rlightning.policy.rsl_rl_policy.ppo_vtrace

PPO policy variant using V-trace value estimation.

Overrides rsl_rl’s PPO to compute value targets and advantages with V-trace corrections.

class rlightning.policy.rsl_rl_policy.ppo_vtrace.Benchmarkable

Bases: object

static register(func)

class rlightning.policy.rsl_rl_policy.ppo_vtrace.PPOVtrace(env, rho_bar: float = 1.0, c_bar: float = 1.0, **kwargs)

Bases: object