==25751== Profiling application: trtexec --loadEngine=best_qat_24_int8_b1_sparse.trt
==25751== Profiling result:
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
 GPU activities:   24.15%  771.22ms     12992  59.361us  2.8160us  5.4506ms  generatedNativePointwise
                   23.49%  750.07ms      7504  99.956us  32.579us  2.3475ms  trt_volta_fp32_i8816cudnn_int8_128x128_ldg16_relu_small_nt_v1
                   13.11%  418.58ms      2800  149.49us  86.535us  731.29us  trt_volta_fp32_i8816cudnn_int8_256x64_ldg16_relu_singleBuffer_small_nt_v1
                    7.12%  227.37ms      3920  58.001us  19.490us  1.7740ms  void CUTENSOR_NAMESPACE::permutationKernelPLC3<CUTENSOR_NAMESPACE::VectorWrite2DTensorView<unsigned char=0, unsigned char=1, bool=0, unsigned int=4, float, char, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::VectorRead2DTensorView<unsigned char=1, unsigned char=0, bool=0, unsigned int=4, float, float, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::ThreadLevelElementwise<CUTENSOR_NAMESPACE::ElementwiseConfig2DCommonCase<CUTENSOR_NAMESPACE::GeneralUnarySmall<float>, CUTENSOR_NAMESPACE::GeneralBinary<float>, int=2, int=32, int=64, int=128, char=4, bool=0, bool=0, bool=1, bool=0, bool=0>, float>, CUTENSOR_NAMESPACE::ElementwiseRuntimePLC3<float, float, char, float>::Params>(unsigned int=4)
                    6.66%  212.65ms      2576  82.552us  41.699us  1.1712ms  void CUTENSOR_NAMESPACE::permutationKernelPLC3<CUTENSOR_NAMESPACE::VectorWrite2DTensorView<unsigned char=0, unsigned char=1, bool=0, unsigned int=4, float, float, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::VectorRead2DTensorView<unsigned char=1, unsigned char=0, bool=0, unsigned int=4, float, float, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::ThreadLevelElementwise<CUTENSOR_NAMESPACE::ElementwiseConfig2DCommonCase<CUTENSOR_NAMESPACE::GeneralUnarySmall<float>, CUTENSOR_NAMESPACE::GeneralBinary<float>, int=2, int=64, int=64, int=256, char=4, bool=0, bool=1, bool=1, bool=1, bool=0>, float>, CUTENSOR_NAMESPACE::ElementwiseRuntimePLC3<float, float, float, float>::Params>(unsigned int=4)
                    5.23%  167.15ms      1008  165.83us  103.40us  1.0296ms  trt_volta_fp32_icudnn_int8x4_128x64_relu_interior_nn_v1
                    5.22%  166.60ms       560  297.50us  127.11us  5.3590ms  trt_volta_fp32_icudnn_int8x4_128x64_relu_small_nn_v1
                    2.36%  75.400ms       224  336.61us  236.05us  1.0020ms  void cudnn::ops::pooling_fw_4d_kernel<float, float, cudnn::maxpooling_func<float, cudnnNanPropagation_t=0>, cudnnPoolingMode_t=0, bool=0>(cudnnTensorStruct, float const *, cudnn::ops::pooling_fw_4d_kernel<float, float, cudnn::maxpooling_func<float, cudnnNanPropagation_t=0>, cudnnPoolingMode_t=0, bool=0>, cudnnTensorStruct*, cudnnPoolingStruct, float, cudnnPoolingStruct, int, cudnn::reduced_divisor, float)
                    2.10%  66.916ms       112  597.46us  532.20us  1.7369ms  trt_volta_fp32_i8816cudnn_int8_256x64_ldg16_relu_singleBuffer_large_nt_v1
                    1.30%  41.653ms       896  46.487us  32.419us  144.24us  void cuSliceLayer::naiveSlice<float, cuSliceLayer::Mode, int=4>(cuSliceLayer::LaunchParams<float>)
                    1.23%  39.306ms       784  50.134us  17.089us  149.71us  cuInt8::nchwToNcqhw4(char const *, unsigned int*, int, int, int, int, int, int, float const *, float const *)
                    1.12%  35.755ms       124  288.35us     928ns  19.597ms  [CUDA memcpy HtoD]
                    0.88%  28.134ms       224  125.60us  82.375us  238.58us  void cuResizeLayer::ResizeNearestGenericKernel<float, int=2, int=128, nvinfer1::ResizeRoundMode>(float*, cuResizeLayer::ResizeNearestGenericKernel<float, int=2, int=128, nvinfer1::ResizeRoundMode> const *, cuResizeLayer::LaunchParams)
                    0.80%  25.536ms       112  228.00us  221.68us  336.09us  void CUTENSOR_NAMESPACE::permutationKernelPLC3<CUTENSOR_NAMESPACE::VectorWrite2DTensorView<unsigned char=0, unsigned char=1, bool=0, unsigned int=16, float, char, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::VectorRead2DTensorView<unsigned char=1, unsigned char=0, bool=0, unsigned int=16, float, char, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::ThreadLevelElementwise<CUTENSOR_NAMESPACE::ElementwiseConfig2DCommonCase<CUTENSOR_NAMESPACE::GeneralUnarySmall<float>, CUTENSOR_NAMESPACE::GeneralBinary<float>, int=2, int=32, int=64, int=128, char=16, bool=0, bool=0, bool=1, bool=0, bool=1>, float>, CUTENSOR_NAMESPACE::ElementwiseRuntimePLC3<char, float, char, float>::Params>(unsigned int=16)
                    0.71%  22.820ms       672  33.958us  12.609us  2.0256ms  cuInt8::nc32hw32ToNcqhw4_block(char4 const *, char4*, int, int, int, int, float const *, float const *)
                    0.55%  17.408ms       672  25.904us  14.657us  51.620us  void cuFillLayer::fill<int>(cuFillLayer::KernelArgs<int>)
                    0.46%  14.674ms       112  131.02us  121.45us  178.93us  trt_volta_scudnn_128x32_relu_interior_nn_v1
                    0.39%  12.403ms       336  36.912us  33.699us  110.89us  void CUTENSOR_NAMESPACE::permutationKernelPLC3<CUTENSOR_NAMESPACE::VectorWrite2DTensorView<unsigned char=0, unsigned char=1, bool=0, unsigned int=4, float, char, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::VectorRead2DTensorView<unsigned char=1, unsigned char=0, bool=0, unsigned int=4, float, char, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::ThreadLevelElementwise<CUTENSOR_NAMESPACE::ElementwiseConfig2DSmallMode0<CUTENSOR_NAMESPACE::GeneralUnarySmall<float>, CUTENSOR_NAMESPACE::GeneralBinary<float>, int=2, int=4, int=256, int=64, char=4, bool=0, bool=0, bool=1, bool=0, bool=1>, float>, CUTENSOR_NAMESPACE::ElementwiseRuntimePLC3<char, float, char, float>::Params>(unsigned int=4)
                    0.36%  11.544ms      1344  8.5890us  1.9840us  28.194us  void genericReformat::copyPackedKernel<float, float, bool=1, bool=1, genericReformat::IdentityCoordMapper<int=4>, int=4>(unsigned int, unsigned int, void const *, genericReformat::ArrayN<genericReformat::IdentityCoordMapper<int=4>>, genericReformat::ArrayNWithReducedDivisors<genericReformat::IdentityCoordMapper<int=4>>, genericReformat::ArrayN, int, int, int, float const *, void*, genericReformat::ArrayN, genericReformat::ArrayNWithReducedDivisors, genericReformat::ArrayNWithReducedDivisors, genericReformat::ArrayN, int, int, int, float const , int=4)
                    0.36%  11.389ms       112  101.68us  95.623us  141.16us  void fused::fusedConvolutionReluKernel<fused::SrcChwcPtr_FltTex_Reader<float, int=1, int=1, int=4, int=1>, fused::KpqkPtrWriter<float, int=1, int=1, int=1>, float, float, int=4, int=7, int=2, int=1, int=1, int=1, int=1>(fused::ConvolutionParams<floatSrcType, int=4, int=1Type>)
                    0.27%  8.7035ms       458  19.003us  1.1200us  81.927us  [CUDA memcpy DtoH]
                    0.26%  8.4329ms       216  39.041us  6.8480us  1.3463ms  [CUDA memcpy DtoD]
                    0.24%  7.7303ms      1344  5.7510us  1.8560us  19.329us  void cuSliceLayer::naiveSlice<float, cuSliceLayer::Mode, int=5>(cuSliceLayer::LaunchParams<float>)
                    0.22%  7.1360ms       336  21.238us  17.441us  57.540us  void CUTENSOR_NAMESPACE::permutationKernelPLC3<CUTENSOR_NAMESPACE::VectorWrite2DTensorView<unsigned char=0, unsigned char=1, bool=1, unsigned int=4, float, char, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::VectorRead2DTensorView<unsigned char=0, unsigned char=1, bool=1, unsigned int=4, float, float, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::ThreadLevelElementwise<CUTENSOR_NAMESPACE::ElementwiseConfig1DCommonCase<CUTENSOR_NAMESPACE::GeneralUnarySmall<float>, CUTENSOR_NAMESPACE::GeneralBinary<float>, int=1, int=2048, int=1, int=128, char=4, bool=1, bool=0, bool=1, bool=0, bool=0>, float>, CUTENSOR_NAMESPACE::ElementwiseRuntimePLC3<float, float, char, float>::Params>(unsigned int=4)
                    0.22%  7.0220ms       112  62.696us  60.293us  88.711us  void nvinfer1::rt::cuda::poolCHW_PQT<int=5, int=5, int=1, int=1, int=1, int=1, int=1024, int=1, int=1, bool=0, nvinfer1::rt::cuda::ITiledPooling::PoolingMode, bool=0>(nvinfer1::rt::cuda::TiledPoolingParams, int)
                    0.22%  6.9970ms       112  62.473us  57.477us  132.94us  cuInt8::ncqhw4ToNc32hw32(char4 const *, char4*, nvinfer1::rt::ReducedDivisor, int, nvinfer1::rt::ReducedDivisor, nvinfer1::rt::ReducedDivisor, int, int, float const *, float const *)
