When trying to run Programming Assignment: Residual Networks locally, I get the error of failing to import tensorflow.keras.
I just installed them and my tensorflow version is 2.3.0
I suspect the problem is about my keras version. my current version of keras is 2.3.1.
What version of keras should we use ?
1 Like
Our courses do not use the latest versions. You may check what version is being used in each assignment by adding some print statements. Also, there are many guides on this forum on how to setup with your local environment. I will share a link here in a few hours when I be on my computer…
1 Like
Here is the link I was talking about: Tips to work on a local environment.
2 Likes
I periodically run the assignments locally. They do not always convert 1 to 1. Usually the compatibility is not with TF/ Keras. Usually has to do with some other dependency/ library being used. I have a local GPU that I leverage and some of the challenges I faced were. cuDNN/ CUDA release compatibility with keras/TF. That was the core dependency. The others like plotting are a bit easier to fix.
2 Likes
Hey, just one more question, is the dependency for every notebook from Class1 to Class5 the same?
That is : If I type !pip list
in notebook, will the result be all the same?
Unfortunately, the answer is No.