Automated ML microservice and function generation for cloud native platforms
One example method includes receiving, as an input, an ML pipeline definition, identifying a group of layers required to be created for the ML pipeline definition, for one of more of the layers, receiving input concerning one or more characteristics of the layer, creating the layers for which input has been received, and packaging the created layers with the ML pipeline definition to create a production-ready ML model.
1. A method, comprising:
receiving, as an input, an ML pipeline definition;
identifying a group of layers required to be created for the ML pipeline definition;
for one of more of the layers, receiving input concerning one or more characteristics of the layer;
creating the layers for which input has been received; and
packaging the created layers with the ML pipeline definition to create a production-ready ML model.
2. The method as recited in claim 1 , wherein the production-ready ML model comprises an inference microservice.
3. The method as recited in claim 2 , further comprising implementing a training microservice operable to implement ML training of the inference microservice.
4. The method as recited in claim 1 , wherein the production-ready ML model is operable in a cloud native environment.
5. The method as recited in claim 1 , wherein the created layers comprise a pipeline execution layer, a microservice/function wrapper layer, an input/output data transformation layer, a test suites layer, a deployment definition layer, and an integration wrapper layer.
6. The method as recited in claim 1 , wherein the production-ready ML model is not constrained for use with any particular vendor.
7. The method as recited in claim 1 , wherein one or more of the layers is created with a respective user-customizable template.
8. The method as recited in claim 1 , further comprising creating the ML pipeline definition.
9. The method as recited in claim 1 , further comprising deploying the production-ready ML model in a production environment.
10. The method as recited in claim 1 , wherein one of the created layers is a microservice/function wrapper layer that specifies an inference deployment platform, a training deployment platform, and an invocation mechanism.
11. A non-transitory storage medium having stored therein instructions that are executable by one or more hardware processors to perform operations comprising:
receiving, as an input, an ML pipeline definition;
identifying a group of layers required to be created for the ML pipeline definition;
for one of more of the layers, receiving input concerning one or more characteristics of the layer;
creating the layers for which input has been received; and
packaging the created layers with the ML pipeline definition to create a production-ready ML model.
12. The non-transitory storage medium as recited in claim 11 , wherein the production-ready ML model comprises an inference microservice.
13. The non-transitory storage medium as recited in claim 12 , wherein the operations further comprise implementing a training microservice operable to implement ML training of the inference microservice.
14. The non-transitory storage medium as recited in claim 11 , wherein the production-ready ML model is operable in a cloud native environment.
15. The non-transitory storage medium as recited in claim 11 , wherein the created layers comprise a pipeline execution layer, a microservice/function wrapper layer, an input/output data transformation layer, a test suites layer, a deployment definition layer, and an integration wrapper layer.
16. The non-transitory storage medium as recited in claim 11 , wherein the production-ready ML model is not constrained for use with any particular vendor.
17. The non-transitory storage medium as recited in claim 1 , wherein one or more of the layers is created with a respective user-customizable template.
18. The non-transitory storage medium as recited in claim 1 , wherein the operations further comprise creating the ML pipeline definition.
19. The non-transitory storage medium as recited in claim 1 , wherein the operations further comprise deploying the production-ready ML model in a production environment.
20. The non-transitory storage medium as recited in claim 1 , wherein one of the created layers is a microservice/function wrapper layer that specifies an inference deployment platform, a training deployment platform, and an invocation mechanism.