                    0.18%  5.8340ms       112  52.088us  48.355us  72.070us  void fused::fusedConvolutionReluKernel<fused::SrcChwcPtr_FltTex_Reader<float, int=1, int=1, int=4, int=1>, fused::KpqkPtrWriter<float, int=1, int=1, int=1>, float, float, int=4, int=7, int=8, int=1, int=1, int=1, int=1>(fused::ConvolutionParams<floatSrcType, int=4, int=1Type>)
                    0.16%  5.1520ms       336  15.333us  5.6320us  37.251us  void genericReformat::copyPackedKernel<float, float, bool=1, bool=1, genericReformat::ArrayN<int=5>, int=5>(unsigned int, unsigned int, void const *, genericReformat::ArrayN<genericReformat::ArrayN<int=5>>, genericReformat::ArrayNWithReducedDivisors<genericReformat::ArrayN<int=5>>, void const *, int, int, int, float const *, void*, void const *, genericReformat::ArrayNWithReducedDivisors, genericReformat::ArrayNWithReducedDivisors, void const *, int, int, int, float const , int=5)
                    0.15%  4.7297ms       112  42.229us  40.100us  121.32us  void CUTENSOR_NAMESPACE::permutationKernelPLC3<CUTENSOR_NAMESPACE::VectorWrite2DTensorView<unsigned char=0, unsigned char=1, bool=0, unsigned int=4, float, char, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::VectorRead2DTensorView<unsigned char=0, unsigned char=1, bool=0, unsigned int=4, float, char, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::ThreadLevelElementwise<CUTENSOR_NAMESPACE::ElementwiseConfig2DSmallMode0<CUTENSOR_NAMESPACE::GeneralUnarySmall<float>, CUTENSOR_NAMESPACE::GeneralBinary<float>, int=2, int=4, int=256, int=64, char=4, bool=1, bool=0, bool=1, bool=0, bool=1>, float>, CUTENSOR_NAMESPACE::ElementwiseRuntimePLC3<char, float, char, float>::Params>(unsigned int=4)
                    0.13%  4.1140ms       112  36.732us  34.659us  101.96us  void CUTENSOR_NAMESPACE::permutationKernelPLC3<CUTENSOR_NAMESPACE::VectorWrite2DTensorView<unsigned char=0, unsigned char=1, bool=0, unsigned int=4, float, char, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::VectorRead2DTensorView<unsigned char=1, unsigned char=0, bool=0, unsigned int=4, float, float, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::ThreadLevelElementwise<CUTENSOR_NAMESPACE::ElementwiseConfig2DSmallMode0<CUTENSOR_NAMESPACE::GeneralUnarySmall<float>, CUTENSOR_NAMESPACE::GeneralBinary<float>, int=2, int=4, int=256, int=128, char=4, bool=0, bool=0, bool=1, bool=0, bool=1>, float>, CUTENSOR_NAMESPACE::ElementwiseRuntimePLC3<float, float, char, float>::Params>(unsigned int=4)
                    0.13%  4.0091ms        10  400.91us  1.3440us  2.6686ms  [CUDA memset]
                    0.07%  2.1084ms       672  3.1370us  1.6960us  7.0400us  void genericReformat::copyPackedKernel<int, int, bool=1, bool=1, genericReformat::IdentityCoordMapper<int=4>, int=4>(unsigned int, unsigned int, void const *, genericReformat::ArrayN<genericReformat::IdentityCoordMapper<int=4>>, genericReformat::ArrayNWithReducedDivisors<genericReformat::IdentityCoordMapper<int=4>>, genericReformat::ArrayN, int, int, int, float const *, void*, genericReformat::ArrayN, genericReformat::ArrayNWithReducedDivisors, genericReformat::ArrayNWithReducedDivisors, genericReformat::ArrayN, int, int, int, float const , int=4)
                    0.07%  2.0914ms       672  3.1120us  1.7930us  5.7930us  void cuSliceLayer::naiveSlice<int, cuSliceLayer::Mode, int=2>(cuSliceLayer::LaunchParams<int>)
                    0.05%  1.5643ms       336  4.6550us  2.6880us  9.5360us  void cuSliceLayer::naiveSlice<int, cuSliceLayer::Mode, int=5>(cuSliceLayer::LaunchParams<int>)
                    0.03%  1.1095ms       336  3.3020us  2.1120us  6.4330us  void cuCastLayer::cast<int, float, float, bool=0, bool=0, int=128, int=4>(float*, int, int const *, int, float const *, float const , int, int, int, int, bool, nvinfer1::rt::ReducedDivisor, nvinfer1::rt, nvinfer1::rt, int)
      API calls:   27.06%  2.62058s        10  262.06ms  9.9200us  2.62040s  cudaMemGetInfo
                   20.77%  2.01150s        16  125.72ms  2.5600us  2.01077s  cudaStreamCreateWithFlags
                   18.97%  1.83720s       112  16.404ms  13.974ms  28.470ms  cudaEventSynchronize
                   14.65%  1.41842s        51  27.812ms  1.7920us  936.87ms  cudaFree
                    6.94%  671.75ms       199  3.3756ms  7.2000us  91.680ms  cuModuleUnload
                    5.92%  573.56ms     28560  20.082us  11.617us  2.6013ms  cudaLaunchKernel
                    2.54%  246.10ms     12992  18.942us  11.872us  1.1696ms  cuLaunchKernel
                    1.73%  167.23ms       116  1.4416ms  936.57us  9.3130ms  cuModuleLoadData
                    0.50%  48.334ms       792  61.027us  11.168us  19.727ms  cudaMemcpyAsync
                    0.38%  36.548ms        39  937.13us  7.4240us  9.7707ms  cudaMalloc
                    0.16%  15.503ms     22522     688ns     288ns  97.701us  cudaGetLastError
                    0.11%  10.917ms       785  13.907us  4.1280us  219.09us  cudaEventRecord
                    0.07%  6.3573ms      7504     847ns     448ns  53.859us  cudaPeekAtLastError
                    0.06%  5.3570ms       784  6.8320us  1.8240us  57.635us  cudaEventElapsedTime
                    0.03%  2.6011ms       225  11.560us  3.9040us  746.09us  cudaStreamWaitEvent
                    0.02%  1.8308ms         5  366.16us  15.905us  1.6572ms  cudaMallocHost
                    0.01%  1.4453ms        14  103.23us  4.7360us  1.3135ms  cudaStreamSynchronize
                    0.01%  1.2439ms       480  2.5910us  1.5040us  38.082us  cudaFuncSetAttribute
                    0.01%  1.2147ms       505  2.4050us  1.5680us  51.747us  cudaEventCreateWithFlags
                    0.01%  1.1928ms         2  596.42us  567.01us  625.83us  cudaHostAlloc
                    0.01%  1.0127ms       505  2.0050us  1.2480us  86.053us  cudaEventDestroy
                    0.01%  904.46us       473  1.9120us     544ns  63.875us  cuDeviceGetAttribute
                    0.01%  779.76us         6  129.96us  63.171us  254.70us  cudaMemcpy
                    0.01%  641.48us       207  3.0980us     960ns  49.955us  cudaDeviceGetAttribute
                    0.00%  462.33us        10  46.233us  10.656us  162.09us  cudaMemsetAsync
                    0.00%  394.65us         1  394.65us  394.65us  394.65us  cudaLaunchHostFunc
                    0.00%  389.27us        11  35.388us  20.417us  72.836us  cudaGetDeviceProperties
                    0.00%  300.50us         2  150.25us  63.107us  237.39us  cudaFreeHost
                    0.00%  245.90us        30  8.1960us  2.6240us  37.698us  cudaStreamDestroy
                    0.00%  243.98us       116  2.1030us  1.2160us  26.562us  cuModuleGetFunction
                    0.00%  133.35us         6  22.225us  13.025us  32.289us  cudaStreamCreate
                    0.00%  113.58us         5  22.715us  8.2560us  38.370us  cuDeviceTotalMem
                    0.00%  107.33us        18  5.9620us  2.8160us  13.505us  cudaDeviceSynchronize
                    0.00%  97.413us         2  48.706us  43.586us  53.827us  cudaCreateTextureObject
                    0.00%  68.388us        16  4.2740us  1.4720us  15.681us  cudaGetDevice
                    0.00%  67.140us         2  33.570us  19.745us  47.395us  cudaSetDevice
                    0.00%  33.954us         2  16.977us  6.2730us  27.681us  cudaDestroyTextureObject
                    0.00%  26.819us         4  6.7040us  4.1930us  7.6810us  cuInit
                    0.00%  25.409us         8  3.1760us  2.6240us  4.5760us  cudaStreamCreateWithPriority
                    0.00%  20.193us         7  2.8840us     544ns  13.825us  cudaGetDeviceCount
                    0.00%  17.089us         2  8.5440us  8.1290us  8.9600us  cudaHostGetDevicePointer
                    0.00%  16.481us         4  4.1200us  1.9520us  6.1440us  cuDriverGetVersion
                    0.00%  15.200us         7  2.1710us  1.2800us  3.5840us  cuDeviceGetCount
                    0.00%  10.240us         5  2.0480us  1.2800us  2.4320us  cuDeviceGetName
                    0.00%  9.0570us         6  1.5090us     929ns  2.2080us  cuDeviceGet
                    0.00%  8.6410us         2  4.3200us  2.6880us  5.9530us  cudaDeviceGetStreamPriorityRange
                    0.00%  6.1120us         5  1.2220us     992ns  1.6320us  cuDeviceGetUuid
                    0.00%  3.9040us         5     780ns     544ns  1.0560us  cudaRuntimeGetVersion
                    0.00%  2.6240us         2  1.3120us     928ns  1.6960us  cudaDriverGetVersion
                    0.00%  2.2400us         1  2.2400us  2.2400us  2.2400us  cuDevicePrimaryCtxRelease
                    0.00%  2.1760us         2  1.0880us     672ns  1.5040us  cudaCreateChannelDesc

