# Audio Embedding Model #  

**MRE Plugin Class**
- Featurer

**Description**:  
This model endpoint uses [OpenL3](https://github.com/marl/openl3) library to embed audio signals into vectors.
This endpoint is created based on a BYOC(Bring Your Own Custom) inference image.


**Use Cases**:  
- Ace serve and double fault serve audio embedding
- Any audio classification use cases based on audio embedding

**Model Type**:  
- Open-source models hosted as Amazon SageMaker endpoint  

**Methods for training data collection and annotation**  
N/A

**Methods for model training**  
N/A  

**Methods for model hosting**  
-Build a docker image based on Multi-Model-Server by runing the build_your_custom_image.sh
-Create SageMaker model based on the BYOC inference image  
-Create SgaeMaker endpoint based on the SageMaker model  
-See details in the notebook