Skip to content

ReinforcementHyperparameters ​

Example Usage ​

typescript
import { ReinforcementHyperparameters } from "@meetkai/mka1/models/components";

let value: ReinforcementHyperparameters = {};

Fields ​

FieldTypeRequiredDescription
batchSizecomponents.ReinforcementHyperparametersBatchSize➖Number of examples in each batch
computeMultipliercomponents.ComputeMultiplier➖Scales compute for exploration during training
evalIntervalcomponents.EvalInterval➖Number of steps between evaluations
evalSamplescomponents.EvalSamples➖Number of samples generated per eval step
learningRateMultipliercomponents.ReinforcementHyperparametersLearningRateMultiplier➖Scaling factor for learning rate
nEpochscomponents.ReinforcementHyperparametersNEpochs➖Number of epochs to train for
reasoningEffortcomponents.ReinforcementHyperparametersReasoningEffort➖Reasoning effort level