A complete implementation of Dreamer v4, including the causal masked-autoencoder tokenizer, block-causal transformer dynamics model, shortcut-forcing objective, task-token finetuning, and imagination-based reinforcement learning.
🎥 Generate high-quality interactive videos in real-time using Causal Forcing, a method that elevates autoregressive diffusion distillation.
Add a description, image, and links to the dreamer-v4 topic page so that developers can more easily learn about it.
To associate your repository with the dreamer-v4 topic, visit your repo's landing page and select "manage topics."