AWS bedrock documentation change
Summary
Updated documentation for model distillation prerequisites, including changes to wording ('choose' to 'decide on'), added details about input dataset preparation, JSONL format requirements, and new sections on IAM roles and security configurations.
Security assessment
The changes include new sections about setting up IAM roles and optional security configurations like encryption and VPC protection, which are security features. However, there is no evidence that these changes address a specific security vulnerability or incident.
Diff
diff --git a/bedrock/latest/userguide/prequisites-model-distillation.md index a5b3f1304..98447262d 100644 --- a/bedrock/latest/userguide/prequisites-model-distillation.md +++ b/bedrock/latest/userguide/prequisites-model-distillation.md @@ -9 +9 @@ Complete the following prerequisites before you start a model distillation job: - 1. **Choose a teacher model** + 1. **Decide on a teacher model** @@ -13 +13 @@ Choose a teacher model that is significantly larger and more capable than the st - 2. **Choose a student model** + 2. **Decide on a student model** @@ -18,0 +19,12 @@ Choose a student model that is significantly smaller in size. For information on +To prepare input datasets for your custom model, you create `.jsonl` files, each line of which is a JSON object corresponding to a record. The files you create must conform to the format for the customization method and model that you choose and the records in it must conform to size requirements. + +###### Note + +If you are using Anthropic or Meta Llama models, continue with this step. + +If you are using Amazon Nova models for distillation, see the following guidelines and then continue with step 4. + + * [Guidelines for preparing your data for Amazon Nova models](https://docs.aws.amazon.com/nova/latest/userguide/customize-fine-tune-prepare.html). + + * [Guidelines for model distillation for Amazon Nova](https://docs.aws.amazon.com/nova/latest/userguide/customize-distill.html#custom-distill-guidance). + @@ -26,0 +39,12 @@ Collect your prompts and store them in a JSON Line (JSONL) format. Each record i + * Include the `schemaVersion` field that must have the value `bedrock-conversion-2024`. + + * [Optional] Include a system prompt that indicates the role assigned to the model. + + * In `messages` field, include the user role containing the input prompt provided to the model. + + * [Optional] In the `messages` field, include assistant role containing the desired response. + +For the preview release Anthropic and Meta Llama models support only single -turn conversation prompts, meaning you can only have one user prompt. The Amazon Nova models support multi-turn conversations, allowing you to provide multiple user and assistant exchanges within one record. + +Example format + @@ -55,6 +78,0 @@ Collect your prompts and store them in a JSON Line (JSONL) format. Each record i -The structure must include the required field `schemaVersion` that must have the value `bedrock-conversion-2024`. You can optionally include a system prompt that indicates the role assigned to the model. In the messages field, the user role is required, containing the input prompt provided to the model, while the assistant role, containing the desired response, is optional. - -For the preview release Anthropic and Meta models support only single -turn conversation prompts, meaning you can only have one user prompt. The Amazon Nova models support multi-turn conversations, allowing you to provide multiple user and assistant exchanges within one record. - -After you have added all your prompts, upload the file to an Amazon Simple Storage Service (Amazon S3) bucket. - @@ -75 +93 @@ You can provide a maximum of 15K prompts or prompt-response pairs to Amazon Bedr -You can optionally add request metadata to the prompt-response pairs in the invocation log using one of the model invocation operations and then later use it to filter the logs. Amazon Bedrock can use the filtered logs to fine-tune the student model. +You can optionally [add request metadata](./add-metadata-to-prompts.html) to the prompt-response pairs in the invocation log using one of the model invocation operations and then later use it to filter the logs. Amazon Bedrock can use the filtered logs to fine-tune the student model. @@ -78,0 +97,8 @@ To filter the logs using multiple request metadata, use a single operation Boole + 4. If you do not already have an IAM service role with proper permissions, create a new custom AWS Identity and Access Management (IAM) [service role](https://docs.aws.amazon.com/IAM/latest/UserGuide/id_roles_terms-and-concepts.html#iam-term-service-role) with the proper permissions by following the instructions at [Create a service role for model customization](./model-customization-iam-role.html) to set up the role. You can skip this prerequisite if you plan to use the AWS Management Console to automatically create a service role for you. + + 5. (Optional) Set up extra security configurations. + + * You can encrypt input and output data, customization jobs, or inference requests made to custom models. For more information, see [Encryption of model customization jobs and artifacts](./encryption-custom-job.html). + + * You can create a virtual private cloud (VPC) to protect your customization jobs. For more information, see [[Optional] Protect your model customization jobs using a VPC](./vpc-model-customization.html). +