W&B Events for the Week of November 1
Publish your model insights with interactive plots for performance metrics, predictions, and hyperparameters. Made by Rajne Kumari using Weights & Biases
Nov 2, 10:00am PT - Paper Reading Group: EfficientNet: Rethinking Model Scaling for Convolutional Neural Networks
In this event along with our fantastic host Aman Arora we explored the “Why” and the "How" section of the paper
based on EfficientNets by Tan et al.
Particularly, we looked at two main contributions from the research paper:
1. Compound Scaling
2. The EfficientNet Architecture (developed using Neural Architecture Search)
We compared the conventional methods of scaling with Compound Scaling approach in Comparing Conventional Methods with Compound Scaling and finally looked at the details of the EfficientNet Architecture in The EfficientNet Architecture using NAS and learn how the authors used Nerual Architecture Search (NAS) to get EfficientNet-B0 architecture and scaled it using Compound Scaling technique to get EfficientNet B1-B7.
Nov 4th, 10am PT - W&B AMA w/Lukas
We’re excited to resume our AMA series. We kicked off the first one with our CEO Lukas Biewald.
Head to our community forum to see what questions the community asked about running startups, building ML models, robots, magnets & more.
Nov 4th, 10:00am PT - Optimizing ML Workflows - how to spend more time on models!
Have you ever built (maybe even deployed!) an amazing model, but when it came time to share you couldn't reproduce the same results? Or maybe you've wondered: why is my model making these strange predictions? In this webinar, we covered how W&B can help you spend more time on models — and less time on ops!
Reach out to us to request the recording.