Deploying generative world models for end-to-end closed-loop evaluation and policy learning requires precise scenario control, consistent observations, and sustained generation quality during extended interactions. This report presents \emph{X-World~1.6}, advancing X-World~1.0 toward a more reliable and versatile drivi...
Video world models generate future visual observations from camera or action inputs, with the goal of synthesizing persistent 3D worlds that can be explored across time and space. Existing benchmarks typically evaluate short interactions or rely on camera trajectories fixed before generation. The performance of video w...
Min-Qi Wang, Jie Peng, He-Jun Dong et al.· 1 citation
To stream long videos while maintaining visual quality and temporal consistency, Self Forcing mitigates exposure bias through self-rollout training on self-generated histories with key-value (KV) caching. To keep memory manageable, it detaches historical caches, preserving forward dependencies between chunks but severi...
Dong-Bin Zhang, Chao-Da Zheng, Kang-Jie Chen et al.· 0 citations
This work presents VGGT-Diff, a geometry-routed multi-view diffusion model for sparse-view novel view synthesis, and introduces robust geometry conditioning, combining training-time regularization with inference-time guidance for improved robustness.