ModuleNotFoundError: No module named 'keras.engine.topology'

[jkortner]

When I try to import keras-vggface in Google Colab I get the error: No module named 'keras.engine.topology'. The same happens on my local machine.

First, I install keras-vggface:

!pip install keras_vggface
!pip install keras_applications

Then, I try to import:

from keras_vggface.vggface import VGGFace

The output is as follows:

---------------------------------------------------------------------------
ModuleNotFoundError                       Traceback (most recent call last)
<ipython-input-20-4dcc6ac1877c> in <module>()
----> 1 from keras_vggface.vggface import VGGFace
      2 print(keras_vggface.__version__)

2 frames
/usr/local/lib/python3.7/dist-packages/keras_vggface/__init__.py in <module>()
----> 1 from keras_vggface.vggface import VGGFace
      2 from keras_vggface.version import __version__

/usr/local/lib/python3.7/dist-packages/keras_vggface/vggface.py in <module>()
      7 '''
      8 from __future__ import print_function
----> 9 from keras_vggface.models import RESNET50, VGG16, SENET50
     10 
     11 

/usr/local/lib/python3.7/dist-packages/keras_vggface/models.py in <module>()
     18 from keras import backend as K
     19 from keras_vggface import utils
---> 20 from keras.engine.topology import get_source_inputs
     21 import warnings
     22 from keras.models import Model

ModuleNotFoundError: No module named 'keras.engine.topology'

[arnoldcvl]

Same error for me when using the latest keras version. Keras 2.4.3/tensorflow 2.5 works fine.

[vmarichkav]

I solved this issue by changing the import from
from keras.engine.topology import get_source_inputs
to
from keras.utils.layer_utils import get_source_inputs in keras_vggface/models.py.

I made a PR for this issue.

[tamis-laan]

Same issue here, can this fix be merged?

[inhyeokJeon]

Same issue here, can this fix be merged?

[braludo]

Same issue here, can this fix be merged?

[Keven1894]

I solved this issue by changing the import from from keras.engine.topology import get_source_inputs to from keras.utils.layer_utils import get_source_inputs in keras_vggface/models.py.

I made a PR for this issue.

same problem solved by this, plz accept the PR

[nidham-sony]

I solved this issue by changing the import from from keras.engine.topology import get_source_inputs to from keras.utils.layer_utils import get_source_inputs in keras_vggface/models.py.

I made a PR for this issue.

Same thing, it worked here.

[PierrunoYT]

import cv2
import numpy as np
from keras.models import load_model
from keras.preprocessing import image
from keras_vggface.utils import preprocess_input
from keras_vggface.vggface import VGGFace
from keras.engine.base_layer import Layer

def age_gender_prediction(image_path):
# load the image using OpenCV
img = cv2.imread(image_path)
# convert the image from BGR to RGB
img = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)
# resize the image to (224, 224) to match the input size of the VGGFace model
img = cv2.resize(img, (224, 224))
# convert the image to an array of floating point values
x = image.img_to_array(img)
# preprocess the image using the preprocess_input function from keras_vggface
x = np.expand_dims(x, axis=0)
x = preprocess_input(x, version=2)
# load the VGGFace model
model = VGGFace(model='resnet50')
# use the model to predict the gender and age of the person in the image
y_pred = model.predict(x)
# extract the gender and age predictions from the output of the model
gender_pred = 'Male' if np.argmax(y_pred[0][0:2]) == 0 else 'Female'
age_pred = int(np.floor(np.clip(y_pred[1], 0, 100))[0][0])
# print the predicted gender and age
print('Predicted Gender:', gender_pred)
print('Predicted Age:', age_pred)

example usage

image_path = r'C:\Users\pierr\Dropbox\pierre.jpg'
age_gender_prediction(image_path)

[sbelharbi]

hi @rcmalli,
can you please merge this pr to avoid editing every install.
very much appreciated.
thanks

I solved this issue by changing the import from from keras.engine.topology import get_source_inputs to from keras.utils.layer_utils import get_source_inputs in keras_vggface/models.py.

I made a PR for this issue.

[bibinkunjumon2020]

Go inside the installed library and change that line there.
from keras_vggface.vggface import VGGFace
To
from keras.utils import get_source_inputs

ANy similar issues can be fixed like this.Even in colab

[bibinkunjumon2020]

from keras.engine.topology import get_source_inputs
from keras.utils.data_utils import get_file

from keras.utils import layer_utils

to

from keras.utils import get_source_inputs
from tensorflow.python.keras.utils.data_utils import get_file
from tensorflow.python.keras.utils import layer_utils