../../../model-artifacts//sp-ort-superpoint_PbDb/artifacts/tempDir/subgraph_0_tidl_net.bin
Layer 0: TIDL_DataLayer "image" weightsElementSizeInBits=16 kernelType=1 multiCoreMode=TIDL_NOT_MULTI_CORE strideOffsetMethod=TIDL_StrideOffsetTopLeft numChannels=0 dataQ=0 actParams: actType=TIDL_NoAct slopeScale=0.000000 clipMin/Max=(0.000000,0.000000) Graph compiler Info: space:DDR baseMem(bytes):33854464-34379776 size(elem):262656 bufferWidth(elem):262144 activeBufferWidth(elem):262144 bufferHeight(elem):1 padRows:0 padCols:0 Outputs: [0] numDim=6 dims=[1,1,1,1,512,512] elementType=TIDL_SignedShort padH/W=[0,0] batchPadH/W=[0,0] numBatchH/W=[1,1] pitch=[262144,262144,262144,262144,512] dataQ=0 roundBits=0 min/maxValue=(0,0) min/maxTensorValue=(0.000000,255.000000) tensorScale=32768.000000 tensorZeroPoint=0
Data [0 0]
Layer 1: TIDL_ConvolutionLayer "/model/relu/Relu_output_0" weightsElementSizeInBits=16 kernelType=1 multiCoreMode=TIDL_NOT_MULTI_CORE strideOffsetMethod=TIDL_StrideOffsetTopLeft convolutionType=0 numInChannels=1 numOutChannels=64 numGroups=1 kernelH/W=[3,3] strideH/W=[1,1] padH/W=[1,1] dilationH/W=[1,1] weightScale=166.134399 biasScale=1849.620483 actParams: actType=TIDL_RelU slopeScale=0.000000 clipMin/Max=(0.000000,0.000000) Graph compiler Info: space:DDR baseMem(bytes):0-33854464 size(elem):16927232 bufferWidth(elem):264480 activeBufferWidth(elem):262144 bufferHeight(elem):64 padRows:0 padCols:0 Outputs: [1] numDim=6 dims=[1,1,1,64,512,512] elementType=TIDL_UnsignedShort padH/W=[0,0] batchPadH/W=[0,0] numBatchH/W=[1,1] pitch=[16926720,16926720,16926720,264480,512] dataQ=0 roundBits=12 min/maxValue=(0,0) min/maxTensorValue=(0.000000,28.772638) tensorScale=1329.075195 tensorZeroPoint=0
Convolution [1 1]
Layer 0: TIDL_DataLayer "image" weightsElementSizeInBits=16 kernelType=1 multiCoreMode=TIDL_NOT_MULTI_CORE strideOffsetMethod=TIDL_StrideOffsetTopLeft numChannels=0 dataQ=0 actParams: actType=TIDL_NoAct slopeScale=0.000000 clipMin/Max=(0.000000,0.000000) Graph compiler Info: space:DDR baseMem(bytes):33854464-34379776 size(elem):262656 bufferWidth(elem):262144 activeBufferWidth(elem):262144 bufferHeight(elem):1 padRows:0 padCols:0 Outputs: [0] numDim=6 dims=[1,1,1,1,512,512] elementType=TIDL_SignedShort padH/W=[0,0] batchPadH/W=[0,0] numBatchH/W=[1,1] pitch=[262144,262144,262144,262144,512] dataQ=0 roundBits=0 min/maxValue=(0,0) min/maxTensorValue=(0.000000,255.000000) tensorScale=32768.000000 tensorZeroPoint=0 ->Layer 1: TIDL_ConvolutionLayer "/model/relu/Relu_output_0" weightsElementSizeInBits=16 kernelType=1 multiCoreMode=TIDL_NOT_MULTI_CORE strideOffsetMethod=TIDL_StrideOffsetTopLeft convolutionType=0 numInChannels=1 numOutChannels=64 numGroups=1 kernelH/W=[3,3] strideH/W=[1,1] padH/W=[1,1] dilationH/W=[1,1] weightScale=166.134399 biasScale=1849.620483 actParams: actType=TIDL_RelU slopeScale=0.000000 clipMin/Max=(0.000000,0.000000) Graph compiler Info: space:DDR baseMem(bytes):0-33854464 size(elem):16927232 bufferWidth(elem):264480 activeBufferWidth(elem):262144 bufferHeight(elem):64 padRows:0 padCols:0 Outputs: [1] numDim=6 dims=[1,1,1,64,512,512] elementType=TIDL_UnsignedShort padH/W=[0,0] batchPadH/W=[0,0] numBatchH/W=[1,1] pitch=[16926720,16926720,16926720,264480,512] dataQ=0 roundBits=12 min/maxValue=(0,0) min/maxTensorValue=(0.000000,28.772638) tensorScale=1329.075195 tensorZeroPoint=0
1x1x1x1x512x512
Layer 2: TIDL_ConvolutionLayer "/model/relu_1/Relu_output_0" weightsElementSizeInBits=16 kernelType=1 multiCoreMode=TIDL_NOT_MULTI_CORE strideOffsetMethod=TIDL_StrideOffsetTopLeft convolutionType=0 numInChannels=64 numOutChannels=64 numGroups=1 kernelH/W=[3,3] strideH/W=[1,1] padH/W=[1,1] dilationH/W=[1,1] weightScale=21438.263672 biasScale=1794.850830 actParams: actType=TIDL_RelU slopeScale=0.000000 clipMin/Max=(0.000000,0.000000) Graph compiler Info: space:DDR baseMem(bytes):33854464-67512320 size(elem):16828928 bufferWidth(elem):262944 activeBufferWidth(elem):262144 bufferHeight(elem):64 padRows:0 padCols:0 Outputs: [2] numDim=6 dims=[1,1,1,64,512,512] elementType=TIDL_UnsignedShort padH/W=[0,0] batchPadH/W=[0,0] numBatchH/W=[1,1] pitch=[16828416,16828416,16828416,262944,512] dataQ=0 roundBits=13 min/maxValue=(0,0) min/maxTensorValue=(0.000000,15.877086) tensorScale=3478.157227 tensorZeroPoint=0
Convolution [2 2]
Layer 1: TIDL_ConvolutionLayer "/model/relu/Relu_output_0" weightsElementSizeInBits=16 kernelType=1 multiCoreMode=TIDL_NOT_MULTI_CORE strideOffsetMethod=TIDL_StrideOffsetTopLeft convolutionType=0 numInChannels=1 numOutChannels=64 numGroups=1 kernelH/W=[3,3] strideH/W=[1,1] padH/W=[1,1] dilationH/W=[1,1] weightScale=166.134399 biasScale=1849.620483 actParams: actType=TIDL_RelU slopeScale=0.000000 clipMin/Max=(0.000000,0.000000) Graph compiler Info: space:DDR baseMem(bytes):0-33854464 size(elem):16927232 bufferWidth(elem):264480 activeBufferWidth(elem):262144 bufferHeight(elem):64 padRows:0 padCols:0 Outputs: [1] numDim=6 dims=[1,1,1,64,512,512] elementType=TIDL_UnsignedShort padH/W=[0,0] batchPadH/W=[0,0] numBatchH/W=[1,1] pitch=[16926720,16926720,16926720,264480,512] dataQ=0 roundBits=12 min/maxValue=(0,0) min/maxTensorValue=(0.000000,28.772638) tensorScale=1329.075195 tensorZeroPoint=0 ->Layer 2: TIDL_ConvolutionLayer "/model/relu_1/Relu_output_0" weightsElementSizeInBits=16 kernelType=1 multiCoreMode=TIDL_NOT_MULTI_CORE strideOffsetMethod=TIDL_StrideOffsetTopLeft convolutionType=0 numInChannels=64 numOutChannels=64 numGroups=1 kernelH/W=[3,3] strideH/W=[1,1] padH/W=[1,1] dilationH/W=[1,1] weightScale=21438.263672 biasScale=1794.850830 actParams: actType=TIDL_RelU slopeScale=0.000000 clipMin/Max=(0.000000,0.000000) Graph compiler Info: space:DDR baseMem(bytes):33854464-67512320 size(elem):16828928 bufferWidth(elem):262944 activeBufferWidth(elem):262144 bufferHeight(elem):64 padRows:0 padCols:0 Outputs: [2] numDim=6 dims=[1,1,1,64,512,512] elementType=TIDL_UnsignedShort padH/W=[0,0] batchPadH/W=[0,0] numBatchH/W=[1,1] pitch=[16828416,16828416,16828416,262944,512] dataQ=0 roundBits=13 min/maxValue=(0,0) min/maxTensorValue=(0.000000,15.877086) tensorScale=3478.157227 tensorZeroPoint=0
