Tiny Moss-image model generates a coherent 40-person crowd photo
Crazy-Repeat-2006 · reddit · 2026-10-20
A Reddit post highlights a viral demo of the small mossimagebaseint8 diffusion model: with a structured JSON prompt and two SFT LoRAs (including an 8-step DMD), it generated a natural group photo of 40 ethnically diverse women — faces, hair, skin tones and body types all distinct and mostly artifact-free, a classic failure mode for diffusion models. The full layered prompt (photo/group/arrangement) for controlling asymmetric crowd layout, natural skin texture and hidden text easter eggs is included, though minor imperfections remain.
More from Multimodal
- Polya AI launches Dialog-RSN-1: realtime voice model with sub-300ms median first-token latency — matthen2 · 2026-10-20
- CueForge v3.4.0 brings ComfyUI 0.42.1 parity to mobile with 3D previews and low-resource mode — galactic_lobster · 2026-10-20
- ASCII Magic launches 634-piece MIT-licensed animated ASCII art library you can wire into coding agents — sull · 2026-10-20
- AI-generated virtual influencer 'Zanita Krakleïn' stars in hyperreal Miami night clip — ovninoir · 2026-10-20
- Training MOSS-Image LoRAs locally on 8GB VRAM: 500 steps beats 8000, full settings inside — AcademiaSD · 2026-10-20
- AI-generated "Verdantism" style: lush hedge mazes and surreal retro illustration — OVolosin82152 · 2026-10-20