Dev's posttrained ref2vid model turns 5 reference images into 15s video in ~6 seconds
Kyrannio · x · 2026-10-30
Developer Kyrannio showed off a self-posttrained ref2vid model: given 5 reference images, it generates a coherent 15-second clip in about 6 seconds, with speed as the main optimization target.
Known issues remain—random objects occasionally appear in the output—and he says quality improvements are coming once he's back in shape.
Related event: Self-Trained ref2vid Model Generates 15-Second Video in Seconds(2 posts)→
More from Multimodal
- Image-understanding model Jev 1.14 lands on Runware at $0.042 per 1M input tokens — aziz4ai · 2026-10-31
- ComfyUI workflow turns MiniMax H3 into a full comic-page generator, open-sourced — gatortux · 2026-10-31
- Runway Agent gains new skills: opens files, runs code, builds brand launch kits — runwayml · 2026-10-31
- FastVideo open-sources FastH3: 4-step distilled MiniMax H3 with 14x Blackwell speedup — victormustar · 2026-10-31
- AI-generated series 'Árpád' hits episode 6 with strikingly lifelike results — PannoniumFilms · 2026-10-31
- Deconstruct AI reverse-engineers reference images into MJ/SDXL/ComfyUI prompts — robert-988 · 2026-10-31