TI Deep Learning Product User Guide
|
This structure define the parameters of Batch Norm layer in TIDL.
Data Fields | |
int32_t | weights |
int32_t | bias |
int32_t | numChannels |
int32_t | biasQ |
int32_t | inDataQ |
int32_t | outDataQ |
int32_t | weightsQ |
float32_tidl | weightScale |
float32_tidl | biasScale |
int32_t | zeroWeightValue |
int32_t sTIDL_BatchNormParams_t::weights |
Offset for kernel parameters
int32_t sTIDL_BatchNormParams_t::bias |
Offset for bias parameters
int32_t sTIDL_BatchNormParams_t::numChannels |
Number of channels channels to be processed
int32_t sTIDL_BatchNormParams_t::biasQ |
Q value of Bias parameter
int32_t sTIDL_BatchNormParams_t::inDataQ |
Q value of the in data
int32_t sTIDL_BatchNormParams_t::outDataQ |
Q value expected for out data
int32_t sTIDL_BatchNormParams_t::weightsQ |
Q value of Kernel weights
float32_tidl sTIDL_BatchNormParams_t::weightScale |
floating point scale for weight
float32_tidl sTIDL_BatchNormParams_t::biasScale |
floating point scale for bias
int32_t sTIDL_BatchNormParams_t::zeroWeightValue |
value of weights added for dynamic quantSytle