TLV DL Bootcamp Revival #1: Demystifying text to image magic 🪄🖼️

Our 1st meetup in the revival series of the Tel Aviv Deep Learning Bootcamp will be run by the venerable Naomi Ken Korem (@ Lightricks) and hosted by AWS Floor28!🥕

The recent success of OpenAI's Dall-E2, Stable diffusion released model, and midjourney, that enabled generating images given text prompts, attracted a lot of attention.

In this hands-on, bring your on laptop (!) session we will try to demystify the text to image magic and help you understand how these models work, while also covering the main innovations that enabled this huge progress! 🧇

Location: AWS Floor28 · Derech Menachem Begin 121 · Tel Aviv-Yafo

𝗔𝗴𝗲𝗻𝗱𝗮:
🍕 18:00 - 18:15 — Mingling, etc.
🔶 18:15 - 19:00 — Talk: Making the next jump in the image generation domain
🔷 19:00 - 19:30 — Workshop: Training a mini-diffusion model
🚽 19:30 - 19:40 — Break
🔶 19:40 - 20:20 — Talk: Text guided diffusion models
🔷 20:20 - 21:00 — Workshop: Text guidance in Diffusion models

***
Part #1: Making the next jump in the image generation domain

Abstract: My talk will cover last year's progress in the field of image generation. I will mention some seminal works in the field of image generation, and I will explain the idea of diffusion models. Then, in the corresponding hands-on workshop you will train a mini diffusion model! 😵

***
Part #2: Text guided diffusion models

Abstract: The second talk will discuss how text is added to DDPMS, how to increase the text effect in the generated image, and how Dall-E2 and Imagen work. We will also talk about Stable diffusion. Then, in the corresponding hands-on workshop, you will implement the text guidance trick, and you use stable diffusion to edit your own images using text prompts! 😱