Package com.gencior.triton.core.pojo

Class TritonModelConfig

java.lang.Object
com.gencior.triton.core.pojo.TritonModelConfig
public final class TritonModelConfig extends Object
Encapsulates configuration information for a Triton model.

This class represents the runtime configuration of a deployed model including its platform, backend, runtime environment, batching capabilities, and associated model files. This configuration is read from Triton's model configuration files and defines how the model is executed and what data formats it expects.

This is an immutable object that wraps the gRPC message ModelConfig.

Since:
1.0.0
Author:
sachachoumiloff

  • Method Details

    • fromProto

      public static TritonModelConfig fromProto(ModelConfigOuterClass.ModelConfig proto)

    • getName

      public String getName()
      Returns the name of the model.

    • getPlatform

      public String getPlatform()
      Returns the platform type (e.g., "tensorflow_savedmodel", "tensorrt_plan").

    • getBackend

      public String getBackend()
      Returns the backend implementation name.

    • getRuntime

      public String getRuntime()
      Returns the runtime environment name.

    • getMaxBatchSize

      public int getMaxBatchSize()
      Returns the maximum batch size supported by the model, or 0 if batching is disabled.

    • getDefaultModelFilename

      public String getDefaultModelFilename()
      Returns the default model filename.

    • getCcModelFilenames

      public Map<String,String> getCcModelFilenames()
      Returns an unmodifiable map of compute capability to model filename mappings.

    • getMetricTags

      public Map<String,String> getMetricTags()
      Returns an unmodifiable map of custom metric tags for this model.

    • toString

      public String toString()
      Overrides:
      toString in class Object