Need help with unity DQN reinforcement learning !!!

I'm learning to use DQN algorithm, which is the banana navigation code in the previous version of mlagents.
This is the code that uses jupyter notebook, and I have set up the environment, but I stopped and reported an error after only training twice during the operation. How to correct this mistake, please, can you point out where this mistake is? Thank you very much.

 

It would help if you looked at the error message more carefully.

it says in line 322 in environment.py, vector_action.keys() is an integer instead of it being a dictionary as you expect it to be.

So you might want to investigate why that might be.

Good luck!