QLinearConvHx4PktStream< TT, TTPARAM, INP_H, INP_W, OUT_W, OUT_W_PAD, STEP_H, STEP_W, B, C, M, KH, KW, GROUP > Class Template Reference

Vector implementation for Hx4 QLinearConv, padding with y_zero, requires data to be arranged in (M,C,KH,KW) -> (M,C,KH',4) where KH' = KH*4 padded to nearest 16, e.g. [a,b,c,d,e,f,g,h,i] -> [a,b,c,0, d,e,f,0, g,h,i,0, 0,0,0,0], requires bias to be shifted, i.e. tbias - tw.reshape(M,-1).sum(1) * X_zero_point, requires KW<=3, INP_W%16=0, OUT_W_PAD%16=0, STEP_H==1|2, STEP_W==1|2,.

#include <qlinearconv.h>

Public Member Functions
	QLinearConvHx4PktStream (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_pktstream *in_s, 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