Today, we are delighted to reveal that DeepSeek R1 distilled Llama and Qwen models are available through Amazon Bedrock Marketplace and Amazon SageMaker JumpStart. With this launch, you can now deploy DeepSeek AI’s first-generation frontier design, DeepSeek-R1, together with the distilled variations ranging from 1.5 to 70 billion criteria to develop, experiment, and responsibly scale your generative AI concepts on AWS.

In this post, we show how to get started with DeepSeek-R1 on Amazon Bedrock Marketplace and SageMaker JumpStart. You can follow comparable actions to release the distilled versions of the models as well.

Overview of DeepSeek-R1

DeepSeek-R1 is a big language model (LLM) established by DeepSeek AI that utilizes reinforcement finding out to enhance thinking capabilities through a multi-stage training procedure from a DeepSeek-V3-Base foundation. A crucial differentiating feature is its support knowing (RL) step, which was used to fine-tune the model’s actions beyond the standard pre-training and fine-tuning procedure. By incorporating RL, DeepSeek-R1 can adapt more successfully to user feedback and goals, eventually improving both importance and clarity. In addition, DeepSeek-R1 uses a chain-of-thought (CoT) method, implying it’s geared up to break down complex queries and factor through them in a detailed way. This directed reasoning process enables the model to produce more precise, transparent, and detailed responses. This design integrates RL-based fine-tuning with CoT capabilities, aiming to produce structured responses while concentrating on interpretability and user interaction. With its extensive abilities DeepSeek-R1 has actually caught the industry’s attention as a flexible text-generation model that can be integrated into numerous workflows such as representatives, rational reasoning and information interpretation jobs.

DeepSeek-R1 utilizes a Mix of Experts (MoE) architecture and is 671 billion criteria in size. The MoE architecture allows activation of 37 billion specifications, enabling efficient inference by routing queries to the most relevant specialist “clusters.” This approach enables the model to concentrate on various issue domains while maintaining total effectiveness. DeepSeek-R1 requires at least 800 GB of HBM memory in FP8 format for reasoning. In this post, we will use an ml.p5e.48 xlarge instance to release the model. ml.p5e.48 xlarge includes 8 Nvidia H200 GPUs supplying 1128 GB of GPU memory.

DeepSeek-R1 distilled models bring the reasoning abilities of the main R1 design to more efficient architectures based on popular open models like Qwen (1.5 B, 7B, 14B, and 32B) and Llama (8B and 70B). Distillation describes a procedure of training smaller sized, more efficient models to mimic the habits and reasoning patterns of the larger DeepSeek-R1 design, using it as an instructor design.

You can deploy DeepSeek-R1 design either through SageMaker JumpStart or Bedrock Marketplace. Because DeepSeek-R1 is an emerging model, we recommend deploying this model with guardrails in location. In this blog, we will use Amazon Bedrock Guardrails to introduce safeguards, avoid harmful material, and evaluate models against essential safety criteria. At the time of composing this blog, for DeepSeek-R1 releases on SageMaker JumpStart and Bedrock Marketplace, Bedrock Guardrails supports just the ApplyGuardrail API. You can develop several guardrails tailored to different usage cases and apply them to the DeepSeek-R1 model, improving user experiences and standardizing security controls throughout your generative AI applications.

Prerequisites

To deploy the DeepSeek-R1 model, you require access to an ml.p5e circumstances. To examine if you have quotas for P5e, open the Service Quotas console and under AWS Services, choose Amazon SageMaker, and verify you’re utilizing ml.p5e.48 xlarge for endpoint use. Make certain that you have at least one ml.P5e.48 xlarge circumstances in the AWS Region you are deploying. To ask for a limit boost, create a limitation increase demand and connect to your account group.

Because you will be releasing this model with Amazon Bedrock Guardrails, make certain you have the right AWS Identity and Gain Access To Management (IAM) consents to utilize Amazon Bedrock Guardrails. For instructions, see Set up authorizations to use guardrails for content filtering.

Implementing guardrails with the ApplyGuardrail API

Amazon Bedrock Guardrails permits you to introduce safeguards, prevent hazardous material, and evaluate designs against crucial security criteria. You can execute security measures for the DeepSeek-R1 model utilizing the Amazon Bedrock ApplyGuardrail API. This allows you to apply guardrails to examine user inputs and design reactions deployed on Amazon Bedrock Marketplace and SageMaker JumpStart. You can produce a guardrail using the Amazon Bedrock console or the API. For the example code to develop the guardrail, see the GitHub repo.

The basic flow includes the following steps: First, the system receives an input for the design. This input is then processed through the ApplyGuardrail API. If the input passes the guardrail check, it’s sent to the model for reasoning. After getting the model’s output, another guardrail check is used. If the output passes this last check, it’s returned as the last outcome. However, if either the input or output is stepped in by the guardrail, a message is returned indicating the nature of the intervention and whether it happened at the input or output stage. The examples showcased in the following areas show inference using this API.

