Collected sources and patterns will appear here. Add from search, explore, or the patterns library.
VectorizedActions -> StepResult<VectorizedObservations, Rewards, Dones, Info>
Automatically reset individual sub-environments within a vectorized environment runner upon termination, storing the final observation in the step information dictionary.
Problem it solves
Manual reset tracking in parallel environment steps complicates training loop logic.
Consumes
Emits
The real projects this mechanism was found in. Attribution is the point — this is how the best teams actually do it.