Fine-tuned model not appearing in Playground or Batch Inference

The issue you're experiencing with your fine-tuned model not appearing in the Playground or Batch Inference options could be due to a few reasons:

1. Model Availability: Even though you've purchased provisional throughput and the model is in-service, there might be a delay in the model becoming fully available across all Bedrock interfaces. Sometimes, it can take a short while for newly created or updated resources to propagate through the system.

2. Permissions: Ensure that your IAM role or user has the necessary permissions to access and use custom models in Amazon Bedrock. You might need specific permissions for viewing and using fine-tuned models in addition to the general Bedrock access.

3. Model Type Compatibility: Not all fine-tuned models are automatically compatible with every Bedrock feature. Some models may have limitations on where they can be used within the Bedrock ecosystem.

4. Cross-Region Inference: If your model requires cross-region inference, you might need to set up an inference profile for it. This is particularly relevant for certain model types, like some Anthropic models.

5. Console Caching: Sometimes, the AWS console can cache older data. Try refreshing your browser or clearing your cache to ensure you're seeing the most up-to-date information.

To resolve this issue:

1. Double-check your IAM permissions to ensure you have full access to Bedrock features, including custom models.

2. Verify in the Bedrock console that your fine-tuned model is indeed listed as available and in-service.

3. If the model is based on a foundation model that requires cross-region inference, create an appropriate inference profile for it.

4. Try accessing the model programmatically using the AWS CLI or SDK to see if it's available through those methods.

5. If the issue persists, it may be worth contacting AWS support for further assistance, as there could be a specific configuration or account-level issue that needs to be addressed.

Remember that the exact steps might vary depending on the specific model you've fine-tuned and the Bedrock features you're trying to use. If you continue to face issues after trying these steps, reaching out to AWS support would be advisable for more tailored assistance.
Sources
Error in AWS Bedrock Batch Inteference | AWS re:Post
Bedrock API invocation error - on demand throughput isn's supported | AWS re:Post
Bedrock API with Cross Region Inference is not working | AWS re:Post