Sometimes there are scenarios where you have way lesser number of samples for some of the classes where as other classes have lots of samples. In such a scenario, you don’t want a training batch to be contain samples just from a few of the classes with lots of samples.
Recently, while working on a video dataset, I noticed that some of the videos contained a few corrupted frames. While running the training, my dataloader used to return an incorrect shaped tensor since it was not able to read the corrupted frame.
In this post we will talk about generating synthetic data from tabular data using Generative adversarial networks(GANs). We will be using the default implementation of CTGAN  model.
Introduction In the last post on GANs we saw how to generate synthetic data on Synthea dataset.
Over the last couple of months, I have been going through a lot of literature about human action recognition using computer vision. In this post, I will share a brief survey of Human Action Recognition.
In this post, we will see how to generate tabular synthetic data using Generative adversarial networks(GANs). The goal is to generate synthetic data that is similar to the actual data in terms of statistics and demographics.
In this notebook I am going to re-implement YOLOV2 as described in the paper YOLO9000: Better, Faster, Stronger. The goal is to replicate the model as described in the paper and train it on the VOC 2012 dataset.
Custom Data Generator, Layer, Loss function and Learning Rate Scheduler In this post, I will demonstrate how you can use custom building blocks for your deep learning model. Specifically, we will see how to use custom data generators, custom Keras layer, custom loss function, and a custom learning rate scheduler.
In this notebook I am going to implement YOLOV1 as described in the paper You Only Look Once. The goal is to replicate the model as described in the paper and in the process, understand the nuances of using Keras on a complex problem.