Vector implementation for 1x1 QLinearConv, requires data to be reshaped from (M,C,1,1) to (M,C') where C' is padded to next multiple of 16, requires bias to be shifted, i.e. tbias - tw_1x1.reshape(M,-1).sum(1) * X_zero_point, requires KH==KW==1, INP_W%16=0, OUT_W_PAD%16=0, STEP_H==1|2, STEP_W==1|2, QLinearConv1x1Stream<a,a,26,32,28,32,1,1,1,3,4,1,1,1> total = 1354 vs 1550 QLinearConv1x1Stream<a,a,26,32,28,16,2,2,1,3,4,1,1,1> total = 2345 vs 2319.

#include <qlinearconv.h>

Public Member Functions
	QLinearConv1x1Stream (int32_t(&b)[M], float x_scale, float w_scale, float y_scale, TT x_zero, TTPARAM w_zero, TT y_zero)

void	filter (input_window< TT > in, input_stream< TTPARAM > weights, output_stream< TT > *out)

Static Public Member Functions
static void	registerKernelClass ()

The documentation for this class was generated from the following files:

design/aie_src/qlinearconv.h
design/aie_src/qlinearconv.cc

Public Member Functions

Static Public Member Functions