https://github.com/slamko/deep-racket

Little Deep Learning framework in Racket
https://github.com/slamko/deep-racket

ai deep-learning functional-programming lisp machine-learning neural-network racket

Last synced: about 1 month ago
JSON representation

Little Deep Learning framework in Racket

• Host: GitHub
• URL: https://github.com/slamko/deep-racket
• Owner: slamko
• Created: 2023-07-08T23:28:01.000Z (over 1 year ago)
• Default Branch: master
• Last Pushed: 2023-07-15T19:17:47.000Z (over 1 year ago)
• Last Synced: 2024-11-20T19:36:00.577Z (3 months ago)
• Topics: ai, deep-learning, functional-programming, lisp, machine-learning, neural-network, racket
• Language: Racket
• Homepage:
• Size: 83 KB
• Stars: 0
• Watchers: 2
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.org

Awesome Lists containing this project

README

        
*** Little deep learning framework based on Perceptron algorithm

Implemented in Typed Racket for better performance with Matrix library

Library part lives in deep.rkt file:

To try it out in your racket code:

#+begin_src lisp

#!/usr/bin/env racket

#lang racket

(require "deep.rkt")

(require math/matrix)

(define xor-output

  (list

    (matrix [[0]])

    (matrix [[1]])

    (matrix [[1]])

    (matrix [[0]])))

(define xor-input

  (list 

    (matrix [[0 0]])

    (matrix [[0 1]])

    (matrix [[1 0]])

    (matrix [[1 1]])))

(define nn (make-nn '(2 5 1)))

(define trained-nn (learn nn xor-input xor-output 1 (* 10 1000)))

(perform trained-nn xor-input xor-output)

#+end_src

app.rkt provides similar interface from the command line:

#+begin_src sh

./app.rkt 10000 -a 2 5 1

#+end_src