Enum QuantizationMode
Specifies the quantization mode for model optimization and export.
public enum QuantizationMode
Fields
Dynamic = 4Dynamic quantization (quantize at runtime)
Float16 = 216-bit floating point quantization
Float32 = 332-bit floating point (full precision, no quantization)
Int8 = 18-bit integer quantization
Mixed = 5Mixed precision (some layers quantized, some not)
None = 0No quantization applied