==25751== NVTX result:
==25751==   Thread "<unnamed>" (id = 2239747200)
==25751==     Domain "TensorRT"
==25751==       Range "(Unnamed Layer* 2154) [Shuffle]"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  165.55us       112  1.4780us  1.1200us  2.4650us  (Unnamed Layer* 2154) [Shuffle]
No kernels were profiled in this range.
No API activities were profiled in this range.

==25751==       Range "(Unnamed Layer* 2361) [Shuffle]"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  192.56us       112  1.7190us  1.0240us  25.858us  (Unnamed Layer* 2361) [Shuffle]
No kernels were profiled in this range.
No API activities were profiled in this range.

==25751==       Range "(Unnamed Layer* 2568) [Shuffle]"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  192.78us       112  1.7210us  1.0240us  21.505us  (Unnamed Layer* 2568) [Shuffle]
No kernels were profiled in this range.
No API activities were profiled in this range.

==25751==       Range "2157 copy"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  7.8724ms       112  70.289us  44.835us  131.14us  2157 copy
 GPU activities:  100.00%  4.3091ms       112  38.473us  36.450us  110.89us  void CUTENSOR_NAMESPACE::permutationKernelPLC3<CUTENSOR_NAMESPACE::VectorWrite2DTensorView<unsigned char=0, unsigned char=1, bool=0, unsigned int=4, float, char, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::VectorRead2DTensorView<unsigned char=1, unsigned char=0, bool=0, unsigned int=4, float, char, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::ThreadLevelElementwise<CUTENSOR_NAMESPACE::ElementwiseConfig2DSmallMode0<CUTENSOR_NAMESPACE::GeneralUnarySmall<float>, CUTENSOR_NAMESPACE::GeneralBinary<float>, int=2, int=4, int=256, int=64, char=4, bool=0, bool=0, bool=1, bool=0, bool=1>, float>, CUTENSOR_NAMESPACE::ElementwiseRuntimePLC3<char, float, char, float>::Params>(unsigned int=4)
      API calls:  100.00%  3.1244ms       112  27.896us  17.889us  80.836us  cudaLaunchKernel

==25751==       Range "2167 copy"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  7.2340ms       112  64.589us  39.490us  122.57us  2167 copy
 GPU activities:  100.00%  3.9976ms       112  35.692us  33.699us  107.82us  void CUTENSOR_NAMESPACE::permutationKernelPLC3<CUTENSOR_NAMESPACE::VectorWrite2DTensorView<unsigned char=0, unsigned char=1, bool=0, unsigned int=4, float, char, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::VectorRead2DTensorView<unsigned char=1, unsigned char=0, bool=0, unsigned int=4, float, char, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::ThreadLevelElementwise<CUTENSOR_NAMESPACE::ElementwiseConfig2DSmallMode0<CUTENSOR_NAMESPACE::GeneralUnarySmall<float>, CUTENSOR_NAMESPACE::GeneralBinary<float>, int=2, int=4, int=256, int=64, char=4, bool=0, bool=0, bool=1, bool=0, bool=1>, float>, CUTENSOR_NAMESPACE::ElementwiseRuntimePLC3<char, float, char, float>::Params>(unsigned int=4)
      API calls:  100.00%  2.8091ms       112  25.081us  15.009us  61.220us  cudaLaunchKernel

==25751==       Range "2177 copy"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  7.1025ms       112  63.415us  38.083us  158.15us  2177 copy
 GPU activities:  100.00%  4.0959ms       112  36.570us  34.754us  107.85us  void CUTENSOR_NAMESPACE::permutationKernelPLC3<CUTENSOR_NAMESPACE::VectorWrite2DTensorView<unsigned char=0, unsigned char=1, bool=0, unsigned int=4, float, char, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::VectorRead2DTensorView<unsigned char=1, unsigned char=0, bool=0, unsigned int=4, float, char, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::ThreadLevelElementwise<CUTENSOR_NAMESPACE::ElementwiseConfig2DSmallMode0<CUTENSOR_NAMESPACE::GeneralUnarySmall<float>, CUTENSOR_NAMESPACE::GeneralBinary<float>, int=2, int=4, int=256, int=64, char=4, bool=0, bool=0, bool=1, bool=0, bool=1>, float>, CUTENSOR_NAMESPACE::ElementwiseRuntimePLC3<char, float, char, float>::Params>(unsigned int=4)
      API calls:  100.00%  2.7913ms       112  24.922us  15.232us  78.213us  cudaLaunchKernel

==25751==       Range "3991 copy"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  4.1106ms       112  36.701us  24.353us  213.07us  3991 copy
 GPU activities:  100.00%  520.23us       112  4.6440us  4.1280us  6.4960us  void genericReformat::copyPackedKernel<int, int, bool=1, bool=1, genericReformat::IdentityCoordMapper<int=4>, int=4>(unsigned int, unsigned int, void const *, genericReformat::ArrayN<genericReformat::IdentityCoordMapper<int=4>>, genericReformat::ArrayNWithReducedDivisors<genericReformat::IdentityCoordMapper<int=4>>, genericReformat::ArrayN, int, int, int, float const *, void*, genericReformat::ArrayN, genericReformat::ArrayNWithReducedDivisors, genericReformat::ArrayNWithReducedDivisors, genericReformat::ArrayN, int, int, int, float const , int=4)
      API calls:  100.00%  2.4039ms       112  21.463us  15.361us  122.28us  cudaLaunchKernel

==25751==       Range "3992 copy"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.6649ms       112  23.793us  19.201us  83.141us  3992 copy
 GPU activities:  100.00%  255.96us       112  2.2850us  2.1440us  3.6170us  void genericReformat::copyPackedKernel<int, int, bool=1, bool=1, genericReformat::IdentityCoordMapper<int=4>, int=4>(unsigned int, unsigned int, void const *, genericReformat::ArrayN<genericReformat::IdentityCoordMapper<int=4>>, genericReformat::ArrayNWithReducedDivisors<genericReformat::IdentityCoordMapper<int=4>>, genericReformat::ArrayN, int, int, int, float const *, void*, genericReformat::ArrayN, genericReformat::ArrayNWithReducedDivisors, genericReformat::ArrayNWithReducedDivisors, genericReformat::ArrayN, int, int, int, float const , int=4)
      API calls:  100.00%  1.8140ms       112  16.196us  13.409us  35.074us  cudaLaunchKernel

==25751==       Range "4050 copy"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  3.4317ms       112  30.639us  22.881us  76.484us  4050 copy
 GPU activities:  100.00%  1.3381ms       112  11.947us  10.753us  15.361us  void genericReformat::copyPackedKernel<float, float, bool=1, bool=1, genericReformat::IdentityCoordMapper<int=4>, int=4>(unsigned int, unsigned int, void const *, genericReformat::ArrayN<genericReformat::IdentityCoordMapper<int=4>>, genericReformat::ArrayNWithReducedDivisors<genericReformat::IdentityCoordMapper<int=4>>, genericReformat::ArrayN, int, int, int, float const *, void*, genericReformat::ArrayN, genericReformat::ArrayNWithReducedDivisors, genericReformat::ArrayNWithReducedDivisors, genericReformat::ArrayN, int, int, int, float const , int=4)
      API calls:  100.00%  2.1818ms       112  19.480us  14.529us  41.250us  cudaLaunchKernel

==25751==       Range "4061 copy"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.6391ms       112  23.563us  18.881us  43.490us  4061 copy
 GPU activities:  100.00%  1.4116ms       112  12.603us  11.169us  15.233us  void genericReformat::copyPackedKernel<float, float, bool=1, bool=1, genericReformat::IdentityCoordMapper<int=4>, int=4>(unsigned int, unsigned int, void const *, genericReformat::ArrayN<genericReformat::IdentityCoordMapper<int=4>>, genericReformat::ArrayNWithReducedDivisors<genericReformat::IdentityCoordMapper<int=4>>, genericReformat::ArrayN, int, int, int, float const *, void*, genericReformat::ArrayN, genericReformat::ArrayNWithReducedDivisors, genericReformat::ArrayNWithReducedDivisors, genericReformat::ArrayN, int, int, int, float const , int=4)
      API calls:  100.00%  1.8704ms       112  16.699us  13.088us  35.202us  cudaLaunchKernel