1x1x1x64x512x512
Layer 3: TIDL_PoolingLayer "/model/pool/MaxPool_output_0" weightsElementSizeInBits=16 kernelType=1 multiCoreMode=TIDL_NOT_MULTI_CORE strideOffsetMethod=TIDL_StrideOffsetTopLeft poolingtype=TIDL_MaxPooling numChannels=64 inDataQ=0 outDataQ=0 useCeil=0 kernelH/W=[2,2] strideH/W=[2,2] padH/W=[0,0] actParams: actType=TIDL_NoAct slopeScale=0.000000 clipMin/Max=(0.000000,0.000000) Graph compiler Info: space:DDR baseMem(bytes):0-9892864 size(elem):4946432 bufferWidth(elem):77280 activeBufferWidth(elem):65536 bufferHeight(elem):64 padRows:0 padCols:0 Outputs: [3] numDim=6 dims=[1,1,1,64,256,256] elementType=TIDL_UnsignedShort padH/W=[0,0] batchPadH/W=[0,0] numBatchH/W=[1,1] pitch=[4945920,4945920,4945920,77280,256] dataQ=0 roundBits=0 min/maxValue=(0,0) min/maxTensorValue=(0.000000,17.478796) tensorScale=3478.157227 tensorZeroPoint=0
Pooling [3 3]
Layer 2: TIDL_ConvolutionLayer "/model/relu_1/Relu_output_0" weightsElementSizeInBits=16 kernelType=1 multiCoreMode=TIDL_NOT_MULTI_CORE strideOffsetMethod=TIDL_StrideOffsetTopLeft convolutionType=0 numInChannels=64 numOutChannels=64 numGroups=1 kernelH/W=[3,3] strideH/W=[1,1] padH/W=[1,1] dilationH/W=[1,1] weightScale=21438.263672 biasScale=1794.850830 actParams: actType=TIDL_RelU slopeScale=0.000000 clipMin/Max=(0.000000,0.000000) Graph compiler Info: space:DDR baseMem(bytes):33854464-67512320 size(elem):16828928 bufferWidth(elem):262944 activeBufferWidth(elem):262144 bufferHeight(elem):64 padRows:0 padCols:0 Outputs: [2] numDim=6 dims=[1,1,1,64,512,512] elementType=TIDL_UnsignedShort padH/W=[0,0] batchPadH/W=[0,0] numBatchH/W=[1,1] pitch=[16828416,16828416,16828416,262944,512] dataQ=0 roundBits=13 min/maxValue=(0,0) min/maxTensorValue=(0.000000,15.877086) tensorScale=3478.157227 tensorZeroPoint=0 ->Layer 3: TIDL_PoolingLayer "/model/pool/MaxPool_output_0" weightsElementSizeInBits=16 kernelType=1 multiCoreMode=TIDL_NOT_MULTI_CORE strideOffsetMethod=TIDL_StrideOffsetTopLeft poolingtype=TIDL_MaxPooling numChannels=64 inDataQ=0 outDataQ=0 useCeil=0 kernelH/W=[2,2] strideH/W=[2,2] padH/W=[0,0] actParams: actType=TIDL_NoAct slopeScale=0.000000 clipMin/Max=(0.000000,0.000000) Graph compiler Info: space:DDR baseMem(bytes):0-9892864 size(elem):4946432 bufferWidth(elem):77280 activeBufferWidth(elem):65536 bufferHeight(elem):64 padRows:0 padCols:0 Outputs: [3] numDim=6 dims=[1,1,1,64,256,256] elementType=TIDL_UnsignedShort padH/W=[0,0] batchPadH/W=[0,0] numBatchH/W=[1,1] pitch=[4945920,4945920,4945920,77280,256] dataQ=0 roundBits=0 min/maxValue=(0,0) min/maxTensorValue=(0.000000,17.478796) tensorScale=3478.157227 tensorZeroPoint=0
1x1x1x64x512x512
Layer 4: TIDL_ConvolutionLayer "/model/relu_2/Relu_output_0" weightsElementSizeInBits=16 kernelType=1 multiCoreMode=TIDL_NOT_MULTI_CORE strideOffsetMethod=TIDL_StrideOffsetTopLeft convolutionType=0 numInChannels=64 numOutChannels=64 numGroups=1 kernelH/W=[3,3] strideH/W=[1,1] padH/W=[1,1] dilationH/W=[1,1] weightScale=31147.523438 biasScale=1653.075928 actParams: actType=TIDL_RelU slopeScale=0.000000 clipMin/Max=(0.000000,0.000000) Graph compiler Info: space:DDR baseMem(bytes):9892864-18388992 size(elem):4248064 bufferWidth(elem):66368 activeBufferWidth(elem):65536 bufferHeight(elem):64 padRows:0 padCols:0 Outputs: [4] numDim=6 dims=[1,1,1,64,256,256] elementType=TIDL_UnsignedShort padH/W=[0,0] batchPadH/W=[0,0] numBatchH/W=[1,1] pitch=[4247552,4247552,4247552,66368,256] dataQ=0 roundBits=16 min/maxValue=(0,0) min/maxTensorValue=(0.000000,24.164906) tensorScale=1653.075928 tensorZeroPoint=0
Convolution [4 4]
Layer 3: TIDL_PoolingLayer "/model/pool/MaxPool_output_0" weightsElementSizeInBits=16 kernelType=1 multiCoreMode=TIDL_NOT_MULTI_CORE strideOffsetMethod=TIDL_StrideOffsetTopLeft poolingtype=TIDL_MaxPooling numChannels=64 inDataQ=0 outDataQ=0 useCeil=0 kernelH/W=[2,2] strideH/W=[2,2] padH/W=[0,0] actParams: actType=TIDL_NoAct slopeScale=0.000000 clipMin/Max=(0.000000,0.000000) Graph compiler Info: space:DDR baseMem(bytes):0-9892864 size(elem):4946432 bufferWidth(elem):77280 activeBufferWidth(elem):65536 bufferHeight(elem):64 padRows:0 padCols:0 Outputs: [3] numDim=6 dims=[1,1,1,64,256,256] elementType=TIDL_UnsignedShort padH/W=[0,0] batchPadH/W=[0,0] numBatchH/W=[1,1] pitch=[4945920,4945920,4945920,77280,256] dataQ=0 roundBits=0 min/maxValue=(0,0) min/maxTensorValue=(0.000000,17.478796) tensorScale=3478.157227 tensorZeroPoint=0 ->Layer 4: TIDL_ConvolutionLayer "/model/relu_2/Relu_output_0" weightsElementSizeInBits=16 kernelType=1 multiCoreMode=TIDL_NOT_MULTI_CORE strideOffsetMethod=TIDL_StrideOffsetTopLeft convolutionType=0 numInChannels=64 numOutChannels=64 numGroups=1 kernelH/W=[3,3] strideH/W=[1,1] padH/W=[1,1] dilationH/W=[1,1] weightScale=31147.523438 biasScale=1653.075928 actParams: actType=TIDL_RelU slopeScale=0.000000 clipMin/Max=(0.000000,0.000000) Graph compiler Info: space:DDR baseMem(bytes):9892864-18388992 size(elem):4248064 bufferWidth(elem):66368 activeBufferWidth(elem):65536 bufferHeight(elem):64 padRows:0 padCols:0 Outputs: [4] numDim=6 dims=[1,1,1,64,256,256] elementType=TIDL_UnsignedShort padH/W=[0,0] batchPadH/W=[0,0] numBatchH/W=[1,1] pitch=[4247552,4247552,4247552,66368,256] dataQ=0 roundBits=16 min/maxValue=(0,0) min/maxTensorValue=(0.000000,24.164906) tensorScale=1653.075928 tensorZeroPoint=0
