ironjr/semantic-draw
A real-time interactive canvas that uses diffusion models to generate images from multiple text prompts drawn on different regions simultaneously.
Velocity · 7d
+0.7
★ / day
Trend
→steady
star history
SemanticDraw provides a large drawing canvas where users can paint masks and assign distinct text prompts to different regions, generating corresponding sections of an image in real time. The system builds on Stable Diffusion and StreamMultiDiffusion to achieve interactive generation speeds. It includes multiple Hugging Face Spaces demos supporting SD1.5, SDXL, and SD3 for users to try without installation.