IntelĀ® Machine Learning Scaling Library
2018
A library providing an efficient implementation of communication patterns used in deep learning.
|
A struct to hold quantization parameters. More...
#include <mlsl.hpp>
Public Attributes | |
char * | lib_path |
char * | quant_buffer_func_name |
char * | dequant_buffer_func_name |
char * | reduce_sum_func_name |
size_t | block_size |
size_t | elem_in_block |
A struct to hold quantization parameters.
Holds information about the quantization library and functions.
size_t MLSL::QuantParams::block_size |
Quantization meta data: block size in bytes.
char* MLSL::QuantParams::dequant_buffer_func_name |
Name of the function for buffer dequantization.
size_t MLSL::QuantParams::elem_in_block |
Quantization meta data: number of elements in a block.
char* MLSL::QuantParams::lib_path |
Quantization library path.
char* MLSL::QuantParams::quant_buffer_func_name |
Name of the function for buffer quantization.
char* MLSL::QuantParams::reduce_sum_func_name |
Name of the function for reduction of quantized buffers.