如何解决「使用的多后端Keras版本不支持TensorFlow 2.0」报错

• 报错核心原因是代码混用了独立安装的第三方Keras包和TensorFlow 2内置的tf.keras包，独立版本的Keras不兼容TensorFlow 2，调用时就会触发该报错。

修改导入规则，所有Keras相关接口统一使用TensorFlow内置的tf.keras下的实现，删除独立Keras的导入、以及tensorflow.python.keras的私有接口导入，清理重复导入项。修正后的完整代码如下：

from sklearn.model_selection import train_test_split
import tensorflow as tf
from tensorflow.keras import layers, optimizers, callbacks
from tensorflow.keras.layers import Input, Dense, Dropout, BatchNormalization, Conv2D, MaxPooling2D, AveragePooling2D, Activation, Flatten
from tensorflow.keras.models import Sequential, load_model, Model
from tensorflow.keras.preprocessing.image import ImageDataGenerator
from tensorflow.keras import datasets
from tensorflow.keras.optimizers import Adam
from tensorflow.keras.utils import plot_model
from tensorflow.keras.callbacks import EarlyStopping

model = Sequential()
model.add(Conv2D(input_shape=(32,32,4),filters=64,kernel_size=(3,3),padding="same", activation="relu"))
model.add(Conv2D(filters=64,kernel_size=(3,3),padding="same", activation="relu"))
model.add(MaxPooling2D(pool_size=(2,2),strides=(2,2)))
model.add(Conv2D(filters=128, kernel_size=(3,3), padding="same", activation="relu"))
model.add(Conv2D(filters=128, kernel_size=(3,3), padding="same", activation="relu"))
model.add(MaxPooling2D(pool_size=(2,2),strides=(2,2)))

model.add(Flatten())
model.add(Dense(units=1024,activation="relu"))
model.add(Dense(units=1024,activation="relu"))
model.add(Dense(units=1, activation="linear"))

#compile model
optimizer = tf.keras.optimizers.Adam(learning_rate=0.0001, beta_1=0.9, beta_2=0.999, epsilon=1e-08, decay=0.0)
model.compile(loss='mean_squared_error',
                optimizer=optimizer,
                metrics=['mae'])

• 后续开发不要调用独立Keras的接口，也不要使用tensorflow.python.keras下的私有接口，全部统一用tf.keras前缀的公开接口即可避免此类兼容性问题。

内容的提问来源于stack exchange，提问作者rayan