Rindow RL Agents is a collection of reinforcement-learning agents, training runners, replay storage, and device-aware environment adapters for Rindow Neural Networks. The implementation supports both host and accelerator linear algebra backends.
Public classes use the Rindow\RL\Agents namespace. Agent implementations are
under Rindow\RL\Agents\Agent, buffers under Rindow\RL\Agents\ReplayBuffer,
and device wrappers under Rindow\RL\Agents\Env.
This manual documents the source tree as it exists in this repository. It is an API reference, not a guarantee that one hyperparameter set will converge for every environment or random seed.