NFNets are faster than EfficientNets and they don’t use normalization — Our smaller models match the test accuracy of an EfficientNet-B7 on ImageNet while being up to 8.7× faster to train, and our largest models attain a new state-of-the-art top-1 accuracy of 86.5%. Source: arxiv One of the most annoying things about training a model is the time it takes to…

A comparison between the latest versions of PyTorch (1.8) and Tensorflow (2.5) — Tensorflow/Keras & Pytorch are by far the 2 most popular major machine learning libraries. Tensorflow is maintained and released by Google while Pytorch is maintained and released by Facebook. In this article, I want to compare them in terms of: What's new in the latest released versions Which one to…

A UCL project to solve the problem of tedious digital pathology labeling — In this article, I am going to talk about one of my dearest and best projects that I have worked on, my undergraduate thesis at University College London. My thesis was about using unsupervised machine learning to automate the process of breaking down huge Whole Slide Images and labeling them…

What are ConvNexts and how do you use them — Introduction Convolutional Neural Networks (CNNs) are a type of neural network architecture that are commonly used in image and video recognition tasks. Convolutional layers are the building blocks of CNNs and are responsible for detecting features in the input data. ConvNeXTv2 is a convolutional neural network architecture that is an improvement…

VOI lookup tables have amazing uses in medical imaging — Introduction A VOI (Value of Interest) lookup table, also known as a transfer function, is a tool used in medical imaging to map the intensity values of a given image to new values that are more suitable for visualization. In medical imaging, the intensity values of an image often do not…

A close look at TensorRT — Advanced TensorRT features Introduction TensorFlow is a great way to build AI models, but it’s not the only way. In this post, I’ll explain why you might want to use TensorRT and how you can use it to improve your models’ performance and reduce power consumption. TensorRT is a high-performance deep-learning inference library developed…

Google uses much more AI than you think — Introduction Google is a company that’s always looking for ways to use artificial intelligence (AI) and machine learning to make its products even better. And it shows — the search giant has built AI into many services you probably use every day, from Gmail to Maps. So here are just some…

Amazon is one of the biggest tech giants, which of their products uses AI? — Introduction Machine learning and artificial intelligence have changed the way we live. Amazon has been at the forefront of this change, leveraging machine learning algorithms to power its recommendation engines and build voice assistants like Alexa. In the past few years, it’s become clear that AI is only going to continue…

From GPT-1 to GPT-3 — The evolution of natural language processing has come a long way since the introduction of the first GPT model, GPT-1. What are language models and transformers? Language models are used in machine translation systems to learn how to map strings from one language to another. For example, if you want to translate “I like dogs” from…