"smithy.api#documentation": "<p> Current status of the deployment update. </p>",
"smithy.api#required": {}
}
}
},
"traits": {
"smithy.api#documentation": "<p> Details about an update to a custom model deployment, including the new custom model resource ARN and current update status. </p>"
"smithy.api#documentation": "<p>The Distillation configuration for the custom model.</p>"
}
},
"rftConfig": {
"target": "com.amazonaws.bedrock#RFTConfig",
"traits": {
"smithy.api#documentation": "<p> Configuration settings for reinforcement fine-tuning (RFT) model customization, including grader configuration and hyperparameters. </p>"
"smithy.api#documentation": "<p> Details about any pending or completed updates to the custom model deployment, including the new model ARN and update status. </p>"
"smithy.api#documentation": "<p> Configuration for using an AWS Lambda function as the grader for evaluating model responses and provide reward signals in reinforcement fine-tuning. </p>"
}
}
},
"traits": {
"smithy.api#documentation": "<p> Configuration for the grader used in reinforcement fine-tuning to evaluate model responses and provide reward signals. </p>"
}
},
"com.amazonaws.bedrock#GuardrailArn": {
"type": "string",
"traits": {
@@ -14785,6 +14874,31 @@
"smithy.api#documentation": "<p>The configuration details for returning the results from the knowledge base vector search.</p>"
"smithy.api#documentation": "<p> ARN of the AWS Lambda function that will evaluate model responses and return reward scores for RFT training. </p>",
"smithy.api#required": {}
}
}
},
"traits": {
"smithy.api#documentation": "<p> Configuration for using an AWS Lambda function to grade model responses during reinforcement fine-tuning training. </p>"
}
},
"com.amazonaws.bedrock#LegalTerm": {
"type": "structure",
"members": {
@@ -19000,6 +19114,124 @@
}
}
},
"com.amazonaws.bedrock#RFTBatchSize": {
"type": "integer",
"traits": {
"smithy.api#range": {
"min": 16,
"max": 512
}
}
},
"com.amazonaws.bedrock#RFTConfig": {
"type": "structure",
"members": {
"graderConfig": {
"target": "com.amazonaws.bedrock#GraderConfig",
"traits": {
"smithy.api#documentation": "<p> Configuration for the grader that evaluates model responses and provides reward signals during RFT training. </p>"
"smithy.api#documentation": "<p> Hyperparameters that control the reinforcement fine-tuning training process, including learning rate, batch size, and epoch count. </p>"
}
}
},
"traits": {
"smithy.api#documentation": "<p> Configuration settings for reinforcement fine-tuning (RFT), including grader configuration and training hyperparameters. </p>"
}
},
"com.amazonaws.bedrock#RFTEvalInterval": {
"type": "integer",
"traits": {
"smithy.api#range": {
"min": 1,
"max": 100
}
}
},
"com.amazonaws.bedrock#RFTHyperParameters": {
"type": "structure",
"members": {
"epochCount": {
"target": "com.amazonaws.bedrock#EpochCount",
"traits": {
"smithy.api#documentation": "<p> Number of training epochs to run during reinforcement fine-tuning. Higher values may improve performance but increase training time. </p>"
}
},
"batchSize": {
"target": "com.amazonaws.bedrock#RFTBatchSize",
"traits": {
"smithy.api#documentation": "<p> Number of training samples processed in each batch during reinforcement fine-tuning (RFT) training. Larger batches may improve training stability. </p>"
"smithy.api#documentation": "<p> Maximum length of input prompts during RFT training, measured in tokens. Longer prompts allow more context but increase memory usage and training-time. </p>"
"smithy.api#documentation": "<p> Number of response samples generated per prompt during RFT training. More samples provide better reward signal estimation. </p>"
"smithy.api#documentation": "<p> Level of reasoning effort applied during RFT training. Higher values may improve response quality but increase training time. </p>"
"smithy.api#documentation": "<p> Interval between evaluation runs during RFT training, measured in training steps. More frequent evaluation provides better monitoring. </p>"
}
}
},
"traits": {
"smithy.api#documentation": "<p> Hyperparameters for controlling the reinforcement fine-tuning training process, including learning settings and evaluation intervals. </p>"
"smithy.api#documentation": "<p> Updates a custom model deployment with a new custom model. This allows you to deploy updated models without creating new deployment endpoints. </p>",
