TI Deep Learning Product User Guide
sTIDL_Layer_t Struct Reference

Detailed Description

This structure define the common layer parameters in TIDL.

Data Fields

sTIDL_LayerParams_t layerParams
 
sTIDL_ActParams_t actParams
 
int32_t layerType
 
int32_t numInBufs
 
int32_t numOutBufs
 
sTIDL_DataParams_t inData [TIDL_NUM_IN_BUFS]
 
sTIDL_DataParams_t outData [TIDL_NUM_OUT_BUFS]
 
int32_t coreID
 
int32_t layersGroupId
 
int32_t weightsElementSizeInBits
 
int32_t strideOffsetMethod
 

Field Documentation

◆ layerParams

sTIDL_LayerParams_t sTIDL_Layer_t::layerParams

◆ actParams

sTIDL_ActParams_t sTIDL_Layer_t::actParams

Feature activation type used by TIDL layer

◆ layerType

int32_t sTIDL_Layer_t::layerType

Layer Type as defined by eTIDL_LayerType

◆ numInBufs

int32_t sTIDL_Layer_t::numInBufs

Number of input data buffers for the layer

◆ numOutBufs

int32_t sTIDL_Layer_t::numOutBufs

Number of output data buffers for the layer

◆ inData

sTIDL_DataParams_t sTIDL_Layer_t::inData[TIDL_NUM_IN_BUFS]

Input data buffers details

◆ outData

sTIDL_DataParams_t sTIDL_Layer_t::outData[TIDL_NUM_OUT_BUFS]

output data buffers details

◆ coreID

int32_t sTIDL_Layer_t::coreID

Processing core ID

◆ layersGroupId

int32_t sTIDL_Layer_t::layersGroupId

Group of layers in the net are processed together. This unique number identify the currently processing group

◆ weightsElementSizeInBits

int32_t sTIDL_Layer_t::weightsElementSizeInBits

Size of compute layer weight parameters in bytes

◆ strideOffsetMethod

int32_t sTIDL_Layer_t::strideOffsetMethod

Offset selection method for stride. eTIDL_StrideOffsetMethod