MOSS SFT image model tested with bounding-box workflow, mixed results
Perfect-Campaign9551 · reddit · 2026-10-18
A Reddit user tested the MOSS SFT image model with a bounding-box workflow; results look mostly fine, though ankle details falter on the 'girl on grass' prompt.
More from Multimodal
- How to Make Professional Food Videos with AI — aziz4ai · 2026-10-19
- One character sheet to a 44s AI-made promo: full workflow with Claude, ElevenLabs and HeyGen — azed_ai · 2026-10-19
- AI-generated "dolphin girlfriend" couch video is peak surreal AI comedy — TradingCardGirl · 2026-10-19
- Surreal AI video "The answer" turns an empty stage into viral ambiance — Electronic_Still_274 · 2026-10-19
- AI-generated family drama skit "I'm leaving the house" is absurdly funny — TradingCardGirl · 2026-10-19
- AI video "Tiger goes to Jail": an orange cat in a #999 prison uniform threatens its cellmate — TradingCardGirl · 2026-10-19