A2C
is implemented as rlpack.actor_critic.a2c.A2C
. It inherits from Agent
class defined as rlpack.utils.base.agent.Agent
.
A2C implements the synchronous Actor-Critic agent which supports gradient accumulation and mean reduction of gradients for training the model. Unlike A3C, only a single agent interacts with environment and mean reduced gradients are synchronously used to train the policy model. For more information refer here
Keyword: agent_name: "a2c"