python - 属性错误 : module 'tensorflow.compat.v2.__internal__' has no attribute 'dispatch'

Question

当我尝试在 Jupyter notebook 中导入 Keras.layers 时，如下所示，我遇到了这个错误。

# Basic Imports
import numpy as np
import matplotlib.pyplot as plt
from glob import glob
from PIL import Image

# Imports for Building CNN
from tensorflow import keras
from keras.layers import Input, Lambda, Dense, Flatten
from keras.models import Model
from keras.applications.vgg16 import VGG16
from keras.applications.vgg16 import preprocess_input
from keras.preprocessing import image
from keras.preprocessing.image import ImageDataGenerator
from keras.models import Sequential, load_model
from tensorflow.keras.preprocessing.image import load_img

# Ignore Warnings
import warnings
warnings.filterwarnings("ignore")

AttributeError: module 'tensorflow.compat.v2.internal' has no attribute 'dispatch'

我从这个错误提要中得到了这个错误:

>     ---------------------------------------------------------------------------
>     AttributeError                            Traceback (most recent call last)
>     Input In [1], in <cell line: 9>()
>           7 # Imports for Building CNN
>           8 from tensorflow import keras
>     ----> 9 from keras.layers import Input, Lambda, Dense, Flatten
>          10 from keras.models import Model
>          11 from keras.applications.vgg16 import VGG16
>     
>     File D:\Black\miniconda3\envs\tensorflow\lib\site-packages\keras\__init__.py:20,
> in <module>
>           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 """Implementation of the Keras API, the high-level API of TensorFlow.
>          16 
>          17 Detailed documentation and user guides are available at
>          18 [keras.io](https://keras.io).
>          19 """
>     ---> 20 from keras import distribute
>          21 from keras import models
>          22 from keras.engine.input_layer import Input
>     
>     File D:\Black\miniconda3\envs\tensorflow\lib\site-packages\keras\distribute\__init__.py:18,
> in <module>
>           1 # Copyright 2019 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 """Keras' Distribution Strategy library."""
>     ---> 18 from keras.distribute import sidecar_evaluator
>     
>     File D:\Black\miniconda3\envs\tensorflow\lib\site-packages\keras\distribute\sidecar_evaluator.py:22,
> in <module>
>          20 from tensorflow.python.platform import tf_logging as logging
>          21 from tensorflow.python.util import deprecation
>     ---> 22 from keras.optimizers.optimizer_experimental import (
>          23     optimizer as optimizer_experimental,
>          24 )
>          25 from tensorflow.python.util.tf_export import keras_export
>          27 _PRINT_EVAL_STEP_EVERY_SEC = 60.0
>     
>     File D:\Black\miniconda3\envs\tensorflow\lib\site-packages\keras\optimizers\__init__.py:25,
> in <module>
>          22 import tensorflow.compat.v2 as tf
>          24 # Imports needed for deserialization.
>     ---> 25 from keras import backend
>          26 from keras.optimizers.legacy import adadelta as adadelta_legacy
>          27 from keras.optimizers.legacy import adagrad as adagrad_legacy
>     
>     File D:\Black\miniconda3\envs\tensorflow\lib\site-packages\keras\backend.py:32,
> in <module>
>          29 import numpy as np
>          30 import tensorflow.compat.v2 as tf
>     ---> 32 from keras import backend_config
>          33 from keras.distribute import distribute_coordinator_utils as dc
>          34 from keras.engine import keras_tensor
>     
>     File D:\Black\miniconda3\envs\tensorflow\lib\site-packages\keras\backend_config.py:33,
> in <module>
>          28 # Default image data format, one of "channels_last", "channels_first".
>          29 _IMAGE_DATA_FORMAT = "channels_last"
>          32 @keras_export("keras.backend.epsilon")
>     ---> 33 @tf.__internal__.dispatch.add_dispatch_support
>          34 def epsilon():
>          35     """Returns the value of the fuzz factor used in numeric expressions.
>          36 
>          37     Returns:
>        (...)
>          42     1e-07
>          43     """
>          44     return _EPSILON
>     
>     AttributeError: module 'tensorflow.compat.v2.__internal__' has no attribute 'dispatch'

我对为什么会收到此错误感到困惑。我只想导入它以便运行我的代码。

Answer 1

这为我解决了(来源:https://www.cnblogs.com/wwj321/p/16832760.html)[必须翻译页面]

在我的例子中，我使用的是 Tensorflow 版本 2.4.0，所以我只是在运行模型构建器之前添加了这个(我使用的是 Google Colab):

!pip install -U -q segmentation-models
!pip install -q keras==2.4.3

from tensorflow import keras
import segmentation_models as sm