- 新しい順
- 投票が多い順
- コメントが多い順
Currently, we only provide ability to use one model per endpoint at a given time. So, it's not possible for now, to use use one endpoint to host multiple AWS Comprehend custom models at same time.
Can you also provide information on what the use case for hosting multiple AWS Comprehend custom models using one endpoint? That'll be helpful feedback for us to further this ask.
If you are looking to change model for the endpoint (From Model A to Model B), you can do that by using https://docs.aws.amazon.com/comprehend/latest/dg/manage-endpoints-update.html
Hi, You create an endpoint to run real-time analysis using a custom model. That means that you can host one custom classification model per Endpoint.
For more details please refer to the documentation or blog post.
I want to classify text using 100s of labels/tags. Since custom comprehend multi-label mode can only support up-to 100 labels per model, I have to build one model for each x labels. Deploying multiple endpoint is very costly, so I started wondering if using one endpoint is an option similar to multi-model endpoint in SageMaker.