1x1x1x64x256x256
Layer 5: TIDL_ConvolutionLayer "/model/relu_3/Relu_output_0" weightsElementSizeInBits=16 kernelType=1 multiCoreMode=TIDL_NOT_MULTI_CORE strideOffsetMethod=TIDL_StrideOffsetTopLeft convolutionType=0 numInChannels=64 numOutChannels=64 numGroups=1 kernelH/W=[3,3] strideH/W=[1,1] padH/W=[1,1] dilationH/W=[1,1] weightScale=21169.777344 biasScale=1075.154663 actParams: actType=TIDL_RelU slopeScale=0.000000 clipMin/Max=(0.000000,0.000000) Graph compiler Info: space:DDR baseMem(bytes):0-8496128 size(elem):4248064 bufferWidth(elem):66368 activeBufferWidth(elem):65536 bufferHeight(elem):64 padRows:0 padCols:0 Outputs: [5] numDim=6 dims=[1,1,1,64,256,256] elementType=TIDL_UnsignedShort padH/W=[0,0] batchPadH/W=[0,0] numBatchH/W=[1,1] pitch=[4247552,4247552,4247552,66368,256] dataQ=0 roundBits=15 min/maxValue=(0,0) min/maxTensorValue=(0.000000,31.540722) tensorScale=1067.970215 tensorZeroPoint=0
Convolution [5 5]
Layer 4: TIDL_ConvolutionLayer "/model/relu_2/Relu_output_0" weightsElementSizeInBits=16 kernelType=1 multiCoreMode=TIDL_NOT_MULTI_CORE strideOffsetMethod=TIDL_StrideOffsetTopLeft convolutionType=0 numInChannels=64 numOutChannels=64 numGroups=1 kernelH/W=[3,3] strideH/W=[1,1] padH/W=[1,1] dilationH/W=[1,1] weightScale=31147.523438 biasScale=1653.075928 actParams: actType=TIDL_RelU slopeScale=0.000000 clipMin/Max=(0.000000,0.000000) Graph compiler Info: space:DDR baseMem(bytes):9892864-18388992 size(elem):4248064 bufferWidth(elem):66368 activeBufferWidth(elem):65536 bufferHeight(elem):64 padRows:0 padCols:0 Outputs: [4] numDim=6 dims=[1,1,1,64,256,256] elementType=TIDL_UnsignedShort padH/W=[0,0] batchPadH/W=[0,0] numBatchH/W=[1,1] pitch=[4247552,4247552,4247552,66368,256] dataQ=0 roundBits=16 min/maxValue=(0,0) min/maxTensorValue=(0.000000,24.164906) tensorScale=1653.075928 tensorZeroPoint=0 ->Layer 5: TIDL_ConvolutionLayer "/model/relu_3/Relu_output_0" weightsElementSizeInBits=16 kernelType=1 multiCoreMode=TIDL_NOT_MULTI_CORE strideOffsetMethod=TIDL_StrideOffsetTopLeft convolutionType=0 numInChannels=64 numOutChannels=64 numGroups=1 kernelH/W=[3,3] strideH/W=[1,1] padH/W=[1,1] dilationH/W=[1,1] weightScale=21169.777344 biasScale=1075.154663 actParams: actType=TIDL_RelU slopeScale=0.000000 clipMin/Max=(0.000000,0.000000) Graph compiler Info: space:DDR baseMem(bytes):0-8496128 size(elem):4248064 bufferWidth(elem):66368 activeBufferWidth(elem):65536 bufferHeight(elem):64 padRows:0 padCols:0 Outputs: [5] numDim=6 dims=[1,1,1,64,256,256] elementType=TIDL_UnsignedShort padH/W=[0,0] batchPadH/W=[0,0] numBatchH/W=[1,1] pitch=[4247552,4247552,4247552,66368,256] dataQ=0 roundBits=15 min/maxValue=(0,0) min/maxTensorValue=(0.000000,31.540722) tensorScale=1067.970215 tensorZeroPoint=0
1x1x1x64x256x256
Layer 6: TIDL_PoolingLayer "/model/pool_1/MaxPool_output_0" weightsElementSizeInBits=16 kernelType=1 multiCoreMode=TIDL_NOT_MULTI_CORE strideOffsetMethod=TIDL_StrideOffsetTopLeft poolingtype=TIDL_MaxPooling numChannels=64 inDataQ=0 outDataQ=0 useCeil=0 kernelH/W=[2,2] strideH/W=[2,2] padH/W=[0,0] actParams: actType=TIDL_NoAct slopeScale=0.000000 clipMin/Max=(0.000000,0.000000) Graph compiler Info: space:DDR baseMem(bytes):8496128-11655168 size(elem):1579520 bufferWidth(elem):24672 activeBufferWidth(elem):16384 bufferHeight(elem):64 padRows:0 padCols:0 Outputs: [6] numDim=6 dims=[1,1,1,64,128,128] elementType=TIDL_UnsignedShort padH/W=[0,0] batchPadH/W=[0,0] numBatchH/W=[1,1] pitch=[1579008,1579008,1579008,24672,128] dataQ=0 roundBits=0 min/maxValue=(0,0) min/maxTensorValue=(0.000000,32.263885) tensorScale=1067.970215 tensorZeroPoint=0
Pooling [6 6]
Layer 5: TIDL_ConvolutionLayer "/model/relu_3/Relu_output_0" weightsElementSizeInBits=16 kernelType=1 multiCoreMode=TIDL_NOT_MULTI_CORE strideOffsetMethod=TIDL_StrideOffsetTopLeft convolutionType=0 numInChannels=64 numOutChannels=64 numGroups=1 kernelH/W=[3,3] strideH/W=[1,1] padH/W=[1,1] dilationH/W=[1,1] weightScale=21169.777344 biasScale=1075.154663 actParams: actType=TIDL_RelU slopeScale=0.000000 clipMin/Max=(0.000000,0.000000) Graph compiler Info: space:DDR baseMem(bytes):0-8496128 size(elem):4248064 bufferWidth(elem):66368 activeBufferWidth(elem):65536 bufferHeight(elem):64 padRows:0 padCols:0 Outputs: [5] numDim=6 dims=[1,1,1,64,256,256] elementType=TIDL_UnsignedShort padH/W=[0,0] batchPadH/W=[0,0] numBatchH/W=[1,1] pitch=[4247552,4247552,4247552,66368,256] dataQ=0 roundBits=15 min/maxValue=(0,0) min/maxTensorValue=(0.000000,31.540722) tensorScale=1067.970215 tensorZeroPoint=0 ->Layer 6: TIDL_PoolingLayer "/model/pool_1/MaxPool_output_0" weightsElementSizeInBits=16 kernelType=1 multiCoreMode=TIDL_NOT_MULTI_CORE strideOffsetMethod=TIDL_StrideOffsetTopLeft poolingtype=TIDL_MaxPooling numChannels=64 inDataQ=0 outDataQ=0 useCeil=0 kernelH/W=[2,2] strideH/W=[2,2] padH/W=[0,0] actParams: actType=TIDL_NoAct slopeScale=0.000000 clipMin/Max=(0.000000,0.000000) Graph compiler Info: space:DDR baseMem(bytes):8496128-11655168 size(elem):1579520 bufferWidth(elem):24672 activeBufferWidth(elem):16384 bufferHeight(elem):64 padRows:0 padCols:0 Outputs: [6] numDim=6 dims=[1,1,1,64,128,128] elementType=TIDL_UnsignedShort padH/W=[0,0] batchPadH/W=[0,0] numBatchH/W=[1,1] pitch=[1579008,1579008,1579008,24672,128] dataQ=0 roundBits=0 min/maxValue=(0,0) min/maxTensorValue=(0.000000,32.263885) tensorScale=1067.970215 tensorZeroPoint=0
