Keras: Keras Identity Layer warning

[sachinruk]

I am trying to create a layer that takes x as is (so that I can merge it with other layers).

I have used the following two methods and they both give the same warning:

/Users/sachin/anaconda/lib/python3.5/site-packages/keras/engine/topology.py:1656: UserWarning: Model inputs must come from a Keras Input layer, they cannot be the output of a previous non-Input layer. Here, a tensor specified as input to "sequential_158_model" was not an Input tensor, it was generated by layer identity_15.
Note that input tensors are instantiated via `tensor = Input(shape)`.
The tensor that caused the issue was: identity_input_10
  str(x.name))

Method 1

x = Sequential()
x.add(Lambda(lambda x: x, input_shape=(x2.shape[1],)))

Method 2

from keras import backend as K
from keras.engine.topology import Layer

class Identity(Layer):
    def __init__(self,output_dim, **kwargs):
        self.output_dim = kwargs['input_shape'][0]
        super(Identity, self).__init__(**kwargs)

    def call(self, x, mask=None):
        return x

    def get_output_shape_for(self, input_shape):
        return (input_shape[0], self.output_dim)

x = Sequential()
x.add(Lambda(lambda x: x, input_shape=(x2.shape[1],)))

My questions are:

1. Is it safe to proceed with this warning?
2. Is there a way of getting rid of the warning?
3. Is the Lambda option preferable simply for readability reasons?

[farizrahman4u]

Is it safe to proceed with this warning?

Yes.

Is there a way of getting rid of the warning?

x.add(Lambda(lambda x: x + 0, input_shape=(x2.shape[1],)))

Is the Lambda option preferable simply for readability reasons?

Yes.

[stale[bot]]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs, but feel free to re-open it if needed.