RL Config#

Resource Config#

rl.utils.AcceleratorResourcesConfig

alias of AcceleratorResourcesConfig

rl.utils.CPUResourcesConfig

alias of CPUResourcesConfig

Rollout Config#

Agent Loop Config#

rl.agent_loop.SingleTurnAgentLoopConfig

alias of SingleTurnAgentLoopConfig

rl.agent_loop_manager.AgentLoopManagerConfig

alias of AgentLoopManagerConfig

rl.agent_loop_manager.TaskSpecConfig

alias of TaskSpecConfig

rl.agent_loop_manager.SamplerConfig

alias of SamplerConfig

rl.agent_loop_manager.SyncProduceStrategyConfig

alias of SyncProduceStrategyConfig

rl.agent_loop_manager.AsyncProduceStrategyConfig

alias of AsyncProduceStrategyConfig

Judger Config#

rl.judger.JudgerConfig

alias of JudgerConfig

rl.judger.GSM8KJudgerConfig

alias of GSM8KJudgerConfig

rl.judger.ComposedJudgerConfig

alias of ComposedJudgerConfig

Replay and Evaluation Config#

Training and Loss Config#

rl.trainer.WorkerConfig

alias of WorkerConfig

rl.loss.BaseRLLossConfig

alias of BaseRLLossConfig

rl.loss.GRPOLossConfig

alias of GRPOLossConfig

rl.loss.OrealLossConfig

alias of OrealLossConfig

rl.rollout_is.RolloutImportanceSampling

alias of RolloutImportanceSampling