Again you mix a lot stuff into this 
So do not mention conda or Docker,as it has nothing do about what you using now.
Now use Python 3.7 for python.org,it has no connection to Anaconda/conda that is a standalone version.
I would not install all of to this on a root original Python version(as you do now),
i could show how this is done in a virtual environment with venv.
virtual environment is used to avoid version problems as you have now,as all most be installed new into environment.
Anacondas could be used,but as mention has nothing to with what use use now,as it's
Anaconda has both
A little frustrating as i have show this before in our Threads,
but you still seems to mix stuff together and do not understand basic concept of this.
An other way here i use CoLab Notebook,just to
then all you import's work,and you don't have to any installation of Keras/Tensorflow as that's already done.
AppData\Local\Programs\Python\Python37Let's make one thing be clear when you have this path to Python,you are not using Anaconda/conda or virtual environment.
So do not mention conda or Docker,as it has nothing do about what you using now.
Now use Python 3.7 for python.org,it has no connection to Anaconda/conda that is a standalone version.
I would not install all of to this on a root original Python version(as you do now),
i could show how this is done in a virtual environment with venv.
virtual environment is used to avoid version problems as you have now,as all most be installed new into environment.
Anacondas could be used,but as mention has nothing to with what use use now,as it's
stand aloneAnaconda has both
conda and pip(tied to only Anaconda and not python.org version).A little frustrating as i have show this before in our Threads,
but you still seems to mix stuff together and do not understand basic concept of this.
An other way here i use CoLab Notebook,just to
pip install yfinance ,then all you import's work,and you don't have to any installation of Keras/Tensorflow as that's already done.