The Cloud ML Engine charges for training models to get predictions but is free for managing resources on the cloud:
- Training:
- Predefined Scales per hour:
- Basic $0.28
- Standard $2.90
- Machine per hour:
- Standard $0.28
- Large $0.70
- Batch processing: $0.92 per node per hour
- Online Prediction: $0.3 per node per hour
- Predefined Scales per hour:
The applications of the ML Engine include:
- To train any model at scale and host these models to generate predictions
- Cloud ML Engine uses the IAM service to manage access to resources
- Additionally, it supports cloud audit log, which is used to generate logs for API operations