lens is always the episode length, including the bootstrap value timestep, and use is_truncated to mask out the bootstrap node from being learned on
Dreamer 4 (wip)
Implementation of Danijar's latest iteration for his Dreamer line of work
Citation
@misc{hafner2025trainingagentsinsidescalable,
title = {Training Agents Inside of Scalable World Models},
author = {Danijar Hafner and Wilson Yan and Timothy Lillicrap},
year = {2025},
eprint = {2509.24527},
archivePrefix = {arXiv},
primaryClass = {cs.AI},
url = {https://arxiv.org/abs/2509.24527},
}
Description
Implementation of Danijar's latest iteration for his Dreamer line of work
artificial-intelligenceattentiondeep-learningmodel-based-reinforcement-learningtransformersworld-models
Readme
MIT
842 KiB
Languages
Python
100%
