Error:Layer ModuleWrapper has arguments in `__init__` and therefore must override `get_config`.

这个报错是因为keras 和 tensorflow库混合使用了，导致model.save时出现问题

原始代码

import os
import pandas as pd
import tensorflow as tf
import numpy as np
import Config
from tensorflow.python.keras.layers import Dense, Activation, Dropout
from keras.layers import LSTM, GRU, Bidirectional
from tensorflow.python.keras.models import Sequential, load_model
from tensorflow.python.keras import optimizers

其中既有keras直接导入的包也有tensorflow.python.keras导入的包，此时训练过程不会受到影响，但在最后一步model.save时会报标题所示的错误。

更改代码

import os
import pandas as pd
import tensorflow as tf
import numpy as np
import Config
from keras.layers import Dense, Activation, Dropout
from keras.layers import LSTM, GRU, Bidirectional
from keras.models import Sequential, load_model
from keras import optimizers

全部改为keras直接导入变不会出现任何保存问题。