The `ml_full_access` role plays a crucial part in the configuration and integration of DeepSeek with OpenSearch Service. This role is part of the OpenSearch Security plugin and is used to manage access to machine learning (ML) features within OpenSearch.
Role Description
The `ml_full_access` role grants full access to all ML features within OpenSearch. This includes the ability to start new ML tasks, read models, and delete models. It provides comprehensive control over ML functionalities, making it essential for users who need to manage and execute ML tasks within OpenSearch[2].
Integration with DeepSeek
When integrating DeepSeek with OpenSearch, the `ml_full_access` role is used to ensure that authenticated entities can execute API calls related to ML tasks. This includes creating connectors to external models like those provided by DeepSeek. By mapping IAM roles to the `ml_full_access` role, users can ensure that their AWS identities have the necessary permissions to interact with DeepSeek models through OpenSearch[1][3].
Configuration Steps
To configure the `ml_full_access` role for use with DeepSeek, you typically follow these steps:
1. Create IAM Roles: Generate IAM roles that allow OpenSearch to interact with SageMaker endpoints and create connectors. These roles are essential for setting up the necessary permissions and trust relationships[1].
2. Map Backend Roles: In OpenSearch Dashboards, navigate to the `ml_full_access` role and map the IAM role ARNs created earlier to backend roles. This step ensures that users with these IAM roles have full access to ML features in OpenSearch[3][6].
3. Verify Setup: After configuring roles and permissions, verify that the setup is working by running test queries through OpenSearch Dashboards. This ensures that you can successfully connect to DeepSeek models and execute ML tasks[1].
In summary, the `ml_full_access` role is vital for managing ML features in OpenSearch and ensuring seamless integration with DeepSeek models by providing the necessary permissions for executing ML tasks and accessing external models.
Citations:
[1] https://aws.amazon.com/blogs/big-data/use-deepseek-with-amazon-opensearch-service-vector-database-and-amazon-sagemaker/
[2] https://opensearch.org/docs/latest/ml-commons-plugin/model-access-control/
[3] https://github.com/opensearch-project/ml-commons/blob/main/docs/tutorials/aws/RAG_with_DeepSeek_R1_model_on_Sagemaker.md
[4] https://github.com/opensearch-project/ml-commons/blob/main/docs/tutorials/aws/RAG_with_DeepSeek_R1_model_on_Bedrock.md
[5] https://github.com/opensearch-project/ml-commons/blob/main/docs/tutorials/aws/DeepSeek_demo_notebook_for_RAG.ipynb
[6] https://opensearch.org/docs/latest/vector-search/tutorials/rag/rag-deepseek-chat/
[7] https://github.com/opensearch-project/ml-commons/blob/main/docs/tutorials/aws/RAG_with_DeepSeek_Chat_model.md
[8] https://zilliz.com/ai-faq/what-apis-does-deepseek-provide-for-model-access