DS1

Appearance

Frequently Asked Questions ​

Models ​

What models are available in DS1? ​

Currently, DS1 offers one production-ready model:

  • DS1-EN-V1: A high-performance English text embedding model with 512 dimensions (L2 normalized), 512 token context length, and a 30k WordPiece tokenizer

Two additional models are scheduled for release:

  • DS1-Multilingual-V1: Optimized for text embedding across multiple languages, providing consistent performance for global applications
  • DS1-Code-V1: Specialized for embedding source code, understanding programming syntax, semantics, and code relationships

When will the multi-lingual and code models be available? ​

We're actively developing these models and will announce release dates as they approach production readiness. Contact us to be notified of updates or to discuss early access opportunities.

Is there an API available? ​

No, DS1 is currently only delivered as a SageMaker Model through AWS Marketplace which you can deploy within your own AWS account.

Getting Started ​

How do I start using DS1? ​

  1. Check out our Quick Start Guide
  2. Review the Introduction for an overview
  3. Try our SageMaker Notebook for hands-on experience

How do I access DS1 after deployment? ​

Once deployed as a SageMaker model in your AWS account, you can invoke it through SageMaker's endpoint interface. See the OpenAPI specification for the complete API reference, or check API Examples for implementation patterns.

Usage & Integration ​

What programming languages are supported? ​

DS1 is accessible from any language supporting AWS SDKs or HTTP requests. See the OpenAPI specification for the complete API reference.

What are common use cases for DS1 embeddings? ​

  • Semantic Search: Find contextually relevant content
  • Retrieval-Augmented Generation (RAG): See our Quickstart Guide
  • Document Similarity: Compare and cluster documents
  • Recommendation Systems: Suggest related content

Are there rate limits? ​

Since DS1 is deployed as a SageMaker model in your own AWS account, the only limits that apply are your AWS account limits and the instance configuration you choose. DS1 itself does not enforce any rate limits.

Deployment ​

Can I deploy DS1 on AWS? ​

Yes! Check our AWS Deployment Guide for detailed instructions on setting up DS1 in your AWS environment.

Is DS1 available in other cloud platforms? ​

Currently, DS1 is exclusively available through AWS Marketplace as a SageMaker model. This ensures optimal integration with AWS services and provides the best performance and security within the AWS ecosystem. For questions about future platform availability, please contact us.

Troubleshooting ​

What should I do if I encounter errors with my deployed model? ​

  1. Check your SageMaker endpoint status and configuration
  2. Verify your request format matches the OpenAPI specification
  3. Review CloudWatch logs for your SageMaker endpoint for detailed error information
  4. Ensure your AWS credentials and permissions are properly configured
  5. If issues persist, contact our support team

How do I optimize embedding performance? ​

  • Use appropriate SageMaker instance types for your workload
  • Batch multiple texts in a single inference request when possible
  • Configure auto-scaling for variable workloads
  • Use appropriate text preprocessing for your use case
  • Consider caching embeddings for frequently used content
  • Monitor CloudWatch metrics to optimize instance configuration

Support ​

How can I get help? ​

  • Review this FAQ and our documentation
  • Consult the OpenAPI specification for API details
  • Check AWS SageMaker documentation for platform-specific questions
  • Contact us for DS1-specific support

Do you offer technical consulting? ​

Yes, we provide technical consulting services to help optimize your embedding implementation. Contact us to discuss your specific needs.