Creating a model

Creating a model in a project that uses ML Runtimes is similar to model creation with a legacy engine, but you must use a different parameter to specify the Runtime ID.

To create a model in a project that uses ML Runtimes you must use the --runtimeId= parameter to specify a Runtime ID (instead of using the --engineImageId= and --kernel= parameters used for a legacy engine).

To create a model in a project that uses ML Runtimes use the --runtimeId= parameter to specify a Runtime ID.

For example:

cdswctl models create --targetFilePath=predict.py --targetFunctionName=predict 
  --projectId=4 --name=created-using-cdswctl --description=created-using-cdswctl 
  --memoryMb=1024 --authEnabled --cpuMillicores=250 --autoBuildModel --autoDeployModel 
  --examples='{"request":{"petal_length":1}}' --runtimeId=1