Synopsis: Deploy a trained ConvNeXt-B model instance on TorchServe for realtime image classification using the categories present in the Food 101 dataset. TorchServe is a framework for serving pyTorch trained models in a seamless, performant manner. In this tutorial, we will walk though the steps for deploying a PyTorch model…

Examining a module consisting of several attention layers running in parallel. — This post refers to the Transformer network architecture (paper). The Transformer model represents a successful attempt to overcome old architectures such as recurrent and convolutional networks. We will only deal with a small Transformer component, the Multi-Head Attention module. The Transformer architecture The first application of Transformer model was language translation. Very briefly…

Multi-Modality is the way forward for efficient and scalable solutions in the Transformer Era — Introduction The famed Transformer [1] architecture released in 2017 was simply outstanding and went on to break almost all SOTA benchmarks in Natural Language Processing. It was then adapted to other domains like Image & Video classification, Object Detection, Audio classification, and even in Generative networks without any major modifications. One…

A Gentle Guide to the Transformer under the hood, and its end-to-end operation. — This is the second article in my series on Transformers. In the first article, we learned about the functionality of Transformers, how they are used, their high-level architecture, and their advantages. In this article, we can now look under the hood and study exactly how they work in detail. We’ll…

CVPR 2021 oral paper — Transformers, proposed originally from natural language processing field, have been increasingly used in computer vision field. In the recent CVPR 2021, some new transformer-based methods had been proposed. In this article, let’s learn together a CVPR 2021 oral paper: UP-DETR: Unsupervised Pre-training for Object Detection with Transformers. …

I am on LinkedIn, come and say hi 👋 Hi guys, happy new year! Today we are going to implement the famous Vi(sion) T(ransformer) proposed in AN IMAGE IS WORTH 16X16 WORDS: TRANSFORMERS FOR IMAGE RECOGNITION AT SCALE. Code is here, an interactive version of this article can be downloaded…

“An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale” May Bring Another Breakthrough to Computer Vision — Lives on earth face a cycle of rise and fall. It is applicable not only for creatures but also for technologies. Technologies in data science have been filled with hypes and biased success stories. Having said that, there are technologies that have lead to the growth of data science: Convolutional…

A complete guide to Facebook’s Detection Transformer (DETR) for Object Detection. Introduction The DEtection TRansformer (DETR) is an object detection model developed by the Facebook Research team which cleverly utilizes the Transformer architecture. In this post, I’ll go through the inner-workings of DETR’s architecture to help provide some intuition on it’s…