Scalable Inference Serving · Schema

TensorMetadata

Metadata describing a single input or output tensor.

AICNCFDeploymentInferenceKubernetesLLMMachine LearningModel ServingMLOpsScalability

Properties

Name Type Description
name string Name of the tensor as defined by the model.
datatype object
shape array Shape of the tensor. Use -1 for dynamic dimensions.
parameters object Optional tensor-specific parameters.
View JSON Schema on GitHub

JSON Schema

scalable-inference-serving-tensormetadata-schema.json Raw ↑ 
{
  "$schema": "https://json-schema.org/draft/2020-12/schema",
  "$id": "#/components/schemas/TensorMetadata",
  "title": "TensorMetadata",
  "type": "object",
  "description": "Metadata describing a single input or output tensor.",
  "required": [
    "name",
    "datatype",
    "shape"
  ],
  "properties": {
    "name": {
      "type": "string",
      "description": "Name of the tensor as defined by the model."
    },
    "datatype": {
      "$ref": "#/components/schemas/TensorDatatype"
    },
    "shape": {
      "type": "array",
      "description": "Shape of the tensor. Use -1 for dynamic dimensions.",
      "items": {
        "type": "integer"
      },
      "example": [
        -1,
        224,
        224,
        3
      ]
    },
    "parameters": {
      "type": "object",
      "additionalProperties": true,
      "description": "Optional tensor-specific parameters."
    }
  }
}