1x1x1x64x256x256
Layer 7: TIDL_ConvolutionLayer "/model/relu_4/Relu_output_0" weightsElementSizeInBits=16 kernelType=1 multiCoreMode=TIDL_NOT_MULTI_CORE strideOffsetMethod=TIDL_StrideOffsetTopLeft convolutionType=0 numInChannels=64 numOutChannels=128 numGroups=1 kernelH/W=[3,3] strideH/W=[1,1] padH/W=[1,1] dilationH/W=[1,1] weightScale=24064.240234 biasScale=1369.173340 actParams: actType=TIDL_RelU slopeScale=0.000000 clipMin/Max=(0.000000,0.000000) Graph compiler Info: space:DDR baseMem(bytes):0-4359168 size(elem):2179584 bufferWidth(elem):17024 activeBufferWidth(elem):16384 bufferHeight(elem):128 padRows:0 padCols:0 Outputs: [7] numDim=6 dims=[1,1,1,128,128,128] elementType=TIDL_UnsignedShort padH/W=[0,0] batchPadH/W=[0,0] numBatchH/W=[1,1] pitch=[2179072,2179072,2179072,17024,128] dataQ=0 roundBits=13 min/maxValue=(0,0) min/maxTensorValue=(0.000000,17.734192) tensorScale=3137.193848 tensorZeroPoint=0
Convolution [7 7]
Layer 6: TIDL_PoolingLayer "/model/pool_1/MaxPool_output_0" weightsElementSizeInBits=16 kernelType=1 multiCoreMode=TIDL_NOT_MULTI_CORE strideOffsetMethod=TIDL_StrideOffsetTopLeft poolingtype=TIDL_MaxPooling numChannels=64 inDataQ=0 outDataQ=0 useCeil=0 kernelH/W=[2,2] strideH/W=[2,2] padH/W=[0,0] actParams: actType=TIDL_NoAct slopeScale=0.000000 clipMin/Max=(0.000000,0.000000) Graph compiler Info: space:DDR baseMem(bytes):8496128-11655168 size(elem):1579520 bufferWidth(elem):24672 activeBufferWidth(elem):16384 bufferHeight(elem):64 padRows:0 padCols:0 Outputs: [6] numDim=6 dims=[1,1,1,64,128,128] elementType=TIDL_UnsignedShort padH/W=[0,0] batchPadH/W=[0,0] numBatchH/W=[1,1] pitch=[1579008,1579008,1579008,24672,128] dataQ=0 roundBits=0 min/maxValue=(0,0) min/maxTensorValue=(0.000000,32.263885) tensorScale=1067.970215 tensorZeroPoint=0 ->Layer 7: TIDL_ConvolutionLayer "/model/relu_4/Relu_output_0" weightsElementSizeInBits=16 kernelType=1 multiCoreMode=TIDL_NOT_MULTI_CORE strideOffsetMethod=TIDL_StrideOffsetTopLeft convolutionType=0 numInChannels=64 numOutChannels=128 numGroups=1 kernelH/W=[3,3] strideH/W=[1,1] padH/W=[1,1] dilationH/W=[1,1] weightScale=24064.240234 biasScale=1369.173340 actParams: actType=TIDL_RelU slopeScale=0.000000 clipMin/Max=(0.000000,0.000000) Graph compiler Info: space:DDR baseMem(bytes):0-4359168 size(elem):2179584 bufferWidth(elem):17024 activeBufferWidth(elem):16384 bufferHeight(elem):128 padRows:0 padCols:0 Outputs: [7] numDim=6 dims=[1,1,1,128,128,128] elementType=TIDL_UnsignedShort padH/W=[0,0] batchPadH/W=[0,0] numBatchH/W=[1,1] pitch=[2179072,2179072,2179072,17024,128] dataQ=0 roundBits=13 min/maxValue=(0,0) min/maxTensorValue=(0.000000,17.734192) tensorScale=3137.193848 tensorZeroPoint=0
1x1x1x64x128x128
Layer 8: TIDL_ConvolutionLayer "/model/relu_5/Relu_output_0" weightsElementSizeInBits=16 kernelType=1 multiCoreMode=TIDL_NOT_MULTI_CORE strideOffsetMethod=TIDL_StrideOffsetTopLeft convolutionType=0 numInChannels=128 numOutChannels=128 numGroups=1 kernelH/W=[3,3] strideH/W=[1,1] padH/W=[1,1] dilationH/W=[1,1] weightScale=25609.787109 biasScale=2612.465332 actParams: actType=TIDL_RelU slopeScale=0.000000 clipMin/Max=(0.000000,0.000000) Graph compiler Info: space:DDR baseMem(bytes):4359168-8636416 size(elem):2138624 bufferWidth(elem):16704 activeBufferWidth(elem):16384 bufferHeight(elem):128 padRows:0 padCols:0 Outputs: [8] numDim=6 dims=[1,1,1,128,128,128] elementType=TIDL_UnsignedShort padH/W=[0,0] batchPadH/W=[0,0] numBatchH/W=[1,1] pitch=[2138112,2138112,2138112,16704,128] dataQ=0 roundBits=14 min/maxValue=(0,0) min/maxTensorValue=(0.000000,12.383701) tensorScale=4903.739258 tensorZeroPoint=0
Convolution [8 8]
Layer 7: TIDL_ConvolutionLayer "/model/relu_4/Relu_output_0" weightsElementSizeInBits=16 kernelType=1 multiCoreMode=TIDL_NOT_MULTI_CORE strideOffsetMethod=TIDL_StrideOffsetTopLeft convolutionType=0 numInChannels=64 numOutChannels=128 numGroups=1 kernelH/W=[3,3] strideH/W=[1,1] padH/W=[1,1] dilationH/W=[1,1] weightScale=24064.240234 biasScale=1369.173340 actParams: actType=TIDL_RelU slopeScale=0.000000 clipMin/Max=(0.000000,0.000000) Graph compiler Info: space:DDR baseMem(bytes):0-4359168 size(elem):2179584 bufferWidth(elem):17024 activeBufferWidth(elem):16384 bufferHeight(elem):128 padRows:0 padCols:0 Outputs: [7] numDim=6 dims=[1,1,1,128,128,128] elementType=TIDL_UnsignedShort padH/W=[0,0] batchPadH/W=[0,0] numBatchH/W=[1,1] pitch=[2179072,2179072,2179072,17024,128] dataQ=0 roundBits=13 min/maxValue=(0,0) min/maxTensorValue=(0.000000,17.734192) tensorScale=3137.193848 tensorZeroPoint=0 ->Layer 8: TIDL_ConvolutionLayer "/model/relu_5/Relu_output_0" weightsElementSizeInBits=16 kernelType=1 multiCoreMode=TIDL_NOT_MULTI_CORE strideOffsetMethod=TIDL_StrideOffsetTopLeft convolutionType=0 numInChannels=128 numOutChannels=128 numGroups=1 kernelH/W=[3,3] strideH/W=[1,1] padH/W=[1,1] dilationH/W=[1,1] weightScale=25609.787109 biasScale=2612.465332 actParams: actType=TIDL_RelU slopeScale=0.000000 clipMin/Max=(0.000000,0.000000) Graph compiler Info: space:DDR baseMem(bytes):4359168-8636416 size(elem):2138624 bufferWidth(elem):16704 activeBufferWidth(elem):16384 bufferHeight(elem):128 padRows:0 padCols:0 Outputs: [8] numDim=6 dims=[1,1,1,128,128,128] elementType=TIDL_UnsignedShort padH/W=[0,0] batchPadH/W=[0,0] numBatchH/W=[1,1] pitch=[2138112,2138112,2138112,16704,128] dataQ=0 roundBits=14 min/maxValue=(0,0) min/maxTensorValue=(0.000000,12.383701) tensorScale=4903.739258 tensorZeroPoint=0
1x1x1x128x128x128
