TI Deep Learning Product User Guide
sTIDL_InnerProductParams_t Struct Reference

Detailed Description

This structure define the parameters of Inner Product (Fully connected) layer in TIDL.

Data Fields

int32_t weights
 
int32_t bias
 
int32_t activationType
 
int32_t numInNodes
 
int32_t numOutNodes
 
int32_t weightsQ
 
float32_tidl weightScale
 
float32_tidl biasScale
 
int32_t zeroWeightValue
 
int32_t biasQ
 
int32_t inDataQ
 
int32_t outDataQ
 
int32_t interDataQ
 
int32_t biasB
 

Field Documentation

◆ weights

int32_t sTIDL_InnerProductParams_t::weights

Offset for kernel parameters

◆ bias

int32_t sTIDL_InnerProductParams_t::bias

Offset for bias parameters

◆ activationType

int32_t sTIDL_InnerProductParams_t::activationType

activation type to be used

◆ numInNodes

int32_t sTIDL_InnerProductParams_t::numInNodes

Number of elements in the flattened input

◆ numOutNodes

int32_t sTIDL_InnerProductParams_t::numOutNodes

Number of elements in the output

◆ weightsQ

int32_t sTIDL_InnerProductParams_t::weightsQ

Q value of Kernel weights

◆ weightScale

float32_tidl sTIDL_InnerProductParams_t::weightScale

floating point scale for weight

◆ biasScale

float32_tidl sTIDL_InnerProductParams_t::biasScale

floating point scale for bias

◆ zeroWeightValue

int32_t sTIDL_InnerProductParams_t::zeroWeightValue

value of weights added for dynamic quantSytle

◆ biasQ

int32_t sTIDL_InnerProductParams_t::biasQ

Q value kernel Bias

◆ inDataQ

int32_t sTIDL_InnerProductParams_t::inDataQ

Q value of the in data

◆ outDataQ

int32_t sTIDL_InnerProductParams_t::outDataQ

Q value expected for out data

◆ interDataQ

int32_t sTIDL_InnerProductParams_t::interDataQ

Q value intermediate output data

◆ biasB

int32_t sTIDL_InnerProductParams_t::biasB

Newly added