==25751==       Range "4066 copy"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.7626ms       112  24.666us  18.914us  87.237us  4066 copy
 GPU activities:  100.00%  2.0710ms       112  18.491us  16.865us  21.506us  void genericReformat::copyPackedKernel<float, float, bool=1, bool=1, genericReformat::IdentityCoordMapper<int=4>, int=4>(unsigned int, unsigned int, void const *, genericReformat::ArrayN<genericReformat::IdentityCoordMapper<int=4>>, genericReformat::ArrayNWithReducedDivisors<genericReformat::IdentityCoordMapper<int=4>>, genericReformat::ArrayN, int, int, int, float const *, void*, genericReformat::ArrayN, genericReformat::ArrayNWithReducedDivisors, genericReformat::ArrayNWithReducedDivisors, genericReformat::ArrayN, int, int, int, float const , int=4)
      API calls:  100.00%  1.9786ms       112  17.666us  13.185us  80.773us  cudaLaunchKernel

==25751==       Range "4074 copy"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  3.3996ms       112  30.353us  19.362us  172.39us  4074 copy
 GPU activities:  100.00%  2.5293ms       112  22.583us  20.769us  28.194us  void genericReformat::copyPackedKernel<float, float, bool=1, bool=1, genericReformat::IdentityCoordMapper<int=4>, int=4>(unsigned int, unsigned int, void const *, genericReformat::ArrayN<genericReformat::IdentityCoordMapper<int=4>>, genericReformat::ArrayNWithReducedDivisors<genericReformat::IdentityCoordMapper<int=4>>, genericReformat::ArrayN, int, int, int, float const *, void*, genericReformat::ArrayN, genericReformat::ArrayNWithReducedDivisors, genericReformat::ArrayNWithReducedDivisors, genericReformat::ArrayN, int, int, int, float const , int=4)
      API calls:  100.00%  2.4449ms       112  21.829us  12.961us  164.97us  cudaLaunchKernel

==25751==       Range "4118 copy"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  3.1076ms       112  27.746us  20.513us  86.917us  4118 copy
 GPU activities:  100.00%  465.98us       112  4.1600us  3.7760us  7.0400us  void genericReformat::copyPackedKernel<int, int, bool=1, bool=1, genericReformat::IdentityCoordMapper<int=4>, int=4>(unsigned int, unsigned int, void const *, genericReformat::ArrayN<genericReformat::IdentityCoordMapper<int=4>>, genericReformat::ArrayNWithReducedDivisors<genericReformat::IdentityCoordMapper<int=4>>, genericReformat::ArrayN, int, int, int, float const *, void*, genericReformat::ArrayN, genericReformat::ArrayNWithReducedDivisors, genericReformat::ArrayNWithReducedDivisors, genericReformat::ArrayN, int, int, int, float const , int=4)
      API calls:  100.00%  1.9147ms       112  17.095us  13.825us  53.315us  cudaLaunchKernel

==25751==       Range "4119 copy"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.7472ms       112  24.528us  19.170us  66.596us  4119 copy
 GPU activities:  100.00%  217.45us       112  1.9410us  1.8240us  2.8800us  void genericReformat::copyPackedKernel<int, int, bool=1, bool=1, genericReformat::IdentityCoordMapper<int=4>, int=4>(unsigned int, unsigned int, void const *, genericReformat::ArrayN<genericReformat::IdentityCoordMapper<int=4>>, genericReformat::ArrayNWithReducedDivisors<genericReformat::IdentityCoordMapper<int=4>>, genericReformat::ArrayN, int, int, int, float const *, void*, genericReformat::ArrayN, genericReformat::ArrayNWithReducedDivisors, genericReformat::ArrayNWithReducedDivisors, genericReformat::ArrayN, int, int, int, float const , int=4)
      API calls:  100.00%  1.9392ms       112  17.314us  13.281us  60.324us  cudaLaunchKernel

==25751==       Range "4177 copy"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.8930ms       112  25.830us  20.193us  117.70us  4177 copy
 GPU activities:  100.00%  699.61us       112  6.2460us  5.2490us  8.1930us  void genericReformat::copyPackedKernel<float, float, bool=1, bool=1, genericReformat::IdentityCoordMapper<int=4>, int=4>(unsigned int, unsigned int, void const *, genericReformat::ArrayN<genericReformat::IdentityCoordMapper<int=4>>, genericReformat::ArrayNWithReducedDivisors<genericReformat::IdentityCoordMapper<int=4>>, genericReformat::ArrayN, int, int, int, float const *, void*, genericReformat::ArrayN, genericReformat::ArrayNWithReducedDivisors, genericReformat::ArrayNWithReducedDivisors, genericReformat::ArrayN, int, int, int, float const , int=4)
      API calls:  100.00%  1.8224ms       112  16.271us  13.313us  43.715us  cudaLaunchKernel

==25751==       Range "4188 copy"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.8209ms       112  25.186us  18.913us  214.83us  4188 copy
 GPU activities:  100.00%  429.12us       112  3.8310us  3.4880us  5.3440us  void genericReformat::copyPackedKernel<float, float, bool=1, bool=1, genericReformat::IdentityCoordMapper<int=4>, int=4>(unsigned int, unsigned int, void const *, genericReformat::ArrayN<genericReformat::IdentityCoordMapper<int=4>>, genericReformat::ArrayNWithReducedDivisors<genericReformat::IdentityCoordMapper<int=4>>, genericReformat::ArrayN, int, int, int, float const *, void*, genericReformat::ArrayN, genericReformat::ArrayNWithReducedDivisors, genericReformat::ArrayNWithReducedDivisors, genericReformat::ArrayN, int, int, int, float const , int=4)
      API calls:  100.00%  2.0753ms       112  18.529us  13.024us  206.70us  cudaLaunchKernel

==25751==       Range "4193 copy"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.9501ms       112  26.340us  18.657us  158.09us  4193 copy
 GPU activities:  100.00%  646.26us       112  5.7700us  5.0560us  7.4890us  void genericReformat::copyPackedKernel<float, float, bool=1, bool=1, genericReformat::IdentityCoordMapper<int=4>, int=4>(unsigned int, unsigned int, void const *, genericReformat::ArrayN<genericReformat::IdentityCoordMapper<int=4>>, genericReformat::ArrayNWithReducedDivisors<genericReformat::IdentityCoordMapper<int=4>>, genericReformat::ArrayN, int, int, int, float const *, void*, genericReformat::ArrayN, genericReformat::ArrayNWithReducedDivisors, genericReformat::ArrayNWithReducedDivisors, genericReformat::ArrayN, int, int, int, float const , int=4)
      API calls:  100.00%  2.1344ms       112  19.057us  12.897us  149.87us  cudaLaunchKernel

==25751==       Range "4201 copy"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  4.5617ms       112  40.729us  19.393us  222.35us  4201 copy
 GPU activities:  100.00%  951.40us       112  8.4940us  7.6800us  11.585us  void genericReformat::copyPackedKernel<float, float, bool=1, bool=1, genericReformat::IdentityCoordMapper<int=4>, int=4>(unsigned int, unsigned int, void const *, genericReformat::ArrayN<genericReformat::IdentityCoordMapper<int=4>>, genericReformat::ArrayNWithReducedDivisors<genericReformat::IdentityCoordMapper<int=4>>, genericReformat::ArrayN, int, int, int, float const *, void*, genericReformat::ArrayN, genericReformat::ArrayNWithReducedDivisors, genericReformat::ArrayNWithReducedDivisors, genericReformat::ArrayN, int, int, int, float const , int=4)
      API calls:  100.00%  3.7710ms       112  33.669us  12.737us  213.61us  cudaLaunchKernel

==25751==       Range "4245 copy"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  3.4523ms       112  30.823us  20.065us  175.53us  4245 copy
 GPU activities:  100.00%  438.46us       112  3.9140us  3.6160us  5.5050us  void genericReformat::copyPackedKernel<int, int, bool=1, bool=1, genericReformat::IdentityCoordMapper<int=4>, int=4>(unsigned int, unsigned int, void const *, genericReformat::ArrayN<genericReformat::IdentityCoordMapper<int=4>>, genericReformat::ArrayNWithReducedDivisors<genericReformat::IdentityCoordMapper<int=4>>, genericReformat::ArrayN, int, int, int, float const *, void*, genericReformat::ArrayN, genericReformat::ArrayNWithReducedDivisors, genericReformat::ArrayNWithReducedDivisors, genericReformat::ArrayN, int, int, int, float const , int=4)
      API calls:  100.00%  2.5282ms       112  22.573us  13.664us  166.89us  cudaLaunchKernel

==25751==       Range "4246 copy"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  3.0559ms       112  27.285us  19.009us  177.93us  4246 copy
 GPU activities:  100.00%  210.33us       112  1.8770us  1.6960us  3.3290us  void genericReformat::copyPackedKernel<int, int, bool=1, bool=1, genericReformat::IdentityCoordMapper<int=4>, int=4>(unsigned int, unsigned int, void const *, genericReformat::ArrayN<genericReformat::IdentityCoordMapper<int=4>>, genericReformat::ArrayNWithReducedDivisors<genericReformat::IdentityCoordMapper<int=4>>, genericReformat::ArrayN, int, int, int, float const *, void*, genericReformat::ArrayN, genericReformat::ArrayNWithReducedDivisors, genericReformat::ArrayNWithReducedDivisors, genericReformat::ArrayN, int, int, int, float const , int=4)
      API calls:  100.00%  2.2082ms       112  19.715us  13.121us  169.13us  cudaLaunchKernel

