world_loss#
- hybrid_aif.train.world_loss(
- core,
- batch,
- key,
- *,
- dyn_scale=0.5,
- rep_scale=0.1,
- free_nats=1.0,
- decoder_scale=1.0,
- reward_scale=1.0,
- cont_scale=1.0,
- return_states=False,
The world-model loss on one batch:
(total, WorldLosses).With
return_states=Truethe auxiliary is(WorldLosses, RSSMState), the posterior states of this forward pass, leaves(B, T, ...). Pure and differentiable incore;keydraws the straight-through posterior samples. Defaults are DreamerV3’s loss scales and one-nat free-bits floor.