Making a simple AI for my raycasting game in Python Pygame - Tutorial Devlog artificial intelligence

[@mrdinoking6626]

Too good !! I would love if you do a 3d from scratch series in python (true 3d)

[@FinFET]

That's something I'm wanting to do for some time, probably my next project when I finish this one

[@nononononononononono2045]

im interested in using deep neural network for ai

[@PeritoProducciones]

Thank you! This is very useful!

[@matthewmatkava133]

Very goooood

[@aintNexus]

Very nice!

[@SkyFly19853]

Oh...
That reminds me monsters from Doom 1 and 2... 😏

[@Diana2-2-2]

Ohh basecly how to make a phloger main -pyro main

[@FinFET]

😂

[@senex_0372]

Is it very complex make a pygame project in 3D?
Because i’m new on pygame, and i made only 2D projects

[@FinFET]

It is a bit harder when compared to game engines like Unity, even more when you do everything from scratch like on my projects, but there is an alternative to pygame better suited for 3d games: Ursina engine. I don't have much experience with it though

[@senex_0372]

@@FinFET cool! Thanks for the reply
Only 1 question, but it is not a bit laggy? I seen you don’t use the pygame screen function you use an alternative one, don’t you?

[@FinFET]

@@senex_0372 Yes, I draw everything except sprites with a custom numba optimized function, also the resolution is scaled, this way it performs better

[@Benavides1283]

I can’t get it out of my mind. I will try to utilize machine learning in a game.

[@FinFET]

I actually tried it out recently, it is pretty interesting!

[@Benavides1283]

@@FinFET can you tell me about you’re attempt?

[@FinFET]

@@Benavides1283 First I tested with the enemies, learning from the old AI, so it behaved more less the same: th-cam.com/video/7GQEV7disi4/w-d-xo.html
Then I tried with the player, taking turns where i gather data from the player and then train the neural network: th-cam.com/video/PfPVUXYgeBQ/w-d-xo.html

[@Benavides1283]

@@FinFET I’m having issues figuring it all out.

[@FinFET]

@@Benavides1283 Super summarized: you need input data, a model and output data
- usually input data reflects the environment where the player is inserted
- output data are the possible actions for the player
- a model turns the inputs into outputs
you will need to train the model, for that you have to have records of inputs and respective outputs
I would recommend learning the basics of machine learning with scikit learn first, I learned mostly from sentdex some time ago, but there are many good videos around youtube