Dev cracks image conditioning on a DIY video model trained on one hour of footage
pixlpa · x · 2026-09-15
Developer pixlpa cracked proper image conditioning of the motion model in a DIY video generation setup and says the results blow him away. Both the motion and image models are fairly simple diffusion models trained on roughly an hour of footage — and the author can hardly believe it works at all.
More from Multimodal
- Artist Gets Old GANs Training Locally on a MacBook After Solving Apple Silicon Compatibility in an Afternoon — Merzmensch · 2026-09-15
- Oyama AI Video Studio tested: MiniMax and LTX 2.5 under one UI — Wild_Ant5693 · 2026-09-15
- "Nothing Was Filmed": a full AI commercial from idea to edit — memo_dreamtide · 2026-09-15
- AI Band Duchess Of The Damned Releases Shock Rock Song 'neewollaH' — LabImpressive1724 · 2026-09-15
- Generating and Animating Fantasy Hanfu Videos with MiniMax H3 — Devajyoti1231 · 2026-09-15
- Seedance 2.5 keeps hiding the cat's fate no matter how you prompt it — techhalla · 2026-09-15