Valeo Derives Video Diffusion Scaling Laws from 5,500 Hours of Driving Data

Valeo AI's VATIX project trains a 9B-parameter video diffusion model on about 5,500 hours of anonymized driving footage from 28 countries, deriving scaling laws for video generation by optimizing model size and training exposure.

2026-09-01 ~ 2026-09-01 · 3 related posts

• Study analyzes scaling laws of video diffusion models using 5,500 hours of driving data — kwangmoo_yi · 2026-09-01
• 5,500 hours of driving video, a 9B diffusion model: Valeo derives scaling laws for video generation — kwangmoo_yi · 2026-09-01
• Valeo Study: Scaling Laws for Video Diffusion on 5,500 Hours of Driving Data — andrew_n_carr · 2026-09-01