LLaVA-OneVision-2.5 released: first clear signal that codec input beats frame sampling
liuziwei7 · x · 2026-11-18
LMMS Lab released LLaVA-OneVision-2.5, presenting the first clear evidence that codec-based video input outperforms uniform frame sampling on general video understanding tasks.
- In OV-2, training with both approaches showed codec worked but mainly helped temporal grounding, not all video tasks.
- The team argues video shouldn't be treated as a pile of independent images — even single images could be stored in compressed views like JPEG, making codec a more unified path.
- OV-2.5 experiments show codec input comes out ahead overall on general video tasks.
More from Multimodal
- LMMS-Lab: codec sampling beats uniform frame sampling on general video tasks — liuziwei7 · 2026-11-18
- AI-generated Robert Pattinson photo set is nearly indistinguishable from real photography — justin_hart · 2026-11-18
- From 1000-step DDPM to 1-step generation: a hands-on turbo diffusion walkthrough — RisingSayak · 2026-11-18
- Clueso v2: AI video editor lets teammates pin comments to the second and the AI makes the edit — aakashgupta · 2026-11-18
- Users seek a simple Director node for Minimax H3 workflows — Kind_Owl2245 · 2026-11-18
- Adobe Firefly ambassador shares full workflow for embroidery-style AI animated short — azed_ai · 2026-11-18