this post was submitted on 16 Jan 2024
37 points (97.4% liked)

Godot

6008 readers
85 users here now

Welcome to the programming.dev Godot community!

This is a place where you can discuss about anything relating to the Godot game engine. Feel free to ask questions, post tutorials, show off your godot game, etc.

Make sure to follow the Godot CoC while chatting

We have a matrix room that can be used for chatting with other members of the community here

Links

Other Communities

Rules

We have a four strike system in this community where you get warned the first time you break a rule, then given a week ban, then given a year ban, then a permanent ban. Certain actions may bypass this and go straight to permanent ban if severe enough and done with malicious intent

Wormhole

!roguelikedev@programming.dev

Credits

founded 2 years ago
MODERATORS
 

Source code: https://github.com/TKperson/badapplenn

It's my first time using Godot, so source code is kinda messy

This project took two weeks of my life. Hope y'all like this

you are viewing a single comment's thread
view the rest of the comments
[–] tkperson@lemm.ee 4 points 11 months ago

Sure, the role of neural network used in this project is recognizing hand written digits. The input is a 81x81 2D grid. There are a few fully-connected layers trying classify all these input pixels into 10 possible outputs and assign each output with a probability. The outputs are the 10 digits in our number system.

This model is trained using pytorch, a machine learning framework, and exported to onnx. The onnx file then gets used by a backend server written in python that will communicate with godot.

The math behind how fully-connected layers work is a little complicated. If you want to get into the details of how exactly machine "learns" you should go watch some youtube tutorials about back-propagation and statistic modeling.

The tricky part here is embedding an entire video (bad apple in this case) into the weights of the fully-connected layers. Training the model with custom-set limitations was really hard. This model's accuracy at guessing a human's handwriting correctly is only about 60% based on the augmented dataset that it was tested on.

I think it's lemmy's parsing issue that the "The" word and the repo link got merged