Around the World in 80 Timesteps
Generate 3D camera trajectories based on text prompts
Retrieve videos of human motions based on text input