Layer 9: TIDL_PoolingLayer "/model/pool_2/MaxPool_output_0" weightsElementSizeInBits=16 kernelType=1 multiCoreMode=TIDL_NOT_MULTI_CORE strideOffsetMethod=TIDL_StrideOffsetTopLeft poolingtype=TIDL_MaxPooling numChannels=128 inDataQ=0 outDataQ=0 useCeil=0 kernelH/W=[2,2] strideH/W=[2,2] padH/W=[0,0] actParams: actType=TIDL_NoAct slopeScale=0.000000 clipMin/Max=(0.000000,0.000000) Graph compiler Info: space:MSMC baseMem(bytes):313728-1370496 size(elem):528384 bufferWidth(elem):4128 activeBufferWidth(elem):4096 bufferHeight(elem):128 padRows:0 padCols:0 Outputs: [9] numDim=6 dims=[1,1,1,128,64,64] elementType=TIDL_UnsignedShort padH/W=[0,0] batchPadH/W=[0,0] numBatchH/W=[1,1] pitch=[528384,528384,528384,4128,64] dataQ=0 roundBits=0 min/maxValue=(0,0) min/maxTensorValue=(0.000000,13.499145) tensorScale=4903.739258 tensorZeroPoint=0
Pooling [9 9]
Layer 8: TIDL_ConvolutionLayer "/model/relu_5/Relu_output_0" weightsElementSizeInBits=16 kernelType=1 multiCoreMode=TIDL_NOT_MULTI_CORE strideOffsetMethod=TIDL_StrideOffsetTopLeft convolutionType=0 numInChannels=128 numOutChannels=128 numGroups=1 kernelH/W=[3,3] strideH/W=[1,1] padH/W=[1,1] dilationH/W=[1,1] weightScale=25609.787109 biasScale=2612.465332 actParams: actType=TIDL_RelU slopeScale=0.000000 clipMin/Max=(0.000000,0.000000) Graph compiler Info: space:DDR baseMem(bytes):4359168-8636416 size(elem):2138624 bufferWidth(elem):16704 activeBufferWidth(elem):16384 bufferHeight(elem):128 padRows:0 padCols:0 Outputs: [8] numDim=6 dims=[1,1,1,128,128,128] elementType=TIDL_UnsignedShort padH/W=[0,0] batchPadH/W=[0,0] numBatchH/W=[1,1] pitch=[2138112,2138112,2138112,16704,128] dataQ=0 roundBits=14 min/maxValue=(0,0) min/maxTensorValue=(0.000000,12.383701) tensorScale=4903.739258 tensorZeroPoint=0 ->Layer 9: TIDL_PoolingLayer "/model/pool_2/MaxPool_output_0" weightsElementSizeInBits=16 kernelType=1 multiCoreMode=TIDL_NOT_MULTI_CORE strideOffsetMethod=TIDL_StrideOffsetTopLeft poolingtype=TIDL_MaxPooling numChannels=128 inDataQ=0 outDataQ=0 useCeil=0 kernelH/W=[2,2] strideH/W=[2,2] padH/W=[0,0] actParams: actType=TIDL_NoAct slopeScale=0.000000 clipMin/Max=(0.000000,0.000000) Graph compiler Info: space:MSMC baseMem(bytes):313728-1370496 size(elem):528384 bufferWidth(elem):4128 activeBufferWidth(elem):4096 bufferHeight(elem):128 padRows:0 padCols:0 Outputs: [9] numDim=6 dims=[1,1,1,128,64,64] elementType=TIDL_UnsignedShort padH/W=[0,0] batchPadH/W=[0,0] numBatchH/W=[1,1] pitch=[528384,528384,528384,4128,64] dataQ=0 roundBits=0 min/maxValue=(0,0) min/maxTensorValue=(0.000000,13.499145) tensorScale=4903.739258 tensorZeroPoint=0
1x1x1x128x128x128
Layer 10: TIDL_ConvolutionLayer "/model/relu_6/Relu_output_0" weightsElementSizeInBits=16 kernelType=1 multiCoreMode=TIDL_NOT_MULTI_CORE strideOffsetMethod=TIDL_StrideOffsetTopLeft convolutionType=0 numInChannels=128 numOutChannels=128 numGroups=1 kernelH/W=[3,3] strideH/W=[1,1] padH/W=[1,1] dilationH/W=[1,1] weightScale=27182.947266 biasScale=2033.967407 actParams: actType=TIDL_RelU slopeScale=0.000000 clipMin/Max=(0.000000,0.000000) Graph compiler Info: space:MSMC baseMem(bytes):313728-1370496 size(elem):528384 bufferWidth(elem):4128 activeBufferWidth(elem):4096 bufferHeight(elem):128 padRows:0 padCols:0 Outputs: [10] numDim=6 dims=[1,1,1,128,64,64] elementType=TIDL_UnsignedShort padH/W=[0,0] batchPadH/W=[0,0] numBatchH/W=[1,1] pitch=[528384,528384,528384,4128,64] dataQ=0 roundBits=17 min/maxValue=(0,0) min/maxTensorValue=(0.000000,20.782539) tensorScale=1016.983704 tensorZeroPoint=0
Convolution [10 10]
Layer 9: TIDL_PoolingLayer "/model/pool_2/MaxPool_output_0" weightsElementSizeInBits=16 kernelType=1 multiCoreMode=TIDL_NOT_MULTI_CORE strideOffsetMethod=TIDL_StrideOffsetTopLeft poolingtype=TIDL_MaxPooling numChannels=128 inDataQ=0 outDataQ=0 useCeil=0 kernelH/W=[2,2] strideH/W=[2,2] padH/W=[0,0] actParams: actType=TIDL_NoAct slopeScale=0.000000 clipMin/Max=(0.000000,0.000000) Graph compiler Info: space:MSMC baseMem(bytes):313728-1370496 size(elem):528384 bufferWidth(elem):4128 activeBufferWidth(elem):4096 bufferHeight(elem):128 padRows:0 padCols:0 Outputs: [9] numDim=6 dims=[1,1,1,128,64,64] elementType=TIDL_UnsignedShort padH/W=[0,0] batchPadH/W=[0,0] numBatchH/W=[1,1] pitch=[528384,528384,528384,4128,64] dataQ=0 roundBits=0 min/maxValue=(0,0) min/maxTensorValue=(0.000000,13.499145) tensorScale=4903.739258 tensorZeroPoint=0 ->Layer 10: TIDL_ConvolutionLayer "/model/relu_6/Relu_output_0" weightsElementSizeInBits=16 kernelType=1 multiCoreMode=TIDL_NOT_MULTI_CORE strideOffsetMethod=TIDL_StrideOffsetTopLeft convolutionType=0 numInChannels=128 numOutChannels=128 numGroups=1 kernelH/W=[3,3] strideH/W=[1,1] padH/W=[1,1] dilationH/W=[1,1] weightScale=27182.947266 biasScale=2033.967407 actParams: actType=TIDL_RelU slopeScale=0.000000 clipMin/Max=(0.000000,0.000000) Graph compiler Info: space:MSMC baseMem(bytes):313728-1370496 size(elem):528384 bufferWidth(elem):4128 activeBufferWidth(elem):4096 bufferHeight(elem):128 padRows:0 padCols:0 Outputs: [10] numDim=6 dims=[1,1,1,128,64,64] elementType=TIDL_UnsignedShort padH/W=[0,0] batchPadH/W=[0,0] numBatchH/W=[1,1] pitch=[528384,528384,528384,4128,64] dataQ=0 roundBits=17 min/maxValue=(0,0) min/maxTensorValue=(0.000000,20.782539) tensorScale=1016.983704 tensorZeroPoint=0
1x1x1x128x64x64
