Python Reinforcement Learning using Stable baselines. Mario PPO
Вставка
- Опубліковано 3 жов 2022
- Code step by step along with me as I show video game AI networks are created. At the end of this tutorial, you will have a working Artificial Intelligence network playing Mario in Python! We use stable baselines 3 which makes this process extremely simple. Stable baselines 3 is built with PyTorch and can be used for several custom problems.
github.com/ClarityCoders/Mari...
👊You got something to say to me?? Join our discord!✌️
/ discord
Python Version used in this video.
www.python.org/downloads/rele... - Наука та технологія
This is the video that got me my first working PPO
Watched all of your videos recently, glad to see a new one :D
Already have another in the works. Stay tuned. Thanks for watching and commenting. Have you joined the discord?
hello
just here to say thank you, i am a french reunion island engineering student i have to make an presentation about deep learning mario.
Unfortunately i am very bad in english, but you speak very well and i am glad to finally understand the thing
so thank you and have a great day.
and just for saying, gym-retro doesn’t work with python 3.10 and 3.9
Great video, I enjoy having your videos pop up in my subscription box and learning new things. I wish if you maybe could explain the concepts in a little more depth in the future, but overall, very informative. Thank you
Of course! Thanks for this feedback. I never now how deep to go in some of these tutorials. Thank you for the view and comment.
A concept ive been considering but college keeps me to busy to explore it. Using Reinforcement leaning to drive the training of Supervised learning. The supervised learning model guesses outcomes and the Reinforcement learning model tweeks the learning rate. Sense supervised learning models are better at being more flexible and better as making decisions. But instead of manually tweaking the learning rate you have a Reinforcement learning model do that for you. And it learns based on how good the guesses are of the supervised model.
It kind of feels like it would be more flexible over time because reinforcement learning models can be so inflexible.
This is very interesting. You should join our discord and we should chat! Thanks for commenting.
population based training by deepmind is kinda like that but its not reinforcement learning
great explanation, helps me a lot. can we have another tutorial on how to trained AI on pyboy emilator?
Please make an video on custom environment with UAV/Dronese application in mind. Maybe a drone following a target using RL etc.
Good job, thanks for share. I saw "Mega man" in example code. i want to train this game :)
Stay tuned I got another one coming with Mega Man!
Great explanation! one question, why didn´t just use make_vec_env ?
Nice Work sir!
Thanks you! I hope you enjoyed we have another crazy one coming right around the corner.
Cool!
Is it possbile to do this with actual online games ? If yes, how :)
How hard would be to include score into consideration for training?
_get_info() returns coins and score as a dict. Im trying to add that information as rewards, if I can find a good answer I will get back to you lol
I can't install gym bros
My python version is 3.8
I don't know what is happening, it throws a lot of errors
python version used is 3.8.10
Hello I have A paid Project for you
You should probably reach out to him through his business email then, instead of comments, bud.
@@HypnoTox can you provide his buisness email id
@@thejasgowda9295 Go to his profile -> About -> View E-Mail Address
@@HypnoTox Not Given In Profile
ClarityCoders, Unfortunately, your code didn't work for me. What could be the problem ? Why does it return a value of type ndarray and not int ?
drive.google.com/file/d/1mPhwCOTgaRDQw12fWbzSgT5SmR5JQxT-/view?usp=share_link - screenshot