Module quantization 

Model quantization support for performance optimization.

This module provides support for quantized ONNX models:

• INT8 quantization for CPU inference
• FP16 quantization for GPU inference
• Dynamic quantization
• Quantization configuration and validation

Quantized models can significantly reduce:

• Model size (2-4x smaller)
• Memory usage (2-4x less)
• Inference latency (1.5-3x faster)

With minimal accuracy loss (typically <1%).

Structs

ModelQuantizer

Model quantizer (placeholder for actual quantization logic)

QuantizationBenefits

Estimated benefits of quantization

QuantizationConfig

Configuration for model quantization

QuantizedModelInfo

Information about a quantized model

Enums

QuantizationMethod

Quantization method

QuantizationPrecision

Quantization precision level