I am trying to contribute to tf-keras. Ihave forked it and cloned it to my local env. I have created a vertual environment and installed the dependencies from requirements.txt
then I have written a simple python file the first two line of which is
import tf_keras
from tf_keras import layers
this throws an error, namely:
ImportError: cannot import name ‘saved_metadata_pb2’ from ‘tf_keras.protobuf’ (unknown location)
any idea on what I am doing wrong?
moreover I would be greatful if you can explain how exactly to make changes in codes and see the effect of doing so.
Hi @ArkadebMisra
Welcome to the TensorFlow Forum!
tf_keras
package has not released yet and will be available along with TensorFlow 2.16 release. Please check the same in this Keras Newsletter (October 6, 2023). Untill then you can install and import keras using below code:
!pip install tensorflow
import keras
print(keras.__version__)
from keras import layers
Yes. I know that. I am trying to use it in Dev environment for opensource. I have forked it and cloned it from tf-keras repo. Then installed the requirements. And then createed a simple python file .
Say I want to use dense layer from Dense.py which is in tf_keras folder. So I’m writing
from tf_keras import layers.
Where is my mistake?
I have not found any document explaining how exactly change the codes and see the results.
Forgive me if these questions are silly. I’m a bigginer.