Scaling GAIA-1: 9-billion parameter generative world model f

Discover how we optimised GAIA-1, enhancing this generative world model with larger-scale training that resulted in higher resolution video generation with fine-grained control over ego-vehicle behaviour and scene features.

Related Keywords

United Kingdom , London , City Of , , Youtube , Language Models , Controllable High Quality Neural , Diffusion Modeling , High Definition Video Generation , Masked Generative Video , High Resolution Video Synthesis , Latent Diffusion ,