Skip to content
Open
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
3 changes: 3 additions & 0 deletions docs/source/reference/modules_models.rst
Original file line number Diff line number Diff line change
Expand Up @@ -14,8 +14,11 @@ Modules for model-based reinforcement learning, including world models and dynam
ObsEncoder
ObsDecoder
RSSMPosterior
RSSMPosteriorV3
RSSMPrior
RSSMPriorV3
RSSMRollout
RSSMRolloutV3

PILCO
-----
Expand Down
25 changes: 25 additions & 0 deletions docs/source/reference/objectives_other.rst
Original file line number Diff line number Diff line change
Expand Up @@ -16,3 +16,28 @@ Additional loss modules for specialized algorithms.
DreamerModelLoss
DreamerValueLoss
ExponentialQuadraticCost

DreamerV3
---------

Loss modules for DreamerV3 (`Mastering Diverse Domains in World Models, Hafner et al. 2023 <https://arxiv.org/abs/2301.04104>`_).
Key differences from V1: discrete categorical latent state, KL balancing, symlog transforms, and two-hot value distributions.

.. autosummary::
:toctree: generated/
:template: rl_template_noinherit.rst

DreamerV3ActorLoss
DreamerV3ModelLoss
DreamerV3ValueLoss

DreamerV3 Utilities
~~~~~~~~~~~~~~~~~~~

.. autosummary::
:toctree: generated/

symlog
symexp
two_hot_encode
two_hot_decode
Loading
Loading