QuantizationMode
Enums · AiDotNet.Enums
Specifies the quantization mode for model optimization and export.
Fields
| Field | Summary |
|---|---|
QuantizationMode Dynamic |
Dynamic quantization (quantize at runtime) |
QuantizationMode Float16 |
16-bit floating point quantization |
QuantizationMode Float32 |
32-bit floating point (full precision, no quantization) |
QuantizationMode Int4 |
4-bit integer quantization (16 levels). |
QuantizationMode Int8 |
8-bit integer quantization |
QuantizationMode Mixed |
Mixed precision (some layers quantized, some not) |
QuantizationMode None |
No quantization applied |
int value__ |