Object Detection Object Detection with PyTorch and Detectron2 In this post, we will show you how to train Detectron2 on Gradient to detect custom objects ie Flowers on Gradient. We will show you how to label custom dataset
Deep Learning Evaluating Object Detection Models Using Mean Average Precision (mAP) To evaluate object detection models like R-CNN and YOLO, the mean average precision (mAP) is used. The mAP compares the ground-truth bounding box to the detected box and returns a
Computer Vision Transpose Convolution Explained for Up-Sampling Images Convolutional neural networks need no introduction when it comes to image processing using Deep Neural Networks (DNNs). CNNs provide a more realistic way to extract and learn features from an
Computer Vision Efficient Channel Attention for Deep Convolutional Neural Networks (ECA-Net) In this article we'll dive into an in-depth discussion of a recently proposed attention mechanism, namely ECA-Net, published at CVPR 2020.
Computer Vision Channel Attention and Squeeze-and-Excitation Networks (SENet) In this article we will cover one of the most influential attention mechanisms proposed in computer vision: channel attention, as seen in Squeeze-and-Excitation Networks (SENet).
Computer Vision 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.
Computer Vision Image Super-Resolution: A Comprehensive Review Image Super Resolution refers to the task of enhancing the resolution of an image from low-resolution (LR) to high (HR). It is popularly used in the following applications: Surveillance: to
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
Computer Vision Attention Mechanisms in Computer Vision: CBAM This post covers an in-depth analysis of the Convolution Block Attention Module (CBAM) with a short introduction to Visual Attention Mechanisms.
Coronavirus Fighting Coronavirus with AI, Part 2: Building a CT Scan COVID-19 Classifier Using PyTorch Using PyTorch, we create a COVID-19 classifier that predicts whether a patient is suffering from coronavirus or not, using chest CT scans of different patients.
YOLO How to Improve YOLOv3 YOLO has been a very popular and fast object detection algorithm, but unfortunately not the best-performing. In this article I will highlight simple training heuristics and small architectural changes that can make YOLOv3 perform better than models like Faster R-CNN and Mask R-CNN.
Series: GauGAN Understanding GauGAN Part 4: Debugging Training & Deciding If GauGAN Is Right For You In this post we cover how to tackle common training issues that may arise with GauGAN. We conclude with advice on whether GauGAN will fit your business needs or not.
Series: GauGAN Understanding GauGAN Part 3: Model Evaluation Techniques Hi folks, welcome to the third part of the GauGAN series. In the last two parts we've covered the components and losses of GauGAN, and how to set up custom
Computer Vision Image Captioning With AI One application that has really caught the attention of many folks in the space of artificial intelligence is image captioning. If you think about it, there is seemingly no way
Series: GauGAN Understanding GauGAN Part 2: Training on Custom Datasets In Part One, we covered the basic components of GauGAN as well as the loss functions it makes use of. In this part, we'll cover the training details and see
Series: GauGAN Understanding GauGAN Part 1: Unraveling Nvidia's Landscape Painting GANs One of the most interesting papers presented at CVPR in 2019 was Nvidia's Semantic Image Synthesis with Spatially-Adaptive Normalization. This features their new algorithm, GauGAN, which can effectively turn doodles
Generative Models A 2020 Guide to Synthetic Media Synthetic media is an exciting new area of research that has seen great advancements in the past few years. This field has the potential to completely revolutionize the way we
Computer Vision Building Image Datasets for Computer Vision Algorithms In my experience, data is one of the most challenging issues that we face in developing any ML/AI pipelines in the field of computer vision. There are two aspects
Generative Models Implementing CycleGAN for Age Conversion If you've been on any sort of social media this year, you've probably seen people uploading a recent picture of themselves right next to another picture of what they'll look
Interviews DeepFake Faceoff w/ Dr. Fakenstein & CtrlShiftFace Update! (December 2019) CtrlShiftFace has a new video Home Stallone produced with Paperspace: Deepfake technology is unlocking a new era of media production. As with all technologies, both positive and
Tutorial Unpaired Image to Image Translation with CycleGAN Yann LeCun, director of AI research at Facebook and Professor at NYU described Generative Adversarial Networks, GANs as the most interesting idea in Machine Learning in the last 10 years.
Deep Learning Building a State of the Art Bacterial Classifier with Paperspace Gradient and Fast.ai One of the great promises of Deep Learning is its applicability in a wide variety of complex tasks. Recent years have seen an explosion in the number of fields Deep
Tutorial Detecting and Localizing Pneumonia from Chest X-Ray Scans with PyTorch Over the years, we have seen very powerful models being built to distinguish between objects. These models keep getting better in terms of performance and latency day by day but
Deep Learning Going beyond torchvision models ResNets, DenseNets and Inception networks are undoubtedly some of the most powerful models out there for performing image classification and object recognition. These models have shown some promising results in
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