Deploy DeepSeek-R1 in Amazon Bedrock Marketplace

Amazon Bedrock Marketplace gives you access to over 100 popular, emerging, and specialized foundation models (FMs) through Amazon Bedrock. To gain access to DeepSeek-R1 in Amazon Bedrock, complete the following actions:

1. On the Amazon Bedrock console, select Model catalog under Foundation designs in the navigation pane. At the time of composing this post, you can utilize the InvokeModel API to invoke the model. It doesn’t support Converse APIs and other Amazon Bedrock tooling.

1. Filter for DeepSeek as a supplier and select the DeepSeek-R1 model.
   
   The model detail page offers essential details about the design’s capabilities, pricing structure, and execution guidelines. You can find detailed use guidelines, including sample API calls and code snippets for combination. The design supports various text generation tasks, including material production, code generation, and concern answering, utilizing its support learning optimization and CoT thinking capabilities. The page also consists of release options and licensing details to help you begin with DeepSeek-R1 in your applications.
2. To start using DeepSeek-R1, choose Deploy.
   
   You will be triggered to configure the release details for DeepSeek-R1. The design ID will be pre-populated.
3. For Endpoint name, enter an endpoint name (between 1-50 alphanumeric characters).
4. For Number of circumstances, go into a number of circumstances (in between 1-100).
5. For example type, select your circumstances type. For optimal efficiency with DeepSeek-R1, a GPU-based circumstances type like ml.p5e.48 xlarge is advised. Optionally, you can set up sophisticated security and infrastructure settings, including virtual private cloud (VPC) networking, service function permissions, and file encryption settings. For most utilize cases, the default settings will work well. However, for production implementations, you might desire to examine these settings to align with your company’s security and compliance requirements.
6. Choose Deploy to begin using the model.
   
   When the implementation is complete, you can check DeepSeek-R1’s capabilities straight in the Amazon Bedrock play area.
7. Choose Open in play area to access an interactive interface where you can experiment with various prompts and adjust design parameters like temperature level and maximum length. When using R1 with Bedrock’s InvokeModel and Playground Console, utilize DeepSeek’s chat design template for optimum results. For example, content for inference.
   
   This is an outstanding way to check out the design’s reasoning and text generation capabilities before incorporating it into your applications. The playground supplies instant feedback, helping you comprehend how the model reacts to various inputs and letting you fine-tune your prompts for optimal outcomes.
   
   You can quickly evaluate the model in the play area through the UI. However, to conjure up the released model programmatically with any Amazon Bedrock APIs, you require to get the endpoint ARN.
   
   Run reasoning utilizing guardrails with the released DeepSeek-R1 endpoint
   
   The following code example demonstrates how to carry out reasoning using a deployed DeepSeek-R1 model through Amazon Bedrock utilizing the invoke_model and ApplyGuardrail API. You can create a guardrail using the Amazon Bedrock console or the API. For the example code to produce the guardrail, see the GitHub repo. After you have actually created the guardrail, use the following code to implement guardrails. The script initializes the bedrock_runtime client, configures inference parameters, and sends out a request to create text based upon a user prompt.
   
   Deploy DeepSeek-R1 with SageMaker JumpStart
   
   SageMaker JumpStart is an artificial intelligence (ML) hub with FMs, built-in algorithms, and prebuilt ML solutions that you can release with simply a few clicks. With SageMaker JumpStart, you can tailor pre-trained designs to your usage case, with your data, and release them into production utilizing either the UI or wiki.myamens.com SDK.
   
   Deploying DeepSeek-R1 design through SageMaker JumpStart provides 2 hassle-free approaches: using the instinctive SageMaker JumpStart UI or carrying out programmatically through the SageMaker Python SDK. Let’s explore both approaches to help you select the approach that best matches your needs.
   
   Deploy DeepSeek-R1 through SageMaker JumpStart UI
   
   Complete the following actions to deploy DeepSeek-R1 using SageMaker JumpStart:
   
   1. On the SageMaker console, pick Studio in the navigation pane.
8. First-time users will be prompted to develop a domain.
9. On the SageMaker Studio console, select JumpStart in the navigation pane.
   
   The design web browser shows available designs, with details like the company name and design abilities.
   
