I Trained an AI with 10,000 Memes

Bright Data: https://brdta.com/codingwithlewis

I tried training an AI model on 10,000+ memes to see if AI can be funny. In this video, we go through the process of creating a meme with AI and learn why we find things like memes and humor funny to begin with.

This video took a very long time to make. Over 10 different models were used. Throughout this video you will see all of the things that I did to get this to work. Let me know what you think in the comments :)

????WE ARE PLANNING A HACKATHON!!! ????
Join the discord to learn more:
https://dsc.gg/lewismenelaws

????‍????DEMO & SOURCE CODE????‍????
Demo: https://memegenerator-ergog3zn3pc7txnm6mgggf.streamlit.app/
GitHub: https://github.com/CodingWithLewis/MemeGenerator (messy lol)
Fine-Tuned Model: https://huggingface.co/codingwithlewis/mistralmemes
OwlV2 Demo: https://huggingface.co/spaces/codingwithlewis/owlv2


If you like this video, check out some of my other videos where I build awesome projects and provide awesome developer resources that you can use in order to be a better developer :)


LINKS
---
MY NEWSLETTER ????
https://thebetter.dev
------

CONNECT WITH ME ON SOCIAL
???? Instagram: https://instagram.com/lewismenelaws

????TikTok: https://tiktok.com/@lewismenelaws

???? Twitter: https://twitter.com/LewisMenelaws
—
My gear ????
https://liinks.co/lewismenelaws
-----

TIMESTAMPS
0:00 Intro
0:22 Why is AI not funny?
0:40 What is a meme... really?
1:51 Getting a LARGE amount of memes
2:32 The Data Collection
3:29 Using Bright Data
4:12 Initial Results
4:45 Prepping for Training
5:11 Getting image descriptions
6:03 Returned image descriptions
6:35 Training the AI Model
7:57 Results... and de-motivation...
8:49 Realizing... I should teach AI how to be funny
9:35 Getting Relevant Context
10:20 Initial Observations
10:37 Labelling Feature
11:27 The right AI model for the job...
12:14 Building out the UI
12:51 Reactions
14:08 Why I made this video...
14:40 SUBSCRIBE