https://github.com/ds1sqe/rust_ml_toolkit

Rust Machine Learing Toolkit from scratch
https://github.com/ds1sqe/rust_ml_toolkit

deep-learning egui framework machine-learning neural-network rust toolkit

Last synced: about 2 months ago
JSON representation

Rust Machine Learing Toolkit from scratch

• Host: GitHub
• URL: https://github.com/ds1sqe/rust_ml_toolkit
• Owner: ds1sqe
• License: gpl-2.0
• Created: 2023-09-26T06:13:37.000Z (over 2 years ago)
• Default Branch: develop
• Last Pushed: 2024-12-18T04:47:40.000Z (over 1 year ago)
• Last Synced: 2025-04-02T11:22:36.733Z (about 1 year ago)
• Topics: deep-learning, egui, framework, machine-learning, neural-network, rust, toolkit
• Language: Rust
• Homepage:
• Size: 6.38 MB
• Stars: 2
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

          
# Rust ML Toolkit

Rust Machine Learning Toolkit from scratch

![manager_main](./doc/img/manager_main.png)

## Purpose

- I created this project to study deep in ml.

- I wanted to learn how various frameworks work.

## Concept

- This project has 3 parts like below

### 1. Core

the core feature of toolkit.

implementation of matrix, neural network,

and various internal methods for that

### 2. Adapter

provide utility feature to user

by interacting with core and

manipulating data

### 3. Manager

GUI layer of project

interact with user

## This project includes

i will call neural network as NN.

### 1. Core

#### 1.1 NN & matrix

1.  matrix operations

2.  CRUD

#### 1.2 learning methods

1.  finite difference

2.  back propagation

### 2. Adapter

#### 1. Context & Session

manage state of model (NN)

save or load model or dataset

#### 2. image

convert img to 2d matrix and vice versa

![png2mat](./doc/img/png2mat.png)

### 3. GUI manager

1.  cost info

    ![cost_view](./doc/img/cost_view.gif)

2.  network view & node state info

    ![node_view](./doc/img/node_view.gif)

# How to run this project?

clone this project and just

```bash

cargo run

```

will launch the GUI manager