federicoarenasl/Regularization-techniques-on-NNs federicoarenasl/Regularization-techniques-on-NNs

  • Stars
    star
    1
  • Language
    Jupyter Notebook
  • Created over 3 years ago
  • Updated over 3 years ago
Twitter logo Share LinkedIn logo Share Facebook logo Share
federicoarenasl/Regularization-techniques-on-NNs
github

federicoarenasl

Reviews

There are no reviews yet. Be the first to send feedback to the community and the maintainers!

Repository Details

During this study we will explore the different regularisation methods that can be used to address the problem of overfitting in a given Neural Network architecture, using the balanced EMNIST dataset.

More Repositories

1

Data-Generation-with-Blender

Step-by-step tutorial on how to create data with Blender for an object detection application, with ressources included.
Python
62
star
2

Evaluating-w-Embeddings

In this paper we compare and evaluate two simple embedding models which can be constructed directly from a given co-occurrence matrix extracted from Twitter data; Positive Pointwise Mutual Information (PPMI), and Hellinger Principal Component Analysis (H-PCA). For each embedding model we consider three alternative metrics for word similarity: cosine, euclidean and manhattan distance.
Jupyter Notebook
4
star
3

SVM-LR-on-Fashion-MNIST

This is a brief tutorial on using Logistic Regression and Support Vector Machines for classification on the Fashion MNIST dataset.
Jupyter Notebook
2
star
4

PCA-on-Fashion-MNIST

This is a concise tutorial on applying PCA in the benchmark dataset Fashion MNIST. I analyse how the data compression process is done in visual information.
Jupyter Notebook
2
star
5

Optmizing-CNNs-w-ResNets

In this repository I explore the effect of applying Residual Connections to a VGG CNN Architecture, as well as applying Batch Normalisation. The networks are tested on the CIFAR100 benchmark dataset.
Jupyter Notebook
1
star
6

BoVW-ResNet18-Robustness-Evaluation

The aim of this project is to explore the classification of images using a ResNet18 Convolutional Neural Network and a Bag of Visual Words (BoVW) method that uses Support Vector Machines (SVM) on a dataset that contains 1200 224 x 224 images of cats and dogs.
Python
1
star