Layer 11: TIDL_ConvolutionLayer "/model/relu_7/Relu_output_0" weightsElementSizeInBits=16 kernelType=1 multiCoreMode=TIDL_NOT_MULTI_CORE strideOffsetMethod=TIDL_StrideOffsetTopLeft convolutionType=0 numInChannels=128 numOutChannels=128 numGroups=1 kernelH/W=[3,3] strideH/W=[1,1] padH/W=[1,1] dilationH/W=[1,1] weightScale=47560.914062 biasScale=2141.770264 actParams: actType=TIDL_RelU slopeScale=0.000000 clipMin/Max=(0.000000,0.000000) Graph compiler Info: space:MSMC baseMem(bytes):0-1056768 size(elem):528384 bufferWidth(elem):4128 activeBufferWidth(elem):4096 bufferHeight(elem):128 padRows:0 padCols:0 Outputs: [11] numDim=6 dims=[1,1,1,128,64,64] elementType=TIDL_UnsignedShort padH/W=[0,0] batchPadH/W=[0,0] numBatchH/W=[1,1] pitch=[528384,528384,528384,4128,64] dataQ=0 roundBits=15 min/maxValue=(0,0) min/maxTensorValue=(0.000000,20.034100) tensorScale=1476.094849 tensorZeroPoint=0
Convolution [11 11]
Layer 10: TIDL_ConvolutionLayer "/model/relu_6/Relu_output_0" weightsElementSizeInBits=16 kernelType=1 multiCoreMode=TIDL_NOT_MULTI_CORE strideOffsetMethod=TIDL_StrideOffsetTopLeft convolutionType=0 numInChannels=128 numOutChannels=128 numGroups=1 kernelH/W=[3,3] strideH/W=[1,1] padH/W=[1,1] dilationH/W=[1,1] weightScale=27182.947266 biasScale=2033.967407 actParams: actType=TIDL_RelU slopeScale=0.000000 clipMin/Max=(0.000000,0.000000) Graph compiler Info: space:MSMC baseMem(bytes):313728-1370496 size(elem):528384 bufferWidth(elem):4128 activeBufferWidth(elem):4096 bufferHeight(elem):128 padRows:0 padCols:0 Outputs: [10] numDim=6 dims=[1,1,1,128,64,64] elementType=TIDL_UnsignedShort padH/W=[0,0] batchPadH/W=[0,0] numBatchH/W=[1,1] pitch=[528384,528384,528384,4128,64] dataQ=0 roundBits=17 min/maxValue=(0,0) min/maxTensorValue=(0.000000,20.782539) tensorScale=1016.983704 tensorZeroPoint=0 ->Layer 11: TIDL_ConvolutionLayer "/model/relu_7/Relu_output_0" weightsElementSizeInBits=16 kernelType=1 multiCoreMode=TIDL_NOT_MULTI_CORE strideOffsetMethod=TIDL_StrideOffsetTopLeft convolutionType=0 numInChannels=128 numOutChannels=128 numGroups=1 kernelH/W=[3,3] strideH/W=[1,1] padH/W=[1,1] dilationH/W=[1,1] weightScale=47560.914062 biasScale=2141.770264 actParams: actType=TIDL_RelU slopeScale=0.000000 clipMin/Max=(0.000000,0.000000) Graph compiler Info: space:MSMC baseMem(bytes):0-1056768 size(elem):528384 bufferWidth(elem):4128 activeBufferWidth(elem):4096 bufferHeight(elem):128 padRows:0 padCols:0 Outputs: [11] numDim=6 dims=[1,1,1,128,64,64] elementType=TIDL_UnsignedShort padH/W=[0,0] batchPadH/W=[0,0] numBatchH/W=[1,1] pitch=[528384,528384,528384,4128,64] dataQ=0 roundBits=15 min/maxValue=(0,0) min/maxTensorValue=(0.000000,20.034100) tensorScale=1476.094849 tensorZeroPoint=0
1x1x1x128x64x64
Layer 12: TIDL_ConvolutionLayer "/model/relu_9/Relu_output_0" weightsElementSizeInBits=16 kernelType=1 multiCoreMode=TIDL_NOT_MULTI_CORE strideOffsetMethod=TIDL_StrideOffsetTopLeft convolutionType=0 numInChannels=128 numOutChannels=256 numGroups=1 kernelH/W=[3,3] strideH/W=[1,1] padH/W=[1,1] dilationH/W=[1,1] weightScale=90728.343750 biasScale=3051.923096 actParams: actType=TIDL_RelU slopeScale=0.000000 clipMin/Max=(0.000000,0.000000) Graph compiler Info: space:DDR baseMem(bytes):0-2343936 size(elem):1171968 bufferWidth(elem):4576 activeBufferWidth(elem):4096 bufferHeight(elem):256 padRows:0 padCols:0 Outputs: [12] numDim=6 dims=[1,1,1,256,64,64] elementType=TIDL_UnsignedShort padH/W=[0,0] batchPadH/W=[0,0] numBatchH/W=[1,1] pitch=[1171456,1171456,1171456,4576,64] dataQ=0 roundBits=16 min/maxValue=(0,0) min/maxTensorValue=(0.000000,16.808695) tensorScale=2043.512573 tensorZeroPoint=0
Convolution [12 12]
Layer 11: TIDL_ConvolutionLayer "/model/relu_7/Relu_output_0" weightsElementSizeInBits=16 kernelType=1 multiCoreMode=TIDL_NOT_MULTI_CORE strideOffsetMethod=TIDL_StrideOffsetTopLeft convolutionType=0 numInChannels=128 numOutChannels=128 numGroups=1 kernelH/W=[3,3] strideH/W=[1,1] padH/W=[1,1] dilationH/W=[1,1] weightScale=47560.914062 biasScale=2141.770264 actParams: actType=TIDL_RelU slopeScale=0.000000 clipMin/Max=(0.000000,0.000000) Graph compiler Info: space:MSMC baseMem(bytes):0-1056768 size(elem):528384 bufferWidth(elem):4128 activeBufferWidth(elem):4096 bufferHeight(elem):128 padRows:0 padCols:0 Outputs: [11] numDim=6 dims=[1,1,1,128,64,64] elementType=TIDL_UnsignedShort padH/W=[0,0] batchPadH/W=[0,0] numBatchH/W=[1,1] pitch=[528384,528384,528384,4128,64] dataQ=0 roundBits=15 min/maxValue=(0,0) min/maxTensorValue=(0.000000,20.034100) tensorScale=1476.094849 tensorZeroPoint=0 ->Layer 12: TIDL_ConvolutionLayer "/model/relu_9/Relu_output_0" weightsElementSizeInBits=16 kernelType=1 multiCoreMode=TIDL_NOT_MULTI_CORE strideOffsetMethod=TIDL_StrideOffsetTopLeft convolutionType=0 numInChannels=128 numOutChannels=256 numGroups=1 kernelH/W=[3,3] strideH/W=[1,1] padH/W=[1,1] dilationH/W=[1,1] weightScale=90728.343750 biasScale=3051.923096 actParams: actType=TIDL_RelU slopeScale=0.000000 clipMin/Max=(0.000000,0.000000) Graph compiler Info: space:DDR baseMem(bytes):0-2343936 size(elem):1171968 bufferWidth(elem):4576 activeBufferWidth(elem):4096 bufferHeight(elem):256 padRows:0 padCols:0 Outputs: [12] numDim=6 dims=[1,1,1,256,64,64] elementType=TIDL_UnsignedShort padH/W=[0,0] batchPadH/W=[0,0] numBatchH/W=[1,1] pitch=[1171456,1171456,1171456,4576,64] dataQ=0 roundBits=16 min/maxValue=(0,0) min/maxTensorValue=(0.000000,16.808695) tensorScale=2043.512573 tensorZeroPoint=0
1x1x1x128x64x64
Layer 13: TIDL_ConvolutionLayer "/model/relu_8/Relu_output_0" weightsElementSizeInBits=16 kernelType=1 multiCoreMode=TIDL_NOT_MULTI_CORE strideOffsetMethod=TIDL_StrideOffsetTopLeft convolutionType=0 numInChannels=128 numOutChannels=256 numGroups=1 kernelH/W=[3,3] strideH/W=[1,1] padH/W=[1,1] dilationH/W=[1,1] weightScale=265958.187500 biasScale=6229.770020 actParams: actType=TIDL_RelU slopeScale=0.000000 clipMin/Max=(0.000000,0.000000) Graph compiler Info: space:DDR baseMem(bytes):0-2343936 size(elem):1171968 bufferWidth(elem):4576 activeBufferWidth(elem):4096 bufferHeight(elem):256 padRows:0 padCols:0 Outputs: [13] numDim=6 dims=[1,1,1,256,64,64] elementType=TIDL_UnsignedShort padH/W=[0,0] batchPadH/W=[0,0] numBatchH/W=[1,1] pitch=[1171456,1171456,1171456,4576,64] dataQ=0 roundBits=15 min/maxValue=(0,0) min/maxTensorValue=(0.000000,2.870904) tensorScale=11980.576172 tensorZeroPoint=0
