See the attached Loom video to listen audio from the demo I am a Beginner Hacker Tracks I am going for:

  • best overall yay :3 ($100 + snack boxes and random swag)
  • most likely meme startup ($50 + snack boxes and random swag)
  • (LAKOG) lowkey actually kind of good ($50 + a snack)
  • my drip >;D ($100)
  • i laughed. (new hackers prize) (10 winning teams) (10)
  • i’m just a chill guy sponsored by Build City ($200)
  • codedex i wanna learn stuff prize ($1041 value)
  • UI/UX most kawaii ($100)
  • PearAI (YC24) prize ($750)

Inspiration

I like to wear clothes with InsPiRaTiOnAl Quotes. Why not do it with GenAI?

What it does

You upload a picture -> Select the clothing piece you want to write the Quote on -> Complete the Prompt Template -> Press Generate.

How we built it

I used Flask to build the App and used ideogram-v2-turbo model from Replicate API for inpainting and used Hugging Face segformer_b2_clothes model for automatically segmenting the clothes and generating masks for the inpainting task.

Challenges we ran into

  • Had challenges with integrating the masks from the segformer_b2_clothes model with the Replicate API.
  • Bad output quality. Utilized Negative prompts to increase output quality.

Accomplishments that we're proud of

  • Being able to complete the project on time.

What we learned

Learnt how to use Hugging Face models and Replicate API's

What's next for MemeDrip

I want to create my own fine-tuned Inpainting models for creating memes on the clothes.

Built With

Share this project:

Updates