==25751==       Range "4304 copy"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.8623ms       112  25.556us  19.681us  71.588us  4304 copy
 GPU activities:  100.00%  476.99us       112  4.2580us  3.9360us  5.6320us  void genericReformat::copyPackedKernel<float, float, bool=1, bool=1, genericReformat::IdentityCoordMapper<int=4>, int=4>(unsigned int, unsigned int, void const *, genericReformat::ArrayN<genericReformat::IdentityCoordMapper<int=4>>, genericReformat::ArrayNWithReducedDivisors<genericReformat::IdentityCoordMapper<int=4>>, genericReformat::ArrayN, int, int, int, float const *, void*, genericReformat::ArrayN, genericReformat::ArrayNWithReducedDivisors, genericReformat::ArrayNWithReducedDivisors, genericReformat::ArrayN, int, int, int, float const , int=4)
      API calls:  100.00%  1.9962ms       112  17.823us  12.929us  64.932us  cudaLaunchKernel

==25751==       Range "4315 copy"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.8586ms       112  25.522us  18.753us  133.74us  4315 copy
 GPU activities:  100.00%  240.02us       112  2.1420us  1.9840us  3.4240us  void genericReformat::copyPackedKernel<float, float, bool=1, bool=1, genericReformat::IdentityCoordMapper<int=4>, int=4>(unsigned int, unsigned int, void const *, genericReformat::ArrayN<genericReformat::IdentityCoordMapper<int=4>>, genericReformat::ArrayNWithReducedDivisors<genericReformat::IdentityCoordMapper<int=4>>, genericReformat::ArrayN, int, int, int, float const *, void*, genericReformat::ArrayN, genericReformat::ArrayNWithReducedDivisors, genericReformat::ArrayNWithReducedDivisors, genericReformat::ArrayN, int, int, int, float const , int=4)
      API calls:  100.00%  2.0533ms       112  18.333us  12.992us  125.83us  cudaLaunchKernel

==25751==       Range "4320 copy"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  3.2203ms       112  28.752us  18.785us  123.14us  4320 copy
 GPU activities:  100.00%  270.27us       112  2.4130us  2.2720us  3.7770us  void genericReformat::copyPackedKernel<float, float, bool=1, bool=1, genericReformat::IdentityCoordMapper<int=4>, int=4>(unsigned int, unsigned int, void const *, genericReformat::ArrayN<genericReformat::IdentityCoordMapper<int=4>>, genericReformat::ArrayNWithReducedDivisors<genericReformat::IdentityCoordMapper<int=4>>, genericReformat::ArrayN, int, int, int, float const *, void*, genericReformat::ArrayN, genericReformat::ArrayNWithReducedDivisors, genericReformat::ArrayNWithReducedDivisors, genericReformat::ArrayN, int, int, int, float const , int=4)
      API calls:  100.00%  2.4461ms       112  21.840us  13.025us  117.16us  cudaLaunchKernel

==25751==       Range "4328 copy"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  4.2410ms       112  37.866us  19.009us  196.43us  4328 copy
 GPU activities:  100.00%  480.75us       112  4.2920us  3.8080us  6.2730us  void genericReformat::copyPackedKernel<float, float, bool=1, bool=1, genericReformat::IdentityCoordMapper<int=4>, int=4>(unsigned int, unsigned int, void const *, genericReformat::ArrayN<genericReformat::IdentityCoordMapper<int=4>>, genericReformat::ArrayNWithReducedDivisors<genericReformat::IdentityCoordMapper<int=4>>, genericReformat::ArrayN, int, int, int, float const *, void*, genericReformat::ArrayN, genericReformat::ArrayNWithReducedDivisors, genericReformat::ArrayNWithReducedDivisors, genericReformat::ArrayN, int, int, int, float const , int=4)
      API calls:  100.00%  3.3682ms       112  30.073us  12.737us  187.50us  cudaLaunchKernel

==25751==       Range "4339"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  274.00us       112  2.4460us  1.2480us  25.730us  4339
No kernels were profiled in this range.
No API activities were profiled in this range.

==25751==       Range "4353"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  121.86us       112  1.0880us     320ns  30.434us  4353
No kernels were profiled in this range.
No API activities were profiled in this range.

==25751==       Range "4367"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  67.108us       112     599ns     320ns  1.0880us  4367
No kernels were profiled in this range.
No API activities were profiled in this range.

==25751==       Range "Cast_1848"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.9577ms       112  26.408us  18.049us  159.05us  Cast_1848
 GPU activities:  100.00%  508.33us       112  4.5380us  3.9050us  6.4330us  void cuCastLayer::cast<int, float, float, bool=0, bool=0, int=128, int=4>(float*, int, int const *, int, float const *, float const , int, int, int, int, bool, nvinfer1::rt::ReducedDivisor, nvinfer1::rt, nvinfer1::rt, int)
      API calls:  100.00%  2.1362ms       112  19.073us  13.313us  121.67us  cudaLaunchKernel

==25751==       Range "Cast_1861"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  208.78us       112  1.8640us  1.1840us  3.2970us  Cast_1861
No kernels were profiled in this range.
No API activities were profiled in this range.

==25751==       Range "Cast_1947"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.3520ms       112  21.000us  16.769us  61.348us  Cast_1947
 GPU activities:  100.00%  311.60us       112  2.7820us  2.3360us  4.2560us  void cuCastLayer::cast<int, float, float, bool=0, bool=0, int=128, int=4>(float*, int, int const *, int, float const *, float const , int, int, int, int, bool, nvinfer1::rt::ReducedDivisor, nvinfer1::rt, nvinfer1::rt, int)
      API calls:  100.00%  1.7628ms       112  15.739us  12.513us  56.131us  cudaLaunchKernel

==25751==       Range "Cast_1960"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  179.83us       112  1.6050us     992ns  18.849us  Cast_1960
No kernels were profiled in this range.
No API activities were profiled in this range.

==25751==       Range "Cast_2046"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  3.1985ms       112  28.558us  16.257us  217.16us  Cast_2046
 GPU activities:  100.00%  289.56us       112  2.5850us  2.1120us  4.6730us  void cuCastLayer::cast<int, float, float, bool=0, bool=0, int=128, int=4>(float*, int, int const *, int, float const *, float const , int, int, int, int, bool, nvinfer1::rt::ReducedDivisor, nvinfer1::rt, nvinfer1::rt, int)
      API calls:  100.00%  2.5544ms       112  22.806us  12.257us  210.16us  cudaLaunchKernel

==25751==       Range "Cast_2059"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  163.14us       112  1.4560us  1.0240us  4.1600us  Cast_2059
No kernels were profiled in this range.
No API activities were profiled in this range.

==25751==       Range "Conv_1794"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  3.3755ms       112  30.138us  22.945us  85.093us  Conv_1794
 GPU activities:  100.00%  14.674ms       112  131.02us  121.45us  178.93us  trt_volta_scudnn_128x32_relu_interior_nn_v1
      API calls:  100.00%  2.5231ms       112  22.527us  16.801us  76.324us  cudaLaunchKernel

==25751==       Range "Conv_1893"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  3.1848ms       112  28.435us  21.441us  67.268us  Conv_1893
 GPU activities:  100.00%  11.389ms       112  101.68us  95.623us  141.16us  void fused::fusedConvolutionReluKernel<fused::SrcChwcPtr_FltTex_Reader<float, int=1, int=1, int=4, int=1>, fused::KpqkPtrWriter<float, int=1, int=1, int=1>, float, float, int=4, int=7, int=2, int=1, int=1, int=1, int=1>(fused::ConvolutionParams<floatSrcType, int=4, int=1Type>)
      API calls:  100.00%  2.2587ms       112  20.167us  15.681us  47.682us  cudaLaunchKernel

==25751==       Range "Conv_1992"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  3.3975ms       112  30.334us  22.081us  118.06us  Conv_1992
 GPU activities:  100.00%  5.8340ms       112  52.088us  48.355us  72.070us  void fused::fusedConvolutionReluKernel<fused::SrcChwcPtr_FltTex_Reader<float, int=1, int=1, int=4, int=1>, fused::KpqkPtrWriter<float, int=1, int=1, int=1>, float, float, int=4, int=7, int=8, int=1, int=1, int=1, int=1>(fused::ConvolutionParams<floatSrcType, int=4, int=1Type>)
      API calls:  100.00%  2.5053ms       112  22.368us  15.329us  104.61us  cudaLaunchKernel

