Python Flappy Bird AI Tutorial (with NEAT) - NEAT Configuration and Explanation

Поделиться
HTML-код
  • Опубликовано: 1 янв 2025

Комментарии •

  • @Soljarag5
    @Soljarag5 5 лет назад +44

    This series is turning out to be exactly what I wanted to learn NEAT!!! Can't wait for the full series!

  • @novicetech1
    @novicetech1 5 лет назад +26

    Thanks, Tim. Another outstanding tutorial. I really appreciate your linking the configuration file and the NEAT documentation. I typed out the entire config file anyway because it helped me better understand the purpose of each line or section. When I got a little confused, having easy access to the documentation really helped. Once again, thank you for the amazing tuts and your professional teaching style. As always, can't wait for the next video in this series.

  • @darkequinox4197
    @darkequinox4197 4 года назад +12

    I am watching this in quarantine and still enjoying

  • @username42
    @username42 5 лет назад +73

    dude dont forget to put the part number on the title !

  • @kamakura9722
    @kamakura9722 Год назад +1

    Tim: I read the whole documentation to do this
    Also Tim: I don't know how it works I stole this config
    xD It's funny
    I am following your tutorial after Covid and it's even more relevant I think! Thank you so much
    🙏🏻

  • @rutvikrana512
    @rutvikrana512 4 года назад +4

    So much info ...... I love you have made this first so we can understand 💕💕💕

  • @technoinfoworldwide2329
    @technoinfoworldwide2329 5 лет назад +2

    I am going watch this series after 1 week finishes of my exam....thanks for video bro

  • @soupnoodles
    @soupnoodles 3 года назад +6

    I recommend implementing Flappy Bird by yourself, then joining the tutorial from this point
    Otherwise it often gets to the point that even tho you understood how it works, when you want to implement an Ai for your own game, you will have a bit of trouble.

  • @lostandfound2893
    @lostandfound2893 4 года назад +4

    Could you please do a series on how to implement neat with multiple outputs rather than a single output?

  • @sharansiva7635
    @sharansiva7635 5 лет назад +18

    These tutorials are amazing but it would be nice if you run some code in it to show and explain

  • @sagittario8796
    @sagittario8796 2 года назад

    This is so interesting, being able to create some form of intelligence and even life.

  • @هانيالعتيبي-ق7ذ
    @هانيالعتيبي-ق7ذ 5 лет назад +4

    Very inspiring 🌹🌹💙🙏

  • @bornaahmadzadeh3408
    @bornaahmadzadeh3408 5 лет назад +2

    Hey! I’m really inspired by your videos to make my own AI to play Flappy Bird but I was wondering would logistic regression be a good algorithm(that’s assuming we have the proper data)? I’ve seen quite a few models for Flappy Bird but none of them used logistic regression. Is there a reason for that?I feel like it’s a good choice for something like this.

    • @fadop3156
      @fadop3156 5 лет назад

      Borna Ahmadzade well try it

    • @camozot
      @camozot 5 лет назад

      Yes, logistic regression is just a neural network without any hidden layers

  • @aryanbhatia6992
    @aryanbhatia6992 5 лет назад +9

    Can you tell in which University you are in ??

  • @jaydankamotho4580
    @jaydankamotho4580 4 года назад +2

    i love you tutorial all the way from episode one but i think you could have gotten better results using relu function

  • @dev.kukata
    @dev.kukata 4 года назад

    This tutorial is pretty neat

  • @rupajchowdhury1877
    @rupajchowdhury1877 3 года назад

    So here's the NEAT part

  • @techfreak3916
    @techfreak3916 4 года назад

    thank you for this tutorial tim

  • @dominickdupuy7891
    @dominickdupuy7891 Год назад

    What if you make a NEAT network to optimize the settings of a NEAT network?

  • @AgentRex42
    @AgentRex42 5 лет назад +1

    7.30 I don't understand "activation function TanH", why ?

    • @maantjemol
      @maantjemol 5 лет назад +2

      with that, you can calculate a number between 1 and -1 from any number you give it. if it is above 0.5, the ai thinks it can jump. if it is under 0.5, the ai thinks it cannot jump

    • @AgentRex42
      @AgentRex42 5 лет назад

      @@maantjemol Thank you.

  • @venetsiakrusteva8893
    @venetsiakrusteva8893 3 года назад

    Any ideas why it is giving AttributeError: module 'neat' has no attribute 'Config' even though neat is installed

  • @codingskillbyharibhushanpa5684
    @codingskillbyharibhushanpa5684 3 года назад

    When I run my code I got error :
    stagnation = config.stagnation_type(config.stagnation_config, self.reporters)
    TypeError: __init__() missing 1 required positional argument: 'stagnation'

  • @leozarni9570
    @leozarni9570 4 года назад

    I'm a little confused by the downloading file as a text part. So I copy-pasted the config file into notepad, and saved it in the same folder where my flappy bird file is. However, after following your instructions in the next video, when I ran the file, I seem to get an error stating, "NameError: name 'config_path' is not defined" . Can anyone help me out here please?

