Module universal_inference 

Universal inference system for diverse neural network architectures

This module provides a flexible, extensible inference engine that can handle various model architectures defined in JSON format, without hardcoding specific layer types or architectures.

Structs

ModelMetadata

Training metadata

OutputSpec

Output specification for the model

UniversalModel

Universal model definition

Enums

Activation

Activation function types supported by the inference engine

InputSpec

Input specification for the model

Layer

Layer type definition for the universal inference system

Tensor

Multi-dimensional tensor representation for intermediate activations