==25751==       Range "ExecutionContext::enqueue"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.75164s       224  12.284ms  11.127ms  27.632ms  ExecutionContext::enqueue
 GPU activities:   24.56%  771.22ms     12992  59.361us  2.8160us  5.4506ms  generatedNativePointwise
                   23.89%  750.07ms      7504  99.956us  32.579us  2.3475ms  trt_volta_fp32_i8816cudnn_int8_128x128_ldg16_relu_small_nt_v1
                   13.33%  418.58ms      2800  149.49us  86.535us  731.29us  trt_volta_fp32_i8816cudnn_int8_256x64_ldg16_relu_singleBuffer_small_nt_v1
                    7.24%  227.37ms      3920  58.001us  19.490us  1.7740ms  void CUTENSOR_NAMESPACE::permutationKernelPLC3<CUTENSOR_NAMESPACE::VectorWrite2DTensorView<unsigned char=0, unsigned char=1, bool=0, unsigned int=4, float, char, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::VectorRead2DTensorView<unsigned char=1, unsigned char=0, bool=0, unsigned int=4, float, float, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::ThreadLevelElementwise<CUTENSOR_NAMESPACE::ElementwiseConfig2DCommonCase<CUTENSOR_NAMESPACE::GeneralUnarySmall<float>, CUTENSOR_NAMESPACE::GeneralBinary<float>, int=2, int=32, int=64, int=128, char=4, bool=0, bool=0, bool=1, bool=0, bool=0>, float>, CUTENSOR_NAMESPACE::ElementwiseRuntimePLC3<float, float, char, float>::Params>(unsigned int=4)
                    6.77%  212.65ms      2576  82.552us  41.699us  1.1712ms  void CUTENSOR_NAMESPACE::permutationKernelPLC3<CUTENSOR_NAMESPACE::VectorWrite2DTensorView<unsigned char=0, unsigned char=1, bool=0, unsigned int=4, float, float, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::VectorRead2DTensorView<unsigned char=1, unsigned char=0, bool=0, unsigned int=4, float, float, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::ThreadLevelElementwise<CUTENSOR_NAMESPACE::ElementwiseConfig2DCommonCase<CUTENSOR_NAMESPACE::GeneralUnarySmall<float>, CUTENSOR_NAMESPACE::GeneralBinary<float>, int=2, int=64, int=64, int=256, char=4, bool=0, bool=1, bool=1, bool=1, bool=0>, float>, CUTENSOR_NAMESPACE::ElementwiseRuntimePLC3<float, float, float, float>::Params>(unsigned int=4)
                    5.32%  167.15ms      1008  165.83us  103.40us  1.0296ms  trt_volta_fp32_icudnn_int8x4_128x64_relu_interior_nn_v1
                    5.31%  166.60ms       560  297.50us  127.11us  5.3590ms  trt_volta_fp32_icudnn_int8x4_128x64_relu_small_nn_v1
                    2.40%  75.400ms       224  336.61us  236.05us  1.0020ms  void cudnn::ops::pooling_fw_4d_kernel<float, float, cudnn::maxpooling_func<float, cudnnNanPropagation_t=0>, cudnnPoolingMode_t=0, bool=0>(cudnnTensorStruct, float const *, cudnn::ops::pooling_fw_4d_kernel<float, float, cudnn::maxpooling_func<float, cudnnNanPropagation_t=0>, cudnnPoolingMode_t=0, bool=0>, cudnnTensorStruct*, cudnnPoolingStruct, float, cudnnPoolingStruct, int, cudnn::reduced_divisor, float)
                    2.13%  66.916ms       112  597.46us  532.20us  1.7369ms  trt_volta_fp32_i8816cudnn_int8_256x64_ldg16_relu_singleBuffer_large_nt_v1
                    1.33%  41.653ms       896  46.487us  32.419us  144.24us  void cuSliceLayer::naiveSlice<float, cuSliceLayer::Mode, int=4>(cuSliceLayer::LaunchParams<float>)
                    1.25%  39.306ms       784  50.134us  17.089us  149.71us  cuInt8::nchwToNcqhw4(char const *, unsigned int*, int, int, int, int, int, int, float const *, float const *)
                    0.90%  28.134ms       224  125.60us  82.375us  238.58us  void cuResizeLayer::ResizeNearestGenericKernel<float, int=2, int=128, nvinfer1::ResizeRoundMode>(float*, cuResizeLayer::ResizeNearestGenericKernel<float, int=2, int=128, nvinfer1::ResizeRoundMode> const *, cuResizeLayer::LaunchParams)
                    0.81%  25.536ms       112  228.00us  221.68us  336.09us  void CUTENSOR_NAMESPACE::permutationKernelPLC3<CUTENSOR_NAMESPACE::VectorWrite2DTensorView<unsigned char=0, unsigned char=1, bool=0, unsigned int=16, float, char, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::VectorRead2DTensorView<unsigned char=1, unsigned char=0, bool=0, unsigned int=16, float, char, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::ThreadLevelElementwise<CUTENSOR_NAMESPACE::ElementwiseConfig2DCommonCase<CUTENSOR_NAMESPACE::GeneralUnarySmall<float>, CUTENSOR_NAMESPACE::GeneralBinary<float>, int=2, int=32, int=64, int=128, char=16, bool=0, bool=0, bool=1, bool=0, bool=1>, float>, CUTENSOR_NAMESPACE::ElementwiseRuntimePLC3<char, float, char, float>::Params>(unsigned int=16)
                    0.73%  22.820ms       672  33.958us  12.609us  2.0256ms  cuInt8::nc32hw32ToNcqhw4_block(char4 const *, char4*, int, int, int, int, float const *, float const *)
                    0.55%  17.408ms       672  25.904us  14.657us  51.620us  void cuFillLayer::fill<int>(cuFillLayer::KernelArgs<int>)
                    0.47%  14.674ms       112  131.02us  121.45us  178.93us  trt_volta_scudnn_128x32_relu_interior_nn_v1
                    0.39%  12.403ms       336  36.912us  33.699us  110.89us  void CUTENSOR_NAMESPACE::permutationKernelPLC3<CUTENSOR_NAMESPACE::VectorWrite2DTensorView<unsigned char=0, unsigned char=1, bool=0, unsigned int=4, float, char, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::VectorRead2DTensorView<unsigned char=1, unsigned char=0, bool=0, unsigned int=4, float, char, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::ThreadLevelElementwise<CUTENSOR_NAMESPACE::ElementwiseConfig2DSmallMode0<CUTENSOR_NAMESPACE::GeneralUnarySmall<float>, CUTENSOR_NAMESPACE::GeneralBinary<float>, int=2, int=4, int=256, int=64, char=4, bool=0, bool=0, bool=1, bool=0, bool=1>, float>, CUTENSOR_NAMESPACE::ElementwiseRuntimePLC3<char, float, char, float>::Params>(unsigned int=4)
                    0.37%  11.544ms      1344  8.5890us  1.9840us  28.194us  void genericReformat::copyPackedKernel<float, float, bool=1, bool=1, genericReformat::IdentityCoordMapper<int=4>, int=4>(unsigned int, unsigned int, void const *, genericReformat::ArrayN<genericReformat::IdentityCoordMapper<int=4>>, genericReformat::ArrayNWithReducedDivisors<genericReformat::IdentityCoordMapper<int=4>>, genericReformat::ArrayN, int, int, int, float const *, void*, genericReformat::ArrayN, genericReformat::ArrayNWithReducedDivisors, genericReformat::ArrayNWithReducedDivisors, genericReformat::ArrayN, int, int, int, float const , int=4)
                    0.36%  11.389ms       112  101.68us  95.623us  141.16us  void fused::fusedConvolutionReluKernel<fused::SrcChwcPtr_FltTex_Reader<float, int=1, int=1, int=4, int=1>, fused::KpqkPtrWriter<float, int=1, int=1, int=1>, float, float, int=4, int=7, int=2, int=1, int=1, int=1, int=1>(fused::ConvolutionParams<floatSrcType, int=4, int=1Type>)
                    0.25%  7.7303ms      1344  5.7510us  1.8560us  19.329us  void cuSliceLayer::naiveSlice<float, cuSliceLayer::Mode, int=5>(cuSliceLayer::LaunchParams<float>)
                    0.23%  7.1360ms       336  21.238us  17.441us  57.540us  void CUTENSOR_NAMESPACE::permutationKernelPLC3<CUTENSOR_NAMESPACE::VectorWrite2DTensorView<unsigned char=0, unsigned char=1, bool=1, unsigned int=4, float, char, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::VectorRead2DTensorView<unsigned char=0, unsigned char=1, bool=1, unsigned int=4, float, float, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::ThreadLevelElementwise<CUTENSOR_NAMESPACE::ElementwiseConfig1DCommonCase<CUTENSOR_NAMESPACE::GeneralUnarySmall<float>, CUTENSOR_NAMESPACE::GeneralBinary<float>, int=1, int=2048, int=1, int=128, char=4, bool=1, bool=0, bool=1, bool=0, bool=0>, float>, CUTENSOR_NAMESPACE::ElementwiseRuntimePLC3<float, float, char, float>::Params>(unsigned int=4)
                    0.22%  7.0220ms       112  62.696us  60.293us  88.711us  void nvinfer1::rt::cuda::poolCHW_PQT<int=5, int=5, int=1, int=1, int=1, int=1, int=1024, int=1, int=1, bool=0, nvinfer1::rt::cuda::ITiledPooling::PoolingMode, bool=0>(nvinfer1::rt::cuda::TiledPoolingParams, int)
                    0.22%  6.9970ms       112  62.473us  57.477us  132.94us  cuInt8::ncqhw4ToNc32hw32(char4 const *, char4*, nvinfer1::rt::ReducedDivisor, int, nvinfer1::rt::ReducedDivisor, nvinfer1::rt::ReducedDivisor, int, int, float const *, float const *)
                    0.19%  5.8340ms       112  52.088us  48.355us  72.070us  void fused::fusedConvolutionReluKernel<fused::SrcChwcPtr_FltTex_Reader<float, int=1, int=1, int=4, int=1>, fused::KpqkPtrWriter<float, int=1, int=1, int=1>, float, float, int=4, int=7, int=8, int=1, int=1, int=1, int=1>(fused::ConvolutionParams<floatSrcType, int=4, int=1Type>)
                    0.16%  5.1520ms       336  15.333us  5.6320us  37.251us  void genericReformat::copyPackedKernel<float, float, bool=1, bool=1, genericReformat::ArrayN<int=5>, int=5>(unsigned int, unsigned int, void const *, genericReformat::ArrayN<genericReformat::ArrayN<int=5>>, genericReformat::ArrayNWithReducedDivisors<genericReformat::ArrayN<int=5>>, void const *, int, int, int, float const *, void*, void const *, genericReformat::ArrayNWithReducedDivisors, genericReformat::ArrayNWithReducedDivisors, void const *, int, int, int, float const , int=5)
                    0.15%  4.7297ms       112  42.229us  40.100us  121.32us  void CUTENSOR_NAMESPACE::permutationKernelPLC3<CUTENSOR_NAMESPACE::VectorWrite2DTensorView<unsigned char=0, unsigned char=1, bool=0, unsigned int=4, float, char, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::VectorRead2DTensorView<unsigned char=0, unsigned char=1, bool=0, unsigned int=4, float, char, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::ThreadLevelElementwise<CUTENSOR_NAMESPACE::ElementwiseConfig2DSmallMode0<CUTENSOR_NAMESPACE::GeneralUnarySmall<float>, CUTENSOR_NAMESPACE::GeneralBinary<float>, int=2, int=4, int=256, int=64, char=4, bool=1, bool=0, bool=1, bool=0, bool=1>, float>, CUTENSOR_NAMESPACE::ElementwiseRuntimePLC3<char, float, char, float>::Params>(unsigned int=4)
                    0.13%  4.1140ms       112  36.732us  34.659us  101.96us  void CUTENSOR_NAMESPACE::permutationKernelPLC3<CUTENSOR_NAMESPACE::VectorWrite2DTensorView<unsigned char=0, unsigned char=1, bool=0, unsigned int=4, float, char, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::VectorRead2DTensorView<unsigned char=1, unsigned char=0, bool=0, unsigned int=4, float, float, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::ThreadLevelElementwise<CUTENSOR_NAMESPACE::ElementwiseConfig2DSmallMode0<CUTENSOR_NAMESPACE::GeneralUnarySmall<float>, CUTENSOR_NAMESPACE::GeneralBinary<float>, int=2, int=4, int=256, int=128, char=4, bool=0, bool=0, bool=1, bool=0, bool=1>, float>, CUTENSOR_NAMESPACE::ElementwiseRuntimePLC3<float, float, char, float>::Params>(unsigned int=4)
                    0.11%  3.4918ms       108  32.331us  6.8800us  280.79us  [CUDA memcpy DtoD]
                    0.07%  2.1084ms       672  3.1370us  1.6960us  7.0400us  void genericReformat::copyPackedKernel<int, int, bool=1, bool=1, genericReformat::IdentityCoordMapper<int=4>, int=4>(unsigned int, unsigned int, void const *, genericReformat::ArrayN<genericReformat::IdentityCoordMapper<int=4>>, genericReformat::ArrayNWithReducedDivisors<genericReformat::IdentityCoordMapper<int=4>>, genericReformat::ArrayN, int, int, int, float const *, void*, genericReformat::ArrayN, genericReformat::ArrayNWithReducedDivisors, genericReformat::ArrayNWithReducedDivisors, genericReformat::ArrayN, int, int, int, float const , int=4)
                    0.07%  2.0914ms       672  3.1120us  1.7930us  5.7930us  void cuSliceLayer::naiveSlice<int, cuSliceLayer::Mode, int=2>(cuSliceLayer::LaunchParams<int>)
                    0.05%  1.5643ms       336  4.6550us  2.6880us  9.5360us  void cuSliceLayer::naiveSlice<int, cuSliceLayer::Mode, int=5>(cuSliceLayer::LaunchParams<int>)
                    0.04%  1.1095ms       336  3.3020us  2.1120us  6.4330us  void cuCastLayer::cast<int, float, float, bool=0, bool=0, int=128, int=4>(float*, int, int const *, int, float const *, float const , int, int, int, int, bool, nvinfer1::rt::ReducedDivisor, nvinfer1::rt, nvinfer1::rt, int)
                    0.00%  14.370us        10  1.4370us  1.1200us  2.7840us  [CUDA memcpy DtoH]
                    0.00%  5.9840us         5  1.1960us     928ns  2.0800us  [CUDA memcpy HtoD]
      API calls:   69.55%  573.56ms     28560  20.082us  11.617us  2.6013ms  cudaLaunchKernel
                   29.84%  246.10ms     12992  18.942us  11.872us  1.1696ms  cuLaunchKernel
                    0.61%  5.0392ms       123  40.968us  14.497us  813.71us  cudaMemcpyAsync

