Search papers, labs, and topics across Lattice.
The paper introduces VaViM, an autoregressive video model, and VaVAM, a video-action model, to explore video generative modeling for autonomous driving. VaViM predicts future frames using spatio-temporal token sequences, capturing driving scene semantics and dynamics, while VaVAM uses VaViM's representations for imitation learning to generate driving trajectories. Experiments in open- and closed-loop driving scenarios demonstrate the potential of video pre-training, highlighting the importance of semantic richness, scaling, and the interplay between model size, data, and safety.
Video pre-training can drive autonomous vehicles, but scaling model size doesn't always guarantee safer closed-loop driving.
We explore the potential of large-scale generative video models for autonomous driving, introducing an open-source auto-regressive video model (VaViM) and its companion video-action model (VaVAM) to investigate how video pre-training transfers to real-world driving. VaViM is a simple auto-regressive video model that predicts frames using spatio-temporal token sequences. We show that it captures the semantics and dynamics of driving scenes. VaVAM, the video-action model, leverages the learned representations of VaViM to generate driving trajectories through imitation learning. Together, the models form a complete perception-to-action pipeline. We evaluate our models in open- and closed-loop driving scenarios, revealing that video-based pre-training holds promise for autonomous driving. Key insights include the semantic richness of the learned representations, the benefits of scaling for video synthesis, and the complex relationship between model size, data, and safety metrics in closed-loop evaluations. We release code and model weights at https://github.com/valeoai/VideoActionModel