scikeras and tensorflow compatibility Problem

[MuhammadShifa]

I want to run your repo., but I am facing the problem in repository. I have tried a lot but not able to find best version of scikeras and tensorflow that works great together.
As you know scikeras required tensorflow>2.4.0. Yeah Its okay but when we are going to install it we are facing this issue.

` \venv\Lib\site-packages\tensorflow\include\external\
cudnn_frontend_archive\_virtual_includes\cudnn_frontend\third_party\cudnn_frontend\include\cudnn_frontend_EngineConfigGenerator.h'
'

And if we try to run app.py we facing this issue
ImportError: No module named 'tensorflow.python

I have remove this error by installing tensorflow==2.0 but there is problem of scikeras which required tensorflow version greater than tensorflow>2.4.0

Kindly look at the problem. Thanks for great work.

[MuhammadShifa]

I have solved this issue by installing pip install tensorflow == 2.5.0 and then install pip install scikeras==0.2.1

[fcakyon]

@MuhammadShifa Its good to see that you managed to solve the issue!

scikeras dependency managment is terrible unfortunately and the maintaners doesnt want to handle this issue. See here for my related issue: adriangb/scikeras#221

[MuhammadShifa]

@fcakyon Yeah I have fixed all the dependency issues but now facing problem in coding part of face_recognition.py

face_recognition.py", line 22, in detect_faces
response_json = response.json()
UnboundLocalError: local variable 'response' referenced before assignment

Can you look into this? Kindly try to make it runnable. I am facing the above issue when I upload the picture and then click on upload and Detect button.

[fcakyon]

@MuhammadShifa have you read the usage section in readme: https://github.com/fcakyon/face-recognition-app-tutorial#app-usage

You have to run a face detection app before using this face recognition app. Can you confirm your face detection app is working properly on the correct port?

[MuhammadShifa]

@fcakyon Sorry I was doing mistake, Thank you so much, now both are running successfully.

[fcakyon]

@MuhammadShifa glad to hear that! Enjoy your apps 💯