==25751==       Range "Expand_1828"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.7881ms       112  24.893us  18.625us  82.373us  Expand_1828
 GPU activities:  100.00%  454.89us       112  4.0610us  3.6480us  5.7930us  void cuSliceLayer::naiveSlice<int, cuSliceLayer::Mode, int=2>(cuSliceLayer::LaunchParams<int>)
      API calls:  100.00%  1.9472ms       112  17.385us  13.281us  53.539us  cudaLaunchKernel

==25751==       Range "Expand_1832"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.4587ms       112  21.952us  15.841us  210.09us  Expand_1832
 GPU activities:  100.00%  361.12us       112  3.2240us  2.8160us  5.4730us  void cuSliceLayer::naiveSlice<int, cuSliceLayer::Mode, int=2>(cuSliceLayer::LaunchParams<int>)
      API calls:  100.00%  1.8069ms       112  16.132us  12.096us  104.61us  cudaLaunchKernel

==25751==       Range "Expand_1847"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.5790ms       112  23.026us  17.121us  142.57us  Expand_1847
 GPU activities:  100.00%  765.79us       112  6.8370us  6.4000us  9.5360us  void cuSliceLayer::naiveSlice<int, cuSliceLayer::Mode, int=5>(cuSliceLayer::LaunchParams<int>)
      API calls:  100.00%  1.9365ms       112  17.289us  12.769us  135.05us  cudaLaunchKernel

==25751==       Range "Expand_1860"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  3.1287ms       112  27.934us  19.073us  69.412us  Expand_1860
 GPU activities:  100.00%  696.60us       112  6.2190us  5.6330us  9.2810us  void cuSliceLayer::naiveSlice<float, cuSliceLayer::Mode, int=5>(cuSliceLayer::LaunchParams<float>)
      API calls:  100.00%  2.0649ms       112  18.437us  13.473us  62.020us  cudaLaunchKernel

==25751==       Range "Expand_1927"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.4169ms       112  21.579us  16.833us  181.23us  Expand_1927
 GPU activities:  100.00%  430.31us       112  3.8420us  3.4560us  5.6970us  void cuSliceLayer::naiveSlice<int, cuSliceLayer::Mode, int=2>(cuSliceLayer::LaunchParams<int>)
      API calls:  100.00%  1.8261ms       112  16.304us  12.545us  167.85us  cudaLaunchKernel

==25751==       Range "Expand_1931"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.2240ms       112  19.857us  15.809us  44.674us  Expand_1931
 GPU activities:  100.00%  328.35us       112  2.9310us  2.5920us  4.7680us  void cuSliceLayer::naiveSlice<int, cuSliceLayer::Mode, int=2>(cuSliceLayer::LaunchParams<int>)
      API calls:  100.00%  1.6630ms       112  14.848us  11.745us  36.322us  cudaLaunchKernel

==25751==       Range "Expand_1946"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.5207ms       112  22.506us  15.937us  217.01us  Expand_1946
 GPU activities:  100.00%  454.33us       112  4.0560us  3.6800us  6.0160us  void cuSliceLayer::naiveSlice<int, cuSliceLayer::Mode, int=5>(cuSliceLayer::LaunchParams<int>)
      API calls:  100.00%  1.8807ms       112  16.791us  12.225us  210.03us  cudaLaunchKernel

==25751==       Range "Expand_1959"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.4967ms       112  22.291us  17.409us  53.027us  Expand_1959
 GPU activities:  100.00%  484.68us       112  4.3270us  3.9360us  6.3680us  void cuSliceLayer::naiveSlice<float, cuSliceLayer::Mode, int=5>(cuSliceLayer::LaunchParams<float>)
      API calls:  100.00%  1.7627ms       112  15.738us  12.512us  46.786us  cudaLaunchKernel

==25751==       Range "Expand_2026"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  3.2683ms       112  29.181us  16.033us  247.02us  Expand_2026
 GPU activities:  100.00%  301.27us       112  2.6890us  2.3360us  3.6480us  void cuSliceLayer::naiveSlice<int, cuSliceLayer::Mode, int=2>(cuSliceLayer::LaunchParams<int>)
      API calls:  100.00%  2.6855ms       112  23.977us  11.905us  239.73us  cudaLaunchKernel

==25751==       Range "Expand_2030"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.5131ms       112  22.438us  15.393us  106.82us  Expand_2030
 GPU activities:  100.00%  215.51us       112  1.9240us  1.7930us  3.2640us  void cuSliceLayer::naiveSlice<int, cuSliceLayer::Mode, int=2>(cuSliceLayer::LaunchParams<int>)
      API calls:  100.00%  1.9959ms       112  17.820us  11.712us  102.21us  cudaLaunchKernel

