Deploy DeepSeek-R1 distilled Llama models with Amazon Bedrock Custom Model Import

I apologize for the difficulty you're experiencing with importing the DeepSeek-R1 distilled Llama model into Amazon Bedrock. Based on the information you've provided, it seems you're following the correct steps, but encountering an error during the import process. Here are a few suggestions that might help resolve the issue:

1. Double-check the model files: Ensure that you have downloaded all the required files for the specific DeepSeek-R1 distilled Llama model you're trying to import (either the 8B or 70B version). Verify that you have the following files:

   • config.json
   • tokenizer.json
   • tokenizer_config.json
   • tokenizer.model
   • All relevant *.safetensors files

2. File integrity: Make sure that all files were downloaded completely and without corruption. You might want to try re-downloading the files to ensure they're not damaged.

3. S3 bucket permissions: Verify that the S3 bucket where you've uploaded the model files has the correct permissions for Amazon Bedrock to access them.

4. Model package preparation: Ensure that you've properly prepared the model package by uploading all the required files to the same S3 bucket and folder.

5. Import process: When importing the model in the Amazon Bedrock console, double-check that you're providing the correct S3 path to your model files.

6. Supported formats: Confirm that you're using a supported model format. Amazon Bedrock Custom Model Import supports specific versions of the DeepSeek-R1 distilled Llama models.

7. Error messages: Pay close attention to any specific error messages you receive during the import process. These can often provide clues about what might be going wrong.

8. AWS support: If you continue to experience issues after verifying all of the above, it may be helpful to contact AWS support for more detailed assistance.

Remember that the process of importing custom models can be complex, and it's not uncommon to encounter challenges. Persistence and careful attention to detail are often key to successfully deploying these models in Amazon Bedrock.
Sources
Deploy DeepSeek-R1 distilled Llama models with Amazon Bedrock Custom Model Import | AWS Machine Learning Blog
Deploy DeepSeek-R1 Distilled Llama models in Amazon Bedrock | AWS Machine Learning Blog
Customized Models - Amazon Bedrock Custom Model Import - AWS
DeepSeek-R1 models now available on AWS | AWS News Blog