MOSS-Image-Base nails small real-world sign text: lost cat flyers and passive-aggressive notes
Impressive-Kick293 · reddit · 2026-11-19
A redditor tested MOSS-Image-Base on small real-world text scenes: lost cat flyers, windshield notes, garage sale signs, cafe chalkboards. No layout JSON needed—just quote each line in the prompt; apostrophes, colons, commas and dollar signs all render correctly. Setup: SLG 22/24, ER-SDE snoise 1.1, cfg 3, 25 steps, 2MP.
More from Multimodal
- HiT image generator learns its own abstraction hierarchy, hits 1.48 FID on ImageNet 256 — LucaAmb · 2026-11-19
- Osmo rebrands as applied AI lab, launches Posty video agent that cuts footage in under 30 min — yawnxyz · 2026-11-19
- Dev ports MiniMax H3 to Swift, cuts 480p video gen to ~4 min on a 16GB M1 MacBook — Lokoto123 · 2026-11-19
- Dreamina Black Friday: unlimited Seedance 2.5 video for 30 days, credits up to 83% off — FellMentKE · 2026-11-19
- How I built a music demo with Gemini 4 and Antigravity — alexanderchen · 2026-11-19
- fal partners with Speri Labs to post-train Idol model, inviting community builds — OdinLovis · 2026-11-19