ianforme / face-prediction

real-time age, gender and emotion prediction based on livestream webcam data

face-prediction

real-time age, gender and emotion prediction based on livestream webcam data

pre-trained models

pre-trained models can be downloaded from here. Due to computational resource constraints, not all images from UTKfaces are used for training. Thus model performance are expected to be improved if more data are used.

worth mentioning that these models are not optimized and built for quick prototyping. There are more areas of improvements

requirements

mtcnn
keras
tensorflow
openCV

face-prediction's Issues

AttributeError

Hi,

thanks for your work.

i get AttributeError: 'Model' object has no attribute '_distribution_strategy'

i figure it has to do with keras and tf, but i'm unable to find a way.

Best,

xxtu

---

AttributeError Traceback (most recent call last)

in
20
21 # Find all the faces in the current frame of video
---> 22 face_locations = detect_face(rgb_frame)
23
24 # Display the results

in detect_face(img)
28
29 # create predictions
---> 30 sex_preds = sex_model.predict(center_img.reshape(1,224,224,3))[0][0]
31 age_preds = age_model.predict(center_img.reshape(1,224,224,3))[0][0]
32

~/.pyenv/versions/3.8.7/lib/python3.8/site-packages/tensorflow/python/keras/engine/training.py in predict(self, x, batch_size, verbose, steps, callbacks, max_queue_size, workers, use_multiprocessing)
1580
1581 outputs = None
-> 1582 with self.distribute_strategy.scope():
1583 # Creates a tf.data.Dataset and handles batch and epoch iteration.
1584 dataset_types = (dataset_ops.DatasetV1, dataset_ops.DatasetV2)

~/.pyenv/versions/3.8.7/lib/python3.8/site-packages/tensorflow/python/keras/engine/training.py in distribute_strategy(self)
687 def distribute_strategy(self):
688 """The tf.distribute.Strategy this model was created under."""
--> 689 return self._distribution_strategy or ds_context.get_strategy()
690
691 @Property

AttributeError: 'Model' object has no attribute '_distribution_strategy'

Help with keras.engine

Hi all,

I am trying to run the stream_prediction-upload.py file and I have been facing this error -

I tried some solutions I found specifically change of import styles, but the error persists. Has anyone faced and managed to solve the error?
Thank you very much in advance!

Model' object has no attribute '_call_spec

Model' object has no attribute '_call_spec'
File "C:\Users\mayur_iyd6xcu\OneDrive\Desktop\Python\models\facepred.py", line 41, in detect_face
sex_preds = sex_model.predict(center_img.reshape(1,224,224,3))[0][0]
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Users\mayur_iyd6xcu\OneDrive\Desktop\Python\models\facepred.py", line 73, in
face_locations = detect_face(rgb_frame)
^^^^^^^^^^^^^^^^^^^^^^
AttributeError: 'Model' object has no attribute '_call_spec'

i can't access link for download pre-trained model

I can't access the link to download the pre-trained model.
please fix it, thank you...