tessera-small: a diffusion-LLM-based pixel-art text-to-image model

thejagstudio · hf · 2026-11-19

thejagstudio released tessera-small, a text-to-image model using masked/discrete diffusion (diffusion-LLM approach) trained on a 1.4M pixel-art dataset, licensed CC-BY-NC 4.0.

Original post →

More from Multimodal

Multimodal channel →