Runtime 8.4 ML still showing TF 2.3

[javierxperez60]

I updated my cluster runtime to use 8.4 ML, I did a !pip list and saw that tensorflow was 2.3. What should I do to resolve this? Not sure if I should just do a pip install Tensorflow==2.5.0

[javierxperez60]

If the answer is the latter, do i need to install the supported versions of cudnn and cuda?