Convolution [14 13]
Layer 11: TIDL_ConvolutionLayer "/model/relu_7/Relu_output_0" weightsElementSizeInBits=16 kernelType=1 multiCoreMode=TIDL_NOT_MULTI_CORE strideOffsetMethod=TIDL_StrideOffsetTopLeft convolutionType=0 numInChannels=128 numOutChannels=128 numGroups=1 kernelH/W=[3,3] strideH/W=[1,1] padH/W=[1,1] dilationH/W=[1,1] weightScale=47560.914062 biasScale=2141.770264 actParams: actType=TIDL_RelU slopeScale=0.000000 clipMin/Max=(0.000000,0.000000) Graph compiler Info: space:MSMC baseMem(bytes):0-1056768 size(elem):528384 bufferWidth(elem):4128 activeBufferWidth(elem):4096 bufferHeight(elem):128 padRows:0 padCols:0 Outputs: [11] numDim=6 dims=[1,1,1,128,64,64] elementType=TIDL_UnsignedShort padH/W=[0,0] batchPadH/W=[0,0] numBatchH/W=[1,1] pitch=[528384,528384,528384,4128,64] dataQ=0 roundBits=15 min/maxValue=(0,0) min/maxTensorValue=(0.000000,20.034100) tensorScale=1476.094849 tensorZeroPoint=0 ->Layer 13: TIDL_ConvolutionLayer "/model/relu_8/Relu_output_0" weightsElementSizeInBits=16 kernelType=1 multiCoreMode=TIDL_NOT_MULTI_CORE strideOffsetMethod=TIDL_StrideOffsetTopLeft convolutionType=0 numInChannels=128 numOutChannels=256 numGroups=1 kernelH/W=[3,3] strideH/W=[1,1] padH/W=[1,1] dilationH/W=[1,1] weightScale=265958.187500 biasScale=6229.770020 actParams: actType=TIDL_RelU slopeScale=0.000000 clipMin/Max=(0.000000,0.000000) Graph compiler Info: space:DDR baseMem(bytes):0-2343936 size(elem):1171968 bufferWidth(elem):4576 activeBufferWidth(elem):4096 bufferHeight(elem):256 padRows:0 padCols:0 Outputs: [13] numDim=6 dims=[1,1,1,256,64,64] elementType=TIDL_UnsignedShort padH/W=[0,0] batchPadH/W=[0,0] numBatchH/W=[1,1] pitch=[1171456,1171456,1171456,4576,64] dataQ=0 roundBits=15 min/maxValue=(0,0) min/maxTensorValue=(0.000000,2.870904) tensorScale=11980.576172 tensorZeroPoint=0
1x1x1x128x64x64
Layer 14: TIDL_ConvolutionLayer "Dbout" weightsElementSizeInBits=16 kernelType=1 multiCoreMode=TIDL_NOT_MULTI_CORE strideOffsetMethod=TIDL_StrideOffsetTopLeft convolutionType=0 numInChannels=256 numOutChannels=256 numGroups=1 kernelH/W=[1,1] strideH/W=[1,1] padH/W=[0,0] dilationH/W=[1,1] weightScale=2199.116211 biasScale=3973.551270 actParams: actType=TIDL_NoAct slopeScale=0.000000 clipMin/Max=(0.000000,0.000000) Graph compiler Info: space:DDR baseMem(bytes):2343936-4638720 size(elem):1147392 bufferWidth(elem):4480 activeBufferWidth(elem):4096 bufferHeight(elem):256 padRows:0 padCols:0 Outputs: [14] numDim=6 dims=[1,1,1,256,64,64] elementType=TIDL_SignedShort padH/W=[0,0] batchPadH/W=[0,0] numBatchH/W=[1,1] pitch=[1146880,1146880,1146880,4480,64] dataQ=0 roundBits=16 min/maxValue=(0,0) min/maxTensorValue=(-388.494781,381.685913) tensorScale=68.571800 tensorZeroPoint=0
Convolution [13 14]
Layer 12: TIDL_ConvolutionLayer "/model/relu_9/Relu_output_0" weightsElementSizeInBits=16 kernelType=1 multiCoreMode=TIDL_NOT_MULTI_CORE strideOffsetMethod=TIDL_StrideOffsetTopLeft convolutionType=0 numInChannels=128 numOutChannels=256 numGroups=1 kernelH/W=[3,3] strideH/W=[1,1] padH/W=[1,1] dilationH/W=[1,1] weightScale=90728.343750 biasScale=3051.923096 actParams: actType=TIDL_RelU slopeScale=0.000000 clipMin/Max=(0.000000,0.000000) Graph compiler Info: space:DDR baseMem(bytes):0-2343936 size(elem):1171968 bufferWidth(elem):4576 activeBufferWidth(elem):4096 bufferHeight(elem):256 padRows:0 padCols:0 Outputs: [12] numDim=6 dims=[1,1,1,256,64,64] elementType=TIDL_UnsignedShort padH/W=[0,0] batchPadH/W=[0,0] numBatchH/W=[1,1] pitch=[1171456,1171456,1171456,4576,64] dataQ=0 roundBits=16 min/maxValue=(0,0) min/maxTensorValue=(0.000000,16.808695) tensorScale=2043.512573 tensorZeroPoint=0 ->Layer 14: TIDL_ConvolutionLayer "Dbout" weightsElementSizeInBits=16 kernelType=1 multiCoreMode=TIDL_NOT_MULTI_CORE strideOffsetMethod=TIDL_StrideOffsetTopLeft convolutionType=0 numInChannels=256 numOutChannels=256 numGroups=1 kernelH/W=[1,1] strideH/W=[1,1] padH/W=[0,0] dilationH/W=[1,1] weightScale=2199.116211 biasScale=3973.551270 actParams: actType=TIDL_NoAct slopeScale=0.000000 clipMin/Max=(0.000000,0.000000) Graph compiler Info: space:DDR baseMem(bytes):2343936-4638720 size(elem):1147392 bufferWidth(elem):4480 activeBufferWidth(elem):4096 bufferHeight(elem):256 padRows:0 padCols:0 Outputs: [14] numDim=6 dims=[1,1,1,256,64,64] elementType=TIDL_SignedShort padH/W=[0,0] batchPadH/W=[0,0] numBatchH/W=[1,1] pitch=[1146880,1146880,1146880,4480,64] dataQ=0 roundBits=16 min/maxValue=(0,0) min/maxTensorValue=(-388.494781,381.685913) tensorScale=68.571800 tensorZeroPoint=0
1x1x1x256x64x64
Layer 16: TIDL_DataLayer "image" weightsElementSizeInBits=16 kernelType=1 multiCoreMode=TIDL_NOT_MULTI_CORE strideOffsetMethod=TIDL_StrideOffsetTopLeft numChannels=0 dataQ=0 actParams: actType=TIDL_NoAct slopeScale=0.000000 clipMin/Max=(0.000000,0.000000) Graph compiler Info: space:DDR baseMem(bytes):33854464-34379776 size(elem):525312 bufferWidth(elem):524288 activeBufferWidth(elem):0 bufferHeight(elem):0 padRows:0 padCols:0
Data [16 0]
