Abubacarr/beam-long-video-world-model

Dataset Analysis

🌌 Beam Long-Video World Model Dataset This dataset was created for the Reflection AI + Hugging Face "Build with Beam" Challenge. It pairs compressed 3D Causal VAE video latents with Correlated Continuous Noise vectors and Hierarchical 1M Long-Context Captions & Predictive World-State Vectors generated by Beam 1.0 (501B-A23B). Designed specifically for next-generation Video World Models, Autoregressive Latent Transformers, and Diffusion Video Generation research.… See the full description on the dataset page: https://huggingface.co/datasets/Abubacarr/beam-long-video-world-model.