Tessera-small: a 47.6M-param masked diffusion LM that paints pixel art in ~16 parallel steps, trained in 5.5h on one GPU
RisingSayak · x · 2026-11-19
Tessera-small treats a 16x16 sprite as a 256-token sentence: each pixel is one token from a learned 512-color palette. This masked diffusion LM starts from a fully masked canvas and unmasks all 256 pixels in parallel in 16 steps, most confident first, conditioned on CLIP text — no VAE, no latents, no autoregression.
The model has just 47.6M parameters and was trained in 5.5 hours on a single GPU over 1.4M sprites. Caleb Fahlgren built a Hugging Face Space running an int8 ONNX build in-browser via WebGPU, letting you watch the canvas fill in live — a neat proof that discrete diffusion can at least do image generation.
More from Multimodal
- MiniMax H3 anime video generation runs locally on an RTX 5070 in ~17 minutes per clip — cyborggp · 2026-11-19
- Verba-TTS-Indic-Zero: a 7.6MB, 12-language TTS claiming 2000x cheaper than ElevenLabs — alexcovo_eth · 2026-11-19
- VEED launches AI video editing agent that trims, subtitles and styles footage from a text prompt — Med1_Ai · 2026-11-19
- Gemini 4 Neon makes a remarkably accurate T cell video, creator plans immunology series — DeryaTR_ · 2026-11-19
- HyperFlow: 14B distilled MiniMax H3 runs video gen on 6GB VRAM at 2x speed — cgpixel23 · 2026-11-19
- One Neon prompt turns a monstera plant into a walking, waving character — fofrAI · 2026-11-19