Skip to content

Latest commit

 

History

History
64 lines (55 loc) · 2.61 KB

README.md

File metadata and controls

64 lines (55 loc) · 2.61 KB

Deep Learning Notebooks

This repository contains a collection of notebooks that I used in tutorials and training.
The list is ordered from the most recent to the oldest

Image Spam Detection Using CNN and Transfer learning

An Keras 3 implementation of part of the Master's project entitled Image Spam Classification using Deep Learning.
You can run this notebook using either TensorFlow or JAX backend.

UK & Ireland Accent Recognition Using Transfer Learning

This notebook uses transfer learning feature extraction from Yamnet audio event classsifier in order to train a model to recognise the accent of a UK or Irish speaker.

Structured Data Classification using TensorFlow Decision Forests

This is an example on how to use TensorFlow Decision Forests on structured data. This notebook was created as part of Kaggle's Tabular Playground Series - Sep 2021 (https://www.kaggle.com/c/tabular-playground-series-sep-2021)

Telecom Customer Churn (Binary Classification)

This is a binary classification problem to detect Telecom customers who are probably going to churn. It uses the following dataset on Kaggle: https://www.kaggle.com/blastchar/telco-customer-churn

Convolutional Neural Networks Tutorial 2

This notebook will show how to:

  • Load a pretrained model (VGG-16)
  • Use it for prediction

Convolutional Neural Networks Tutorial 1

This is a simple notebook that will train the CIFAR10 dataset using a convolutional neural network
It will teach how to:

  • Load and Preprocess data
  • Create a Sequential Model
  • Compile and view the model
  • Calculate the number of parameters in the model
  • Train
  • Visualise training results
  • Test the model on an uploaded image
  • Compare CPU and GPU training times

Standard Neural Networks Tutorial 2

This notebooks's aim is to show how to create a custom layer. It uses the MNIST dataset.

We will create a layer that we will call Linear which is the same as the Keras Dense one i.e. calculates y = wx + b
This notebook will teach how to:

  • Load and Preprocess data
  • Create a custom layer
  • Use Keras Functional Model
  • Create a Callback to stop a model training when accuracy reaches a desired value

Standard Neural Networks Tutorial 1

This is a simple notebook that will train the MNIST dataset using a standard neural network.
It will teach how to:

  • Load and Preprocess data
  • Create a Sequential Model
  • Compile and view the model
  • Calculate the number of parameters in the model
  • Train
  • Visualise training results
  • Predict an uploaded image