Draw2Photo

Inspiration

This tool was born from the simple idea of taking a drawing of a person, adding their real face, and seeing them as a real photo.

What it does

Draw2Photo transforms a simple drawing and a face photo into a single, photorealistic image. You just upload your drawing, add a face. Then the model turns your drawing to a photo with the identity.

How we built it

We used FLUX.1 Kontext [dev] as base model We trained a specialized LoRA. For lora training, we made a special dataset for this. We built a interactive Gradio app and hosted it on Hugging Face Spaces so anyone can try it instantly.

Challenges we ran into

Creating perfect dataset was challenging. We need to draw person drawing manually as I didn't found this type specific dataset for my need. Creating Gradio web app was also challenging in this short time. But at last, the results went well.

Categories that aligns with the project

Best LoRA (I trained a custom lora) Best Creative Vision(Made an interactive web app to use it) Best local use case (Provided clear instruction in model page to run the gradio app locally)

Accomplishments that we're proud of

It actually works! We created a LoRA for a completely novel creative task, not just another style transfer. We are most proud of building a fully functional, public demo that people can use and have fun with right now.

What we learned

This project was a deep dive into the creative potential of the FLUX.1 Kontext [dev] model. We learned that fine-tuning a custom LoRA can unlock entirely new abilities in a model.

What's next for Draw2Photo

I am going to make a AI image generator and photo editor SaaS. So, we can add very good features on the SaaS.

Built With

• flux-kontext-dev
• flux-kontext-lora
• gradio
• lora-training
• python