AppData\Roaming\Python\Python310\site-packages\tensorflow\python\distribute\tpu_replicated_variable.py:23
     20 from collections import abc
     21 import contextlib
---> 23 from tensorflow.python.compiler.xla.experimental import xla_sharding
     24 from tensorflow.python.distribute import tpu_util
     25 from tensorflow.python.eager import context

File ~\AppData\Roaming\Python\Python310\site-packages\tensorflow\python\compiler\xla\__init__.py:19
     17 # pylint: disable=unused-import
     18 from tensorflow.python.compiler.xla import jit
---> 19 from tensorflow.python.compiler.xla import xla

File ~\AppData\Roaming\Python\Python310\site-packages\tensorflow\python\compiler\xla\xla.py:20
     15 """xla is an experimental library that provides XLA support APIs."""
     17 import contextlib
---> 20 from tensorflow.compiler.jit.ops import xla_ops
     21 from tensorflow.compiler.jit.ops import xla_ops_grad  # pylint: disable=unused-import
     22 from tensorflow.core.framework import attr_value_pb2

File ~\AppData\Roaming\Python\Python310\site-packages\tensorflow\compiler\jit\ops\xla_ops.py:13
     11 from tensorflow.python.eager import execute as _execute
     12 from tensorflow.python.framework import dtypes as _dtypes
---> 13 from tensorflow.security.fuzzing.py import annotation_types as _atypes
     15 from tensorflow.python.framework import op_def_registry as _op_def_registry
     16 from tensorflow.python.framework import ops as _ops

ModuleNotFoundError: No module named 'tensorflow.security'

已执行的排查操作：

• 多次重新安装TensorFlow，pip show tensorflow显示安装状态正常
• 重装Anaconda并清理安装缓存
• 目标是使用GPU运行TensorFlow，但当前连基础导入都无法完成

解决步骤 #

1. 彻底清理TensorFlow残留 #

• 卸载所有TensorFlow相关包：

  pip uninstall -y tensorflow tensorflow-gpu
  

• 手动删除Python site-packages中的TensorFlow残留文件夹，路径示例：~\AppData\Roaming\Python\Python310\site-packages\tensorflow，同时删除所有以tensorflow-开头的文件夹
• 清理pip缓存：

  pip cache purge
  

2. 安装正确版本的TensorFlow #

TensorFlow 2.x之后已将GPU支持整合到主包，tensorflow-gpu包已被弃用，无需单独安装。需根据本地CUDA和cuDNN版本选择匹配的TensorFlow版本：

• 确认CUDA/cuDNN版本后，执行对应版本的安装命令，例如适配CUDA 11.8的版本：

  pip install tensorflow==2.15.0
  

• 若安装速度慢，可使用国内镜像源：

  pip install tensorflow==2.15.0 -i https://pypi.tuna.tsinghua.edu.cn/simple
  

3. 对齐Jupyter环境与TensorFlow安装环境 #

确保Jupyter使用的Python环境和安装TensorFlow的环境一致：

• 在Jupyter Notebook中执行以下代码查看当前环境路径：

  import sys
  print(sys.executable)
  

• 在终端中执行where python（Windows）或which python（Linux/macOS）查看TensorFlow安装环境的路径
• 若路径不一致，需将TensorFlow安装到Jupyter对应的环境，或在Jupyter中切换内核到正确的Python环境

4. 验证安装成功 #

安装完成后，在终端或Jupyter中执行以下代码验证：

import tensorflow as tf
print(tf.__version__)
print(tf.config.list_physical_devices('GPU'))

若输出GPU设备信息，则说明GPU配置成功

内容的提问来源于stack exchange，提问作者Ahmad Syauqi