Layer 14: TIDL_ConvolutionLayer "Dbout" weightsElementSizeInBits=16 kernelType=1 multiCoreMode=TIDL_NOT_MULTI_CORE strideOffsetMethod=TIDL_StrideOffsetTopLeft convolutionType=0 numInChannels=256 numOutChannels=256 numGroups=1 kernelH/W=[1,1] strideH/W=[1,1] padH/W=[0,0] dilationH/W=[1,1] weightScale=2199.116211 biasScale=3973.551270 actParams: actType=TIDL_NoAct slopeScale=0.000000 clipMin/Max=(0.000000,0.000000) Graph compiler Info: space:DDR baseMem(bytes):2343936-4638720 size(elem):1147392 bufferWidth(elem):4480 activeBufferWidth(elem):4096 bufferHeight(elem):256 padRows:0 padCols:0 Outputs: [14] numDim=6 dims=[1,1,1,256,64,64] elementType=TIDL_SignedShort padH/W=[0,0] batchPadH/W=[0,0] numBatchH/W=[1,1] pitch=[1146880,1146880,1146880,4480,64] dataQ=0 roundBits=16 min/maxValue=(0,0) min/maxTensorValue=(-388.494781,381.685913) tensorScale=68.571800 tensorZeroPoint=0 ->Layer 16: TIDL_DataLayer "image" weightsElementSizeInBits=16 kernelType=1 multiCoreMode=TIDL_NOT_MULTI_CORE strideOffsetMethod=TIDL_StrideOffsetTopLeft numChannels=0 dataQ=0 actParams: actType=TIDL_NoAct slopeScale=0.000000 clipMin/Max=(0.000000,0.000000) Graph compiler Info: space:DDR baseMem(bytes):33854464-34379776 size(elem):525312 bufferWidth(elem):524288 activeBufferWidth(elem):0 bufferHeight(elem):0 padRows:0 padCols:0
1x1x1x256x64x64
Layer 15: TIDL_ConvolutionLayer "Pbout" weightsElementSizeInBits=16 kernelType=1 multiCoreMode=TIDL_NOT_MULTI_CORE strideOffsetMethod=TIDL_StrideOffsetTopLeft convolutionType=0 numInChannels=256 numOutChannels=65 numGroups=1 kernelH/W=[1,1] strideH/W=[1,1] padH/W=[0,0] dilationH/W=[1,1] weightScale=3471.911133 biasScale=19168.011719 actParams: actType=TIDL_NoAct slopeScale=0.000000 clipMin/Max=(0.000000,0.000000) Graph compiler Info: space:DDR baseMem(bytes):2343936-2927360 size(elem):291712 bufferWidth(elem):4480 activeBufferWidth(elem):4096 bufferHeight(elem):65 padRows:0 padCols:0 Outputs: [15] numDim=6 dims=[1,1,1,65,64,64] elementType=TIDL_SignedShort padH/W=[0,0] batchPadH/W=[0,0] numBatchH/W=[1,1] pitch=[291200,291200,291200,4480,64] dataQ=0 roundBits=16 min/maxValue=(0,0) min/maxTensorValue=(-39.648476,11.018888) tensorScale=634.696899 tensorZeroPoint=0
Convolution [15 15]
Layer 13: TIDL_ConvolutionLayer "/model/relu_8/Relu_output_0" weightsElementSizeInBits=16 kernelType=1 multiCoreMode=TIDL_NOT_MULTI_CORE strideOffsetMethod=TIDL_StrideOffsetTopLeft convolutionType=0 numInChannels=128 numOutChannels=256 numGroups=1 kernelH/W=[3,3] strideH/W=[1,1] padH/W=[1,1] dilationH/W=[1,1] weightScale=265958.187500 biasScale=6229.770020 actParams: actType=TIDL_RelU slopeScale=0.000000 clipMin/Max=(0.000000,0.000000) Graph compiler Info: space:DDR baseMem(bytes):0-2343936 size(elem):1171968 bufferWidth(elem):4576 activeBufferWidth(elem):4096 bufferHeight(elem):256 padRows:0 padCols:0 Outputs: [13] numDim=6 dims=[1,1,1,256,64,64] elementType=TIDL_UnsignedShort padH/W=[0,0] batchPadH/W=[0,0] numBatchH/W=[1,1] pitch=[1171456,1171456,1171456,4576,64] dataQ=0 roundBits=15 min/maxValue=(0,0) min/maxTensorValue=(0.000000,2.870904) tensorScale=11980.576172 tensorZeroPoint=0 ->Layer 15: TIDL_ConvolutionLayer "Pbout" weightsElementSizeInBits=16 kernelType=1 multiCoreMode=TIDL_NOT_MULTI_CORE strideOffsetMethod=TIDL_StrideOffsetTopLeft convolutionType=0 numInChannels=256 numOutChannels=65 numGroups=1 kernelH/W=[1,1] strideH/W=[1,1] padH/W=[0,0] dilationH/W=[1,1] weightScale=3471.911133 biasScale=19168.011719 actParams: actType=TIDL_NoAct slopeScale=0.000000 clipMin/Max=(0.000000,0.000000) Graph compiler Info: space:DDR baseMem(bytes):2343936-2927360 size(elem):291712 bufferWidth(elem):4480 activeBufferWidth(elem):4096 bufferHeight(elem):65 padRows:0 padCols:0 Outputs: [15] numDim=6 dims=[1,1,1,65,64,64] elementType=TIDL_SignedShort padH/W=[0,0] batchPadH/W=[0,0] numBatchH/W=[1,1] pitch=[291200,291200,291200,4480,64] dataQ=0 roundBits=16 min/maxValue=(0,0) min/maxTensorValue=(-39.648476,11.018888) tensorScale=634.696899 tensorZeroPoint=0
1x1x1x256x64x64
Layer 17: TIDL_DataLayer "image" weightsElementSizeInBits=16 kernelType=1 multiCoreMode=TIDL_NOT_MULTI_CORE strideOffsetMethod=TIDL_StrideOffsetTopLeft numChannels=0 dataQ=0 actParams: actType=TIDL_NoAct slopeScale=0.000000 clipMin/Max=(0.000000,0.000000) Graph compiler Info: space:DDR baseMem(bytes):33854464-34379776 size(elem):525312 bufferWidth(elem):524288 activeBufferWidth(elem):0 bufferHeight(elem):0 padRows:0 padCols:0
Data [17 0]
Layer 15: TIDL_ConvolutionLayer "Pbout" weightsElementSizeInBits=16 kernelType=1 multiCoreMode=TIDL_NOT_MULTI_CORE strideOffsetMethod=TIDL_StrideOffsetTopLeft convolutionType=0 numInChannels=256 numOutChannels=65 numGroups=1 kernelH/W=[1,1] strideH/W=[1,1] padH/W=[0,0] dilationH/W=[1,1] weightScale=3471.911133 biasScale=19168.011719 actParams: actType=TIDL_NoAct slopeScale=0.000000 clipMin/Max=(0.000000,0.000000) Graph compiler Info: space:DDR baseMem(bytes):2343936-2927360 size(elem):291712 bufferWidth(elem):4480 activeBufferWidth(elem):4096 bufferHeight(elem):65 padRows:0 padCols:0 Outputs: [15] numDim=6 dims=[1,1,1,65,64,64] elementType=TIDL_SignedShort padH/W=[0,0] batchPadH/W=[0,0] numBatchH/W=[1,1] pitch=[291200,291200,291200,4480,64] dataQ=0 roundBits=16 min/maxValue=(0,0) min/maxTensorValue=(-39.648476,11.018888) tensorScale=634.696899 tensorZeroPoint=0 ->Layer 17: TIDL_DataLayer "image" weightsElementSizeInBits=16 kernelType=1 multiCoreMode=TIDL_NOT_MULTI_CORE strideOffsetMethod=TIDL_StrideOffsetTopLeft numChannels=0 dataQ=0 actParams: actType=TIDL_NoAct slopeScale=0.000000 clipMin/Max=(0.000000,0.000000) Graph compiler Info: space:DDR baseMem(bytes):33854464-34379776 size(elem):525312 bufferWidth(elem):524288 activeBufferWidth(elem):0 bufferHeight(elem):0 padRows:0 padCols:0
1x1x1x65x64x64
Search
✕
×
NODE PROPERTIES
layerNum
type
name
INPUTS
OUTPUTS
ATTRIBUTES