My students built five models during the final exam. This solution was submitted by Shih-Yu Ma (Shih-Yu Ma (馬世宇)). He finished in 97 minutes with 221 edits. The five models are: 👇

1. AutoEncoder (AE) 2. Convolutional Neural Network (CNN) 3. Transformer 4. U-Net 5. Contrastive Language-Image Pretraining (CLIP) I am pleased that many other students did just as well! If you are my students and are reading this, thank you for studying and practicing hard this semester! 🙏 I believe you are now prepared to do advanced works in AI. You are setting an example for other students that it is possible to understand the math of deep learning and build models from scratch without the help of any framework. To my fantastic TAs and co-instructor, Anna Rahn, Nolan Brady, @JubairTheSloth, Leopold Beuken: Thank you for all your hard works to make this course a huge success! --- How can you support AI by Hand ✍️? 1. Subscribe Newsletter: 2. Repost ♻️

@horse_is_fish where can i get an excel with these calculation to check out ?

@horse_is_fish Please check out:

@horse_is_fish How to become your student so I can do all this with hands-on 😅

@horse_is_fish You can become my students simply by following my posts or newsletter for free! 😀

@horse_is_fish Excellent work, and congratulations on your students. Do you release them online so our students can use them?

@horse_is_fish Hi Reza, Yes. I will clean up and release them online. But I will just email you now.

@horse_is_fish I would like to learn these things, but I don't have the option to attend your classes. Are there any books or some online sources from where I could start at least? Thank you in advance for anyone who can guide me to the right direction.

@horse_is_fish Very nice and helpful work!

@horse_is_fish Very interesting stuff. Book marking to review later.