  • @cheshire1
    @cheshire1 5 лет назад +4

    Why do you say that you're going to let NEAT pick the activation function and then fix it to tanh?

    • @erwinb.7711
      @erwinb.7711 5 лет назад +5

      tanh is only for the output activation function. The activation functions for the hidden layers is figured out by NEAT

  • @LicenseToPlay
    @LicenseToPlay 5 лет назад

    Hey Tim fantastic content. I tried to subscribe at the website but when i click on the button nothing happens. How cn I sub ? thanks

  • @dark_lord98
    @dark_lord98 4 года назад

    will it require GPU to run the algo??

  • @chilirum1061
    @chilirum1061 5 лет назад

    Hey everyone, I'm currently trying to find out how one could make the lives of busy coders a lot easier ;)
    Which is why I have two very simple questions:
    1. As a career-oriented individual, what are the 2 biggest issues you're dealing with when trying to stay healthy?
    2. When it comes to balancing career with peace of mind and physical health, what would you wish for more than anything else?
    Thanks so much in advance - looking forward to reading your answers

    • @yoelc6009
      @yoelc6009 5 лет назад +1

      1. I've started taking programming seriously about 6 months ago. I've had no problems staying healthy because I've always had a good diet and gone to the gym consistently. However, the only thing that could be considered as 'unhealthy' is that I don't really sleep 8 hours a day. I sleep usually around 5-6 hours due to the amount of time I dedicate to programming secondary to school.
      2. Honestly, I only wish that I am able to pay my gym subscription, because it gives me both physical health and peace of mind.
      A little made up conclusion, going to the gym is essential for any person; not only programmers.
      Good luck on your project, and sorry for the late answer!

    • @chilirum1061
      @chilirum1061 5 лет назад

      @@yoelc6009 Thank you so much! I appreciate this

  • @hiddemarsman1117
    @hiddemarsman1117 3 года назад

    Bruh mine doesnt work lol, it just yeets itself in the first and second pipes on purpose it seems. Isn't really learning at all. How can I fix this?

  • @yaboyJJJJJ
    @yaboyJJJJJ Год назад

    it doesn't make any sense to use the birds Y position as neuron because it has 0 impact on decision making. It actually hurts the AI quite a bit by including it. Either only give 2 neurons or give the birds velocity as the third neuron.

  • @mrfrozen97-despicable
    @mrfrozen97-despicable 4 года назад

    Sounds interesting.....
    Thanks))))

  • @jakubd
    @jakubd 5 лет назад

    thanks to youtube for video speed change

  • @giacomoarienti
    @giacomoarienti 4 года назад

    I didn't get the difference between the tanh and the sigmoid activation function.

    • @fortunegaming1114
      @fortunegaming1114 4 года назад

      The TanH function gives the value between - 1 and 1
      The Sigmoid Function gives the value between 0 and 1

  • @sunhouhua
    @sunhouhua 5 лет назад +1

    Why tanh and not relu?

    • @mohamedaminebenbouali2941
      @mohamedaminebenbouali2941 5 лет назад +1

      because tanh is for Ai with two outputs, for ex jump or don't jump, relu is for Ai with many outputs and tanh is way more accurate cuz it tackles two specifically, relu would still work doe, just not symmetrically.

  • @techmatrix2103
    @techmatrix2103 5 лет назад +2

    Please do a tutorial on real time face detection using ML...

  • @himankjerolia9282
    @himankjerolia9282 4 года назад

    Do we have to make this NEAT config differently for some other game or is it pre compiled and universal for all the games? please can somebody explain

    • @thethisismyaccounful
      @thethisismyaccounful 4 года назад

      Himank Jerolia No you have to change it up for your project

    • @jelle7937
      @jelle7937 4 года назад

      I am not very experienced, but I think that you have to change the amount of inputs and outputs for each project, as he explains 5:30

  • @arthurrobert2135
    @arthurrobert2135 5 лет назад

    Need the next episode

  • @neuron8186
    @neuron8186 3 года назад +1

    i literally learned neural net for reinforcement learning then this module comes in :/

  • @lol_moon0
    @lol_moon0 5 лет назад

    What if I set population on 1?

    • @ApexHarvestR
      @ApexHarvestR 4 года назад +2

      you get a 50/50 chance of getting a good bird or bad bird but if u end up with shit bird it is unlikely to evolve well and will probably become a shitter bird

  • @TheEnderOcelot
    @TheEnderOcelot 2 года назад

    nice

  • @sainco3036
    @sainco3036 5 лет назад

    thanks.

  • @kklol07
    @kklol07 2 года назад

    woah

  • @devonk298
    @devonk298 3 года назад

    Something tells me you might be from Canada..

  • @big123lak
    @big123lak 11 месяцев назад

    I wonder when hospitals are gonna start using the neat package