Abbie Petulante
In this notebook, we dive deep into the architecture of LLaVA, with the goal of fine-tuning it (both with and without LoRA) to adapt it to determining the morphologies of GalaxyZoo2 images.
Mathematical foundations and practical implementation of LoRA, one of the most common paremeter efficient fine tuning methods.
An in-depth guide to reinforcement learning using proximal policy optimization (PPO), with a focus as it applies to modern large language models (and with human feedback!).
A guide to how prompt caching - a modular approach to KV caching. We’ll talk through how and why this modular implementation works, and include a practical example of implementing it for LLaMa 3.2 1B.
A guide to how KV caching is implemented for LLMs, including a practical example of implementing it for LLaMa 3.2 1B.
A complete guide to the U-Net architecture. Covers all aspects of a layer, from the convolution operation, to pooling, to what makes U-Nets so special.
