Skip to main content

compute_mc_returns

Function compute_mc_returns 

Source
pub fn compute_mc_returns(buffer: &mut RolloutBuffer)
Expand description

Compute Monte Carlo returns (full episode returns)

Computes returns using the full trajectory, which provides unbiased but high-variance estimates. Only works for complete episodes.

ยงArguments

  • buffer - Rollout buffer to compute returns for