https://github.com/ahmetfurkandemir/deep-learning-4-regularization

Deep Learning(4) - Regularization
https://github.com/ahmetfurkandemir/deep-learning-4-regularization

Last synced: 6 months ago
JSON representation

Deep Learning(4) - Regularization

• Host: GitHub
• URL: https://github.com/ahmetfurkandemir/deep-learning-4-regularization
• Owner: AhmetFurkanDEMIR
• Created: 2020-07-29T09:06:24.000Z (about 5 years ago)
• Default Branch: master
• Last Pushed: 2020-07-29T09:33:52.000Z (about 5 years ago)
• Last Synced: 2025-04-15T21:52:18.613Z (6 months ago)
• Language: Jupyter Notebook
• Homepage:
• Size: 3.56 MB
• Stars: 13
• Watchers: 2
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

          
# Regularization (Düzenleştirme)

* Derin Öğrenme modelleri o kadar fazla esnekliğe ve kapasiteye sahiptir ki, eğitim veri seti yeterince büyük değilse aşırı öğrenme ciddi bir sorun olabilir. Elbette eğitim setinde iyi sonuç veriyor, ancak öğrenilen ağ, daha önce hiç görmediği yeni örneklere genelleme yapamoyor!

* Şunları öğreneceksiniz: Derin öğrenme modellerinizde düzenlemeyi kullanmak.

* Yapacağımız adımlar : 

    

     * Düzenlenmemiş model

     

     ![0](https://user-images.githubusercontent.com/54184905/88782199-f93bce80-d195-11ea-90a1-065679a75aaa.png)

     

     ![index](https://user-images.githubusercontent.com/54184905/88781963-b4179c80-d195-11ea-80ab-7a7859d960be.png)

     * L2 Düzenleme

     

     ![1](https://user-images.githubusercontent.com/54184905/88782204-f9d46500-d195-11ea-8f77-233bd9cb653c.png)

     

     ![index1](https://user-images.githubusercontent.com/54184905/88781959-b37f0600-d195-11ea-92af-3780392d6a6b.png)

     * Dropout

     

     ![2](https://user-images.githubusercontent.com/54184905/88782207-fa6cfb80-d195-11ea-9ce1-1eb1ffc63f49.png)

     

     ![index3](https://user-images.githubusercontent.com/54184905/88781955-b24dd900-d195-11ea-8ba0-89fee5a299cc.png)

* Kaynak : Improving Deep Neural Networks: Hyperparameter tuning, Regularization and Optimization (Coursera) - https://www.coursera.org/learn/deep-neural-network