An AI clothes recommender trained on TikTok best fashion influencers
The current TikTok's recommendation algorithm is using inputs on how users interact with videos to recommend products from TikTok shop, which is already very effective. ClothesTok hopes to elevate this even more and enhance users' tailored discovery of clothing shops by taking the user's visual features into consideration when recommending clothes products.
Clone this repository
git clone https://github.com/weebao/clothestok.git
Install the necessary modules for the backend
cd backend
pip install -r requirements.txt
Do the same for the frontend
cd frontend
npm install
To run this app in development mode, you will need to run both the backend and the frontend at the same time locally.
- Running the backend:
cd backend
uvicorn main:app --reload
- Running the frontend:
cd frontend
npm run dev
- Backend: FastAPI, SegFormer, Torch, Scikit-learn, Deployed on Render - 512 MB RAM so it might not run if too many people are using
- Frontend: Next.js, TailwindCSS, Deployed on Vercel
- APIs used: TikTok's API, HuggingFace's API
Special shoutout for @levihsu for the powerful OOT diffusion model which was a great help when building our products!