Blog posts

2022

Different implementations of the ubiquitous convolution

less than 1 minute read

Published:

Throughout the Deep Learning field, nn.Conv2d is being used left-right-center for building efficient convolution layers in PyTorch without worrying much about how they are implemented under the hood. In this post, we will specifically gain some insights into different convolution implementations like a naive nested for-loop, Im2Col, Winograd, Strassen and FFT algorithms and infer their pros & cons based on latencies incurred on a N1 CPU and a T4 GPU. We will also relate Strassen’s algorithm to DeepMind’s recent computing breakthrough with AlphaTensor.

Unboxing ChatGPT: A Deep-Dive on How This AI-Driven Chatbot Was Trained

less than 1 minute read

Published:

ChatGPT, OpenAI’s latest dialogue model, has taken the internet by storm, surpassing 1 million users in just 5 days. From seamless chatting to creating poetry and from writing code to conceiving an imaginary OS, its performance is truly mind-blowing. How did conversational AI become so much better so quickly? OpenAI appears to have cracked the nut using Reinforcement Learning with Human Feedback (RLHF) – a method that uses human demonstrations to guide the model toward desired behavior. In this article, we’ll unpack ChatGPT’s training techniques and take a deeper look at what goes on under the hood. Find the wandb article written by me here.

2020

Augmented Reality, the reality disrupter

less than 1 minute read

Published:

Reality is becoming more and more elusive in our lives. Unless you have been living under a rock, you would have heard of Augmented Reality (AR), the technology that drives Snapchat filters, Pokemon Go, IKEA furniture place app etc. which superimposes a computer-generated image on a user’s real world view. This post is a personal take on the technological and business trends in the AR field. It might be the next biggest disrupter that seamlessly amplifies human capabilities.