add

0: FP32 quantization.
1: SQ8U quantization.
2: SQ4U quantization.
3: FP16 quantization.
4: SQ8 quantization.
5: SQ4 quantization.

def add(n: int, input: numpy.ndarray, consecutive: int, level: int) -> int

Build a graph index.

Parameter	Description	Data Type	Value Range
n	Data volume.	int	≥ 1 It is recommended that the value be less than 1 billion.
input	Data.	numpy.ndarray	The value cannot be null. The expected length is n × K_features (data volume × feature dimension).
consecutive	Block size.	int	[1, 31]
level	Quantization level.	int	[0, 5] 0: FP32 quantization. 1: SQ8U quantization. 2: SQ4U quantization. 3: FP16 quantization. 4: SQ8 quantization. 5: SQ4 quantization.

Data Type	Description
int	The API is encapsulated by the underlying C++ API, which controls the exception throwing.

Parent topic: Python