   4. Search for DeepSeek-R1 to view the DeepSeek-R1 model card. Each design card reveals key details, including:
   
   - Model name
10. Provider name
11. Task category (for instance, Text Generation). Bedrock Ready badge (if relevant), showing that this model can be registered with Amazon Bedrock, permitting you to use Amazon Bedrock APIs to invoke the model
    
    5. Choose the design card to see the model details page.
    
    The design details page consists of the following details:
    
    - The model name and company details. Deploy button to deploy the design. About and Notebooks tabs with detailed details
    
    The About tab consists of important details, such as:
    
    - Model description.
12. License details.
13. Technical requirements.
14. Usage standards
    
    Before you deploy the design, it’s recommended to review the design details and license terms to confirm compatibility with your usage case.
    
    6. Choose Deploy to continue with release.
    
    7. For Endpoint name, utilize the instantly generated name or create a custom one.
15. For Instance type ¸ pick a circumstances type (default: ml.p5e.48 xlarge).
16. For Initial circumstances count, get in the number of circumstances (default: 1). Selecting suitable circumstances types and counts is vital for expense and efficiency optimization. Monitor your deployment to change these settings as needed.Under Inference type, Real-time reasoning is selected by default. This is optimized for sustained traffic and low latency.
17. Review all setups for precision. For this design, surgiteams.com we highly suggest sticking to SageMaker JumpStart default settings and making certain that network isolation remains in location.
18. Choose Deploy to release the design.
    
    The deployment procedure can take a number of minutes to complete.
    
    When release is complete, your endpoint status will change to InService. At this moment, the design is prepared to accept reasoning requests through the endpoint. You can keep track of the release development on the SageMaker console Endpoints page, which will show appropriate metrics and status details. When the implementation is complete, you can conjure up the design utilizing a SageMaker runtime client and integrate it with your applications.
    
    Deploy DeepSeek-R1 utilizing the SageMaker Python SDK
    
    To begin with DeepSeek-R1 utilizing the SageMaker Python SDK, you will require to set up the SageMaker Python SDK and make certain you have the needed AWS authorizations and environment setup. The following is a detailed code example that demonstrates how to deploy and use DeepSeek-R1 for inference programmatically. The code for deploying the model is offered in the Github here. You can clone the note pad and run from SageMaker Studio.
    
    You can run extra requests against the predictor:
    
    Implement guardrails and run inference with your SageMaker JumpStart predictor
    
    Similar to Amazon Bedrock, you can likewise use the ApplyGuardrail API with your SageMaker JumpStart predictor. You can develop a guardrail utilizing the Amazon Bedrock console or the API, and implement it as revealed in the following code:
    
    Clean up
    
    To avoid unwanted charges, complete the steps in this section to tidy up your resources.
    
    Delete the Amazon Bedrock Marketplace implementation
    
    If you deployed the design utilizing Amazon Bedrock Marketplace, complete the following actions:
    
    1. On the Amazon Bedrock console, under Foundation models in the navigation pane, choose Marketplace deployments.
19. In the Managed deployments section, locate the endpoint you want to delete.
20. Select the endpoint, and on the Actions menu, pick Delete.
21. Verify the endpoint details to make certain you’re erasing the right implementation: 1. name.
22. Model name.
23. Endpoint status
    
    Delete the SageMaker JumpStart predictor
    
    The SageMaker JumpStart design you released will sustain expenses if you leave it running. Use the following code to delete the endpoint if you want to stop sustaining charges. For hb9lc.org more details, see Delete Endpoints and Resources.
    
    Conclusion
    
    In this post, we checked out how you can access and release the DeepSeek-R1 model using Bedrock Marketplace and SageMaker JumpStart. Visit SageMaker JumpStart in SageMaker Studio or Amazon Bedrock Marketplace now to begin. For more details, refer to Use Amazon Bedrock tooling with Amazon SageMaker JumpStart designs, SageMaker JumpStart pretrained models, Amazon SageMaker JumpStart Foundation Models, Amazon Bedrock Marketplace, and Getting going with Amazon SageMaker JumpStart.
    
    About the Authors
    
    Vivek Gangasani is a Lead Specialist Solutions Architect for Inference at AWS. He assists emerging generative AI business construct innovative solutions using AWS services and accelerated calculate. Currently, he is focused on developing strategies for fine-tuning and enhancing the inference performance of big language designs. In his spare time, Vivek takes pleasure in hiking, enjoying motion pictures, and trying different foods.
    
    Niithiyn Vijeaswaran is a Generative AI Specialist Solutions Architect with the Third-Party Model Science group at AWS. His location of focus is AWS AI accelerators (AWS Neuron). He holds a Bachelor’s degree in Computer Science and Bioinformatics.
    
    Jonathan Evans is an Expert Solutions Architect working on generative AI with the Third-Party Model Science group at AWS.
    
    Banu Nagasundaram leads product, engineering, and strategic collaborations for Amazon SageMaker JumpStart, SageMaker’s artificial intelligence and generative AI center. She is passionate about building solutions that help customers accelerate their AI journey and unlock service worth.