Model Serving

Fire Insights allows you to save your models. These models can be saved to:

• HDFS : when running on a Hadoop Cluster
• S3 : when running on AWS
• ADLS : when running on Azure
• Local file system : when running on your laptop or independent machine

Once these models are saved, they can be served in various ways.

• Scoring with Workflows
  • Data Preparation and Scoring Environments
  • Workflow Patterns for Scoring Models
  • Using 3 Workflows
  • Using 2 workflows
• Serving Spark MLlib Models
  • Batch Model Scoring:
  • Online Scoring with Kafka and Spark Streaming:
• Serving H2O Models
  • Serving H2O MOJO models
  • Serving H2O POJO models
• Serving AWS SageMaker models
• Serving Scikit Learn Models
• Serving Tensorflow Models
• Integration with MLflow