Neural Network - Paperspace Blog

Paper Review: GhostNet (CVPR 2020) in PyTorch and TensorFlow

In this post we'll take an in-depth look at feature maps in convolutional neural networks, do a thorough review of GhostNet, and break down the code in PyTorch and TensorFlow.

Neural Network

Neural Architecture Search Part 4: REINFORCE Gradient and Evaluation

So far we've built two major components of our neural architecture search (NAS) pipeline. In the second part of this series we created a model generator which takes encoded sequences

Neural Network

Capsule Networks: A Quick Primer

Whether you're new to deep learning or a serious researcher, you've surely encountered the term Convolutional Neural Networks. They are one of the most researched and top-performing architectures in the

Deep Learning

A Review of Popular Deep Learning Architectures: DenseNet, ResNeXt, MnasNet, and ShuffleNet v2

The aim of this three-part series has been to shed light on the landscape and development of deep learning models that have defined the field and improved our ability to

Deep Learning

A Review of Popular Deep Learning Architectures: ResNet, InceptionV3, and SqueezeNet

Previously we looked at the field-defining deep learning models from 2012-2014, namely AlexNet, VGG16, and GoogleNet. This period was characterized by large models, long training times, and difficulties carrying over

Deep Learning

A Review of Popular Deep Learning Architectures: AlexNet, VGG16, and GoogleNet

Problems ranging from image recognition to image generation and tagging have benefited greatly from various deep learning (DL) architectural advancements. Understanding the intricacies of different DL models will help you

Neural Network

Neural Architecture Search Part 3: Controllers and Accuracy Predictors

In the first part of this series we saw an overview of neural architecture search, including a state of the art review of the literature. In Part 2 we then

Neural Network

Neural Architecture Search Part 2: Search Space, Architecture Design and One-Shot Training

In the first part of the series, we took a look at all the different angles the problem of neural architecture is being approached from. With the foundation covered, we'll

Research

Neural Architecture Search Part 1: An Overview

The hyperparameter optimization problem has been solved in many different ways for classical machine learning algorithms. Some examples include the use of grid search, random search, Bayesian optimization, meta-learning, and

Keras

Working With The Lambda Layer in Keras

In this tutorial we'll cover how to use the Lambda layer in Keras to build, save, and load models which perform custom operations on your data.

Deep Learning

Image Compression Using Autoencoders in Keras

Autoencoders are a deep learning model for transforming data from a high-dimensional space to a lower-dimensional space. They work by encoding the data, whatever its size, to a 1-D vector.

Tutorial

Generating an interactive Pix2Pix model with Gradient° and ml5.js

This post will go through the process of training a generative image model using Gradient° and then porting the model to ml5.js, so you can interact with it in

Series: Optimization

Intro to Optimization in Deep Learning: Vanishing Gradients and Choosing the Right Activation Function

An look into how various activation functions like ReLU, PReLU, RReLU and ELU are used to address the vanishing gradient problem, and how to chose one amongst them for your network.

Computer Vision

Style Transfer Part 2: Face Swap!

Face Transfer Using Deep Neural Nets Introduction In the previous article we looked at what style transfer was and how to use it to create your own art. Now, we will focus on how to extend the capabilities of style transfer beyond the art

Generative Models

Recurrent Language Models with RNNs

Building an intuition There is a vast amount of data which is inherently sequential, such as speech, time series (weather, financial, etc.), sensor data, video, and text, just to mention some. Recurrent Neural Networks (RNNs) are a family of neural networks designed specifically for