Karthik Yearning Deep Learning

tflite inference in raspberry pi

Raspberry pi OS pi@raspberrypi:~ $ cat /etc/os-release PRETTY_NAME="Raspbian GNU/Linux 11 (bullseye)" NAME="Raspbian GNU/Linux" VERSION_ID="11" VERSION="11 (bullseye)" VERSION_CODENAME=bullseye ID=raspbian ID_LIKE=debian HOME_URL="http://www.raspbian.org/" SUPPORT_URL="http://www.raspbian.org/RaspbianForums" BUG_REPORT_URL="http://www.raspb... Read more

Machine Learning and Deep Learning comparison

A Comparison of Traditional Machine Learning and Deep Learning in Image Recognition This paper compares the accuracy by using machine learning and the Deep convolution neural network. The paper used a computationally simpler DCNN architecture. The ML technique: Bag of words model in computer vision In computer vision, the bag-of-words mo... Read more

Siamese Network for duplicate detection

In this post, I will list out the application of Siamese network from the paper Using Siamese CNNs for Removing Duplicate Entries From Real-Estate Listing Databases. Real estate databases are geo-specific. If a house to be put up for sale is located close to the geo boundary, a real estate listing agent will often list it in both databases. F... Read more

Low dimension neural network training

Deep neural networks can be optimized in randomly-projected subspaces of much smaller dimensionality than their native parameter space. While such training is promising for more efficient and scalable optimization schemes, its practical application is limited by inferior optimization performance. Here, we improve on recent random subspace appro... Read more

Gradio Tutorial

I recently had a POC to present to my team. It was an image similarity project in the retail space. I was looking for a simple tool to demo the top N similar image for a given query image. I did not want to spend much time in learning the tool, but I started with Streamlit, since it was famous than Gradio. Streamlit is a good package, which prov... Read more

Segmentation data annotation cvat

In this article, I will walk you through the process of annotating images for instance segmentation using the cvat.org website. Before we begin I will give you a brief introduction about instance segmentation. Segmentation of each instance of objects in the image is called instance segmentation. In the below dataset, the application was to segm... Read more

Serving Tensorflow to 300 million predictions per second

This paper presents the process of transitioning machine learning models to the Tensorflow framework at a large scale in an online advertising ecosystem. I will outline the important practices to follow while training and serving a machine learning model The process of scaling machine learning models implemented in the Tensorflow machi... Read more

Statistically stopping of neural network training

Paper: Statistically Significant Stopping of Neural Network Training Github: Code Repository Much learning of neural network does not take place once the optimal values are found, the condition does not impact the final accuracy of the model. According to the runtime perspective, this is of great significance when numerous neural networ... Read more

Visualize feature maps

In this article, I will visualize the feature maps in a neural network. Since the focus of this article is to visualize the feature maps, I am using a tutorial neural network training script from PyTorch official website. This tutorial uses Transfer learning with Resnet50 architecture. The complete tutorial script can be found here. Visualizin... Read more