QuantType

Description

Quantization type

Definition

1
2
3
4
5
6
7
8
9
enum QuantType: int {
    QUANT_UNDEFINED = 0,
    QUANT_UNQUANT = 0,
    QUANT_INT4,
    QUANT_INT8,
    QUANT_INT16,
    QUANT_FLOAT8,
    QUANT_FLOAT16
};

Members

Member

Description

QUANT_UNDEFINED

No quantization

QUANT_UNQUANT

No quantization

QUANT_INT4

INT4 quantization. Not supported yet.

QUANT_INT8

INT8 quantization.

QUANT_INT16

INT16 quantization. Not supported yet.

QUANT_FLOAT8

FLOAT8 quantization. Not supported yet.

QUANT_FLOAT16

FLOAT16 quantization. Not supported yet.