Welcome to our tutorial presented at the ACM MultiMedia conference 2024, happening in Melbourne, Australia from October 28th to November 1st 2024.
📅 Slides Availability: Slide are available here
For a visual summary of what this tutorial covers, check out our outline:
Explore the Jupyter Notebooks we used during the tutorial in our Notebooks Folder. These interactive notebooks are perfect for hands-on learning and are also available on Google Colab:
- d-Simplex with K pre-allocated classes
- Effects of HOC loss during model fine-tuning
- Large Model Replacements
Access our notebooks directly on Google Colab from the following links to dive right into the code:
- d-Simplex with K pre-allocated classes
- Effects of HOC loss during model fine-tuning
- Large Model Replacements
This repository is inspired by groundbreaking work in the field of backward-compatible representations. Check out these resources for more insights:
🤝 Join us at ACM MultiMedia 2024 to learn how we can achieving compatibility in evolving machine learning models without sacrificing performance!