All Classes and Interfaces
Class
Description
Configuration for the AutoEncoder Compressor model.
A passive utility model for dimension reduction via learned compression.
A VariableContextElement that wraps a compressor model internally.
Configuration for the Curiosity-Driven Recurrent Soft Actor-Critic model.
An exploration-driven variant of RSAC that learns through intrinsic curiosity.
Configuration for the Double Deep Q-Network (DDQN) model.
A classic deep reinforcement learning agent for discrete action spaces.
Descriptor for a nested AutoEncoder slot within an RSAC agent.
Configuration for the Proximal Policy Optimization (PPO) model.
PPO Agent interface for continuous action spaces using clipped surrogate objectives.
Configuration for the Recurrent Soft Actor-Critic (RSAC) model.
Builder for
RecurrentSoftActorCriticConfig.A sophisticated deep reinforcement learning agent for continuous control with temporal memory.
Represents a single atomic capability of an Agent.
Manages the registry of actions available to an agent.
Represents an active "Decision Maker" agent.
The main entry point for the Spartan Deep Learning engine.
Represents a passive "Data Processor" model.
Service Provider Interface (SPI) for creating concrete configuration implementations.
Registry to hold the concrete implementation of configuration creation.
Represents the agent's perception of the world.
Base contract for a single context element (sensor) that contributes values to a SpartanContext.
Base class for fixed-size hot-encoded elements.
Represents an instance of a Machine Learning brain.
Base configuration interface for all Spartan ML models.
Discriminator identifying which model family a config belongs to.
Configuration for a Multi-Agent group container.
Group of N homogeneous agents sharing a single SpartanContext.
Exception thrown when a native Spartan engine operation fails.
Exception thrown when model persistence operations fail.
Convenience base class for a single scalar context element.
Base class for variable-length context elements.
Represents a fixed block of memory allocated to a context element.