Skip to content

Combinations of "optimizers and activation functions" were created with Numpy and their performances were compared.

Notifications You must be signed in to change notification settings

oomeryk/Creating_Neural_Network_using_just_Numpy

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

7 Commits
 
 
 
 

Repository files navigation

Neural_Network_only_using_Numpy

-EN

The MNIST dataset in this project, contains 70,000 handwritten digit images, 60,000 for training and 10,000 for testing, each grayscale with pixel values ​​from 0 to 255.

"ReLU and sigmoid" activation functions have been implemented. As optimizers, "SGD, Momentum, Adam and RMSProp" optimizers are available.

Combinations of optimizer and activation functions were created and their performances were compared.

Additionally, "L1 and L2 regularization" was applied and performance comparison was made.

-TR

Bu projedeki MNIST veri seti, her biri 0'dan 255'e kadar piksel değerlerine sahip gri tonlamalı, 60.000'i eğitim ve 10.000'i test olmak üzere 70.000 el yazısı rakamlı görüntü içeriyor.

"ReLU ve sigmoid" aktivasyon fonksiyonları uygulanmıştır. Optimizer olarak da "SGD, Momentum, Adam ve RMSProp" optimizer mevcuttur.

Optimizer ve aktivasyon fonksiyonları kombinasyonları oluşturulup performansları kıyaslanmıştır.

Ayrıca "L1 ve L2 regularizasyon" uygulanıp performans kıyaslama yapılmıştır.

image

image

About

Combinations of "optimizers and activation functions" were created with Numpy and their performances were compared.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published