We start from an initial video context and unroll the scene for two videos. We see that in both cases, the model predicts a different future. The two different futures contain different traffic and shows the model is able to generate a diverse set of scenes. Read the blog here: https://wayve.ai/thinking/scaling-gaia-1/

(video) Holger Caesar – Autolabeling Everything Everywhere | Nuro Technical Talks

GIPHY App Key not set. Please check settings