Neural Networks: Zero to Hero - Activations & Gradients, BatchNorm.

Building makemore Part 3: Activations & Gradients, BatchNorm

We following a course created by Andrej Karpathy on building neural networks, from scratch, in code.

We start with the basics of backpropagation and build up to modern deep neural networks, like GPT. In my opinion language models are an excellent place to learn deep learning, even if your intention is to eventually go to other areas like computer vision because most of what you learn will be immediately transferable. This is why we dive into and focus on languade models.

Prerequisites: solid programming (Python), intro-level math (e.g. derivative, gaussian).

---

This is forth event from this series.

We dive into some of the internals of MLPs with multiple layers and scrutinize the statistics of the forward pass activations, backward pass gradients, and some of the pitfalls when they are improperly scaled. We also look at the typical diagnostic tools and visualizations you'd want to use to understand the health of your deep network. We learn why training deep neural nets can be fragile and introduce the first modern innovation that made doing so much easier: Batch Normalization. Residual connections and the Adam optimizer remain notable todos for later video.

If you was absent on previous you can watch all previous lessons.

Full curse is presented on page:
https://karpathy.ai/zero-to-hero.html

---

✅ We will follow course program, discuss code, explain fragments that are unclear and learn together.
✅ After course there will be time to eat dinner making new connections in AI worlds and sharing what you think.

Basically you can just view video at your home, but learning in group you can grasp some ideas faster asking questions and learn deeper explaining something that you understand to others.