Keras: Conv2D TypeError: __init()__ takes at least 4 arguments (4 given)

[ankitsharma285]

I am learning to use Keras's Functional API . I tried running one of code snippet (ex: inception module ) from the documentation (https://keras.io/getting-started/functional-api-guide/). I get the above mentioned error. I tried running other code snippets as well from the documentation. I get same error for each one.

---

Inception Module code snippet from the docs:

from keras.layers import Conv2D, MaxPooling2D, Input

input_img = Input(shape=(256, 256,3))

tower_1 = Conv2D(64, (1, 1), padding='same', activation='relu')(input_img)
tower_1 = Conv2D(64, (3, 3), padding='same', activation='relu')(tower_1)

tower_2 = Conv2D(64, (1, 1), padding='same', activation='relu')(input_img)
tower_2 = Conv2D(64, (5, 5), padding='same', activation='relu')(tower_2)

tower_3 = MaxPooling2D((3, 3), strides=(1, 1), padding='same')(input_img)
tower_3 = Conv2D(64, (1, 1), padding='same', activation='relu')(tower_3)

output = keras.layers.concatenate([tower_1, tower_2, tower_3], axis=1)

I am trying to use Conv2D as the first layer in functional API setting. Can anyone please provide me some suggestion on how to resolve this issue?

Thank you.

Ankit Sharma.

[ramanb25]

Have you checked your keras version?

Consider upgrading as the API call has been changed.

sudo pip install --upgrade keras

[ankitsharma285]

@ramanb25: I did the same few hours back. The issue got resolved. Thank you for your suggestion.

[danae2hyr]

I have the same problem as you， could you tell me how to solve this issue? Thanks

[SilverJacket]

@ramanb25 Where do I type that command? I'm using a Jupyter notebook on AWS. In my shell, I shut down the notebook in order to access my ubuntu command line (while ssh'ed into the AWS server) and typed that line, but it said "sudo: pip: command not found". Thanks.

[SabariRaaj]

x1 = Conv2D(16, (4, 4), strides=(2, 2), activation='relu')(input_1)
TypeError: __init__() takes at least 4 arguments (4 given)
This is the error I'm getting, I tried upgrading keras and tensorflow versions but nothing seems to be effective. Please give me some solution

[brotheryun]

@SabariRaaj Me too!Could you tell me how to solve this issue?

[SabariRaaj]

@SabariRaaj Me too!Could you tell me how to solve this issue?

@brotheryun Bro just use the following versions, keras 2.1.5 and tensorflow 1.5.0. only these versions matched perfectly and solved my issue. These versions are for python 3.