13th PyData Lisbon Meetup

Heya everyone,

The next PyData Lisbon Meetup is just around the corner! This time we're partnering with Feedzai who will sponsor the event and host it in their office in Parque das Nações. We have fewer spots than usual, so we request you to unregister if you know you won't make it.

We'll have two great presentations followed by the usual pizza, drinks, and networking. Check the details below 👇

Aequitas Flow: A Fair ML optimization framework - Sérgio Jesus (Feedzai)
Join us for an overview of Aequitas Flow, the latest version of an open-source package designed to evaluate fairness in Machine Learning. This session will introduce both the basic concepts of Fair ML, and the new features that empower users to create their custom experiments with a variety of bias mitigation techniques. With a strong focus on extensibility and reproducibility, Aequitas Flow allows researchers and practitioners to easily integrate their own methods and datasets, while providing several checkpoints throughout the experiments. Whether you're new to Fair ML or looking to deepen your expertise, this presentation will provide the insights and tools you need to start building fairer models for the future.

Llama to Llama 3.1 -- a year and a half of open-access LLMs in retrospective - João Gante (HuggingFace)
Llama was originally released in February 2023, sparking a revolution in non-commercial usage of LLMs. However, Llama's quality was subpar when compared to the best closed LLMs. Fast-forward to July 2024, and Llama 3.1 is very close to the best closed LLMs, despite their significant improvements. How did we get here? This talk will cover recent major advances in LLMs, focusing on the evolution of the Llama family.

Program:
18:00 - 18:30: Walk-in
18:30 - 19:15: “Aequitas Flow” by Sérgio Jesus
19:15 - 19:30: Quick Break
19:30 - 20:30: “Llama to Llama 3.1” by João Gante
20:30 Onwards: Networking, Pizza and Drinks

Looking forward to seeing you,
Telmo @ Loka