How agents execute human instructions with varying degrees of agency, from predefined steps to complete autonomy. Smolagents, a Huggingface library, which uses CodeAgent to enhance agent capabilities.
Author: krsnewwave
I'm a software engineer and a data science guy on recommender systems, natural language processing, and computer vision.
Categories
Team Recos
Hey team, I’ll miss the standups and hangouts. Keep in touch and let’s hang out soon! Here’s a playlist for your enjoyment! And also a team pic, courtesy of gen-ai.
Categories
Happy 1st Year in Canada
We’ve been here for a year now and how time flies. As we grow stronger as a family every day, we find new things to be truly grateful for in this new chapter of our life. In this blog, I sit down and write some reflections on the year past. The Greats As you might […]
Categories
An AI Looks for Friends
In this post, I had an idea for a story where I had ChatGPT help my creative process. It’s a fun experience. With some tweaks and regenerating some responses, I present to you a human-AI collaborative story. Please enjoy!
Categories
Reactions on Generative Models
Generative models like ChatGPT and Stable Diffusion have made waves around the world. Experts on machine learning and even different fields are weighing in on the implications of such powerful models. How will these change industries? I’ve talked about it in a previous blog, but now, more than ever, ethicists are required to bridge the […]
I just thought to write this to celebrate my parting from my longest material possession, my PC. It has been a good 10 years. It’s closing time, finally. Modifications done on this rig 2012 – Start of service. Bought a BIG chassis to support whatever fantastic thing I can think of. Machine learning, dual boot […]
I’m going to use NVTabular with PyTorch Lightning to train a wide and deep recommender model on MovieLens 25M. It’s quite a chimera implementation as you shall see.
Categories
NVIDIA Merlin on a Potato PC Setup
NVIDIA Merlin is a high-level open-source library to accelerate recommender systems on NVIDIA GPUs at scale. I’m on a potato PC setup relying on free instances. Can I make it work?
Header image note: I typed ‘variation’ in Pexels and it popped up. Masarap! We can turn a simple autoencoder into something sophisticated. Autoencoders discover a latent mapping z, which as a lower-dimensional representation of the input x, can be useful for pre-training networks and creating recommender models. However, how autoencoders compress information may come in […]
It is the REAL DYL 