Study finds text-to-image diffusion models cannot control clock times in generated images
A new arXiv paper reports that text-to-image diffusion models can render photorealistic analog clocks but cannot control the time displayed, a limitation observed across four pretrained models.
Read the full story at arXiv cs.CV ↗
Timeline · 1 report
- 2026-11-11 05:00 · arXiv cs.CV
Continuous Angular Control in Text-to-Image Generation via Geometric Token Encoding