unilab.tasks.locomotion.g1.manager_terms.G1PenaltyCurriculum

class unilab.tasks.locomotion.g1.manager_terms.G1PenaltyCurriculum[source]

Bases: ManagerTermBase

Scale negative-weight reward terms by average episode length.

Ports the G1Walk task-family penalty curriculum (64c7f8b8, roadmap #1042): penalty weights start at initial_scale of their configured value and relax toward max_scale as the tracked average episode length crosses the configured thresholds.

Parameters:

  • cfg (ManagerTermBaseCfg)

  • env (_G1Env)

Methods

__init__(cfg, env)

reset(env_ids)

Row boundary for the global curriculum state.

reset_scoped(env, env_ids, **params)

Selected-row curriculum update at reset.

Attributes

name

num_envs

__init__(cfg, env)[source]
Parameters:

  • cfg (ManagerTermBaseCfg)

  • env (_G1Env)

__call__(env, env_ids, **params)[source]

Returns the value of the term required by the manager.

Parameters:
Return type:

dict[str, float]

reset(env_ids)[source]

Row boundary for the global curriculum state.

Parameters:

env_ids (Tensor)

Return type:

None

reset_scoped(env, env_ids, **params)[source]

Selected-row curriculum update at reset.

Parameters:
Return type:

dict[str, float]

property name: str
property num_envs: int