Two repositories were released. next-state/open-dreamer holds the training pipeline: a causal video tokenizer, an action-conditioned latent dynamics model, rollout ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results