MLOps_VideoAnomalyDetection
MLOps_VideoAnomalyDetection copied to clipboard
“Keras requires TensorFlow 2.2 or higher”!!
The config.yml pins tensorflow==1.15. However no Keras version is pinned. Thus, after execution of conda env create -f config/environment.yml
latest version of Keras (i.e. 2.4.3) gets installed.
While executing train.py locally, it throws Error “Keras requires TensorFlow 2.2 or higher”
I think pinning appropriate version of keras would help. What it is?
I got the same problem. Downgrade Keras to 2.2.4 still not working. Now we want to know what Keras version in the original environment?
the solution is to install Keras 2.3.1
the solution is to install Keras 2.3.1
Have tried installing Keras 2.3.1? Did it work for you in conda environment?
Yes. Worked! Should try it. using conda!