RL
Attributes
attributenamestr= 'rl'attributeConfigtype= RLConfigFunctions
func_check_inputs(self, ctx) -> NoneparamselfparamctxRunContextReturns
Nonefuncsetup(self, ctx, backend) -> NoneparamselfparamctxRunContextparambackendTinkerBackendReturns
Nonefuncbuild_batch(self, step_idx) -> TrainingBatchparamselfparamstep_idxintReturns
evsys_sdk.training.loop.TrainingBatchfunc_hyperparams_extra(self) -> dict[str, Any]paramselfReturns
dict[str, typing.Any]func_prep_task(self, t) -> HarborTaskTemplate the instruction + fill the verifier fn_name fallback, so the materialized harbor task is self-contained.
paramselfparamtHarborTaskReturns
evsys_sdk.data_types.HarborTaskfunc_slice(self, step_idx) -> list[HarborTask]paramselfparamstep_idxintReturns
list[evsys_sdk.data_types.HarborTask]