New Pruning Section to the Course.

[peremartra]

I've added a new section to the course: Pruning of Large Language Models.

I was debating whether to create an advanced section on Agents or this one on Pruning. Although Agents are becoming increasingly relevant, there are already many courses and resources available. On the other hand, there are very few places where you can find examples of model pruning.

Pruning is a technique that allows us to reduce the size of a model by eliminating redundant parts, whether neurons or entire layers. It is one of the most important and widely used steps in model optimization.

Recently, NVIDIA published work explaining how they create versions of LLaMA models that are 50% smaller yet outperform the base model on specific tasks.

In this course, we’ll attempt to replicate NVIDIA's work, starting with simpler examples, gradually leading to a process that enables us to create a smaller model in a similar way.

https://github.com/peremartra/Large-Language-Model-Notebooks-Course/blob/main/6-PRUNING/readme.md

I hope you enjoy it!