Heartbeat Newsletter: Volume 28
Happy New Year!
Dear Heartbeat Readers,
Happy 2023! We hope you’ve been able to get some rest and reset for the new year. The beginning of 2023 brings with it a shift in focus for Heartbeat and we’re excited to dive deeper into Comet related tutorials, more deep learning content, and some NLP and computer vision projects. Submissions are once again open and you can learn more on our Call for Contributors page.
The cadence of pieces will also be slowing but that’s so we can focus on bringing you the highest quality content possible! Also be sure to follow Comet’s Twitter and LinkedIn so you’re up-to-date on all the new product releases, in-person and virtual events, job listings, and more. And don’t forget to subscribe to Deep Learning Weekly for the best in DL delivered to your inbox each week.
Happy Reading,
Emilie, Abby & the Heartbeat Team
2022 Comet Wrap-Up
2022 was a big year for Comet! Not only did we release our very first open-source tool, but we were named a Gartner Cool Vendor, expanded the team by 50+ folks, and added multiple integrations. Here’s a wrap-up of the most exciting year in Comet’s history!
Heartbeat 2022 Wrap-Up
This year we published roughly 318 articles from about 72 international writers. This was huge growth from the previous year and couldn’t have been done without Abby Morgan and the Heartbeat team and all the talented members of our community. To celebrate, we compiled the top ten most read articles published this year.
Exploring Image Classification With Kangas
— by Oluseye Jeremiah
Data developers have always had trouble working with large datasets, and this difficulty still exists in the modern data industry. It is no longer necessary to worry about how to run a dataset with millions of images without crashing the notebook because the team at Comet have created a software that enables you to load enormous volumes of multimedia images: Kangas. Millions of images can be loaded simultaneously, and you can also sort, group, and visualize everything at once.
Pythae + Comet
The Pythae library, which brings together many Variational Autoencoder models and enables researchers to make comparisons and conduct reproducible research, is now integrated with Comet ML! The Comet ML experiment tracking tool is very useful for researchers to store their experiment configs, track their training, and compare the results in an easy and understandable way through a visual interface.
Using the EfficientNet Architecture to Classify Car Brands
— by Harpreet Sahota
How do you measure how big a convolutional neural network is? You can’t weigh it or use a ruler to measure it. And if you can’t measure it…then how can you scale it? Until 2020, the process of measuring a convolutional neural network was never well understood. That is until researchers set out to answer an important question: Is there a principled method to scale up ConvNets, so they achieve better accuracy and efficiency?