==25751==       Range "Expand_2045"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.8974ms       112  25.870us  15.489us  171.98us  Expand_2045
 GPU activities:  100.00%  344.15us       112  3.0720us  2.6880us  5.0240us  void cuSliceLayer::naiveSlice<int, cuSliceLayer::Mode, int=5>(cuSliceLayer::LaunchParams<int>)
      API calls:  100.00%  2.3285ms       112  20.790us  11.745us  167.40us  cudaLaunchKernel

==25751==       Range "Expand_2058"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.8769ms       112  25.686us  16.545us  239.57us  Expand_2058
 GPU activities:  100.00%  351.45us       112  3.1370us  2.5920us  5.3450us  void cuSliceLayer::naiveSlice<float, cuSliceLayer::Mode, int=5>(cuSliceLayer::LaunchParams<float>)
      API calls:  100.00%  2.2036ms       112  19.674us  12.257us  230.35us  cudaLaunchKernel

==25751==       Range "MaxPool_977"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  4.0630ms       112  36.276us  28.162us  128.81us  MaxPool_977
 GPU activities:  100.00%  7.0220ms       112  62.696us  60.293us  88.711us  void nvinfer1::rt::cuda::poolCHW_PQT<int=5, int=5, int=1, int=1, int=1, int=1, int=1024, int=1, int=1, bool=0, nvinfer1::rt::cuda::ITiledPooling::PoolingMode, bool=0>(nvinfer1::rt::cuda::TiledPoolingParams, int)
      API calls:  100.00%  2.3264ms       112  20.771us  17.089us  46.659us  cudaLaunchKernel

==25751==       Range "MaxPool_978"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  6.1765ms       112  55.147us  40.962us  211.02us  MaxPool_978
 GPU activities:  100.00%  27.478ms       112  245.34us  236.05us  574.96us  void cudnn::ops::pooling_fw_4d_kernel<float, float, cudnn::maxpooling_func<float, cudnnNanPropagation_t=0>, cudnnPoolingMode_t=0, bool=0>(cudnnTensorStruct, float const *, cudnn::ops::pooling_fw_4d_kernel<float, float, cudnn::maxpooling_func<float, cudnnNanPropagation_t=0>, cudnnPoolingMode_t=0, bool=0>, cudnnTensorStruct*, cudnnPoolingStruct, float, cudnnPoolingStruct, int, cudnn::reduced_divisor, float)
      API calls:  100.00%  2.7481ms       112  24.536us  19.617us  55.971us  cudaLaunchKernel

==25751==       Range "MaxPool_979"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.8849ms       112  25.758us  19.937us  49.411us  MaxPool_979
 GPU activities:  100.00%  47.923ms       112  427.88us  411.84us  1.0020ms  void cudnn::ops::pooling_fw_4d_kernel<float, float, cudnn::maxpooling_func<float, cudnnNanPropagation_t=0>, cudnnPoolingMode_t=0, bool=0>(cudnnTensorStruct, float const *, cudnn::ops::pooling_fw_4d_kernel<float, float, cudnn::maxpooling_func<float, cudnnNanPropagation_t=0>, cudnnPoolingMode_t=0, bool=0>, cudnnTensorStruct*, cudnnPoolingStruct, float, cudnnPoolingStruct, int, cudnn::reduced_divisor, float)
      API calls:  100.00%  1.8555ms       112  16.566us  13.185us  33.602us  cudaLaunchKernel

==25751==       Range "PWN(PWN(4056 + (Unnamed Layer* 2247) [Shuffle] + Mul_1881, PWN(4343 + (Unnamed Layer* 2250) [Shuffle], Pow_1882)), Mul_1883)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.3912ms       112  21.349us  16.417us  68.452us  PWN(PWN(4056 + (Unnamed Layer* 2247) [Shuffle] + Mul_1881, PWN(4343 + (Unnamed Layer* 2250) [Shuffle], Pow_1882)), Mul_1883)
 GPU activities:  100.00%  901.63us       112  8.0500us  7.2000us  11.105us  generatedNativePointwise
      API calls:  100.00%  1.8645ms       112  16.647us  12.833us  63.588us  cuLaunchKernel

==25751==       Range "PWN(PWN(4183 + (Unnamed Layer* 2454) [Shuffle] + Mul_1980, PWN(4357 + (Unnamed Layer* 2457) [Shuffle], Pow_1981)), Mul_1982)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.4530ms       112  21.901us  15.841us  117.42us  PWN(PWN(4183 + (Unnamed Layer* 2454) [Shuffle] + Mul_1980, PWN(4357 + (Unnamed Layer* 2457) [Shuffle], Pow_1981)), Mul_1982)
 GPU activities:  100.00%  484.58us       112  4.3260us  3.9360us  6.2400us  generatedNativePointwise
      API calls:  100.00%  1.9394ms       112  17.316us  12.321us  107.91us  cuLaunchKernel

==25751==       Range "PWN(PWN(4310 + (Unnamed Layer* 2661) [Shuffle] + Mul_2079, PWN(4371 + (Unnamed Layer* 2664) [Shuffle], Pow_2080)), Mul_2081)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  3.1521ms       112  28.143us  15.457us  166.06us  PWN(PWN(4310 + (Unnamed Layer* 2661) [Shuffle] + Mul_2079, PWN(4371 + (Unnamed Layer* 2664) [Shuffle], Pow_2080)), Mul_2081)
 GPU activities:  100.00%  387.39us       112  3.4580us  3.0080us  4.9600us  generatedNativePointwise
      API calls:  100.00%  2.6149ms       112  23.347us  12.065us  159.88us  cuLaunchKernel

==25751==       Range "PWN(PWN(PWN(4044 + (Unnamed Layer* 2214) [Shuffle] + Mul_1869, PWN(4046 + (Unnamed Layer* 2217) [Shuffle], Sub_1871)), Add_1872), 4049 + (Unnamed Layer* 2221) [Shuffle] + Mul_1874)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.4440ms       112  21.821us  15.617us  89.125us  PWN(PWN(PWN(4044 + (Unnamed Layer* 2214) [Shuffle] + Mul_1869, PWN(4046 + (Unnamed Layer* 2217) [Shuffle], Sub_1871)), Add_1872), 4049 + (Unnamed Layer* 2221) [Shuffle] + Mul_1874)
 GPU activities:  100.00%  830.65us       112  7.4160us  6.6560us  11.329us  generatedNativePointwise
      API calls:  100.00%  1.8709ms       112  16.704us  12.417us  68.612us  cuLaunchKernel

==25751==       Range "PWN(PWN(PWN(4171 + (Unnamed Layer* 2421) [Shuffle] + Mul_1968, PWN(4173 + (Unnamed Layer* 2424) [Shuffle], Sub_1970)), Add_1971), 4176 + (Unnamed Layer* 2428) [Shuffle] + Mul_1973)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.3881ms       112  21.322us  15.841us  95.397us  PWN(PWN(PWN(4171 + (Unnamed Layer* 2421) [Shuffle] + Mul_1968, PWN(4173 + (Unnamed Layer* 2424) [Shuffle], Sub_1970)), Add_1971), 4176 + (Unnamed Layer* 2428) [Shuffle] + Mul_1973)
 GPU activities:  100.00%  500.49us       112  4.4680us  3.9370us  6.6890us  generatedNativePointwise
      API calls:  100.00%  1.9188ms       112  17.132us  12.448us  88.581us  cuLaunchKernel

==25751==       Range "PWN(PWN(PWN(4298 + (Unnamed Layer* 2628) [Shuffle] + Mul_2067, PWN(4300 + (Unnamed Layer* 2631) [Shuffle], Sub_2069)), Add_2070), 4303 + (Unnamed Layer* 2635) [Shuffle] + Mul_2072)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.7445ms       112  24.504us  15.232us  120.36us  PWN(PWN(PWN(4298 + (Unnamed Layer* 2628) [Shuffle] + Mul_2067, PWN(4300 + (Unnamed Layer* 2631) [Shuffle], Sub_2069)), Add_2070), 4303 + (Unnamed Layer* 2635) [Shuffle] + Mul_2072)
 GPU activities:  100.00%  356.79us       112  3.1850us  2.8160us  4.9280us  generatedNativePointwise
      API calls:  100.00%  2.2876ms       112  20.425us  12.064us  114.25us  cuLaunchKernel

==25751==       Range "PWN(PWN(Sigmoid_119, Mul_120), Add_121)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  3.2101ms       112  28.661us  18.561us  61.476us  PWN(PWN(Sigmoid_119, Mul_120), Add_121)
 GPU activities:  100.00%  19.256ms       112  171.93us  160.53us  515.11us  generatedNativePointwise
      API calls:  100.00%  2.4928ms       112  22.256us  14.305us  45.378us  cuLaunchKernel

==25751==       Range "PWN(PWN(Sigmoid_152, Mul_153), Add_154)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  3.0808ms       112  27.506us  18.433us  116.49us  PWN(PWN(Sigmoid_152, Mul_153), Add_154)
 GPU activities:  100.00%  18.989ms       112  169.54us  160.40us  248.21us  generatedNativePointwise
      API calls:  100.00%  2.3472ms       112  20.957us  14.112us  49.187us  cuLaunchKernel

==25751==       Range "PWN(PWN(Sigmoid_185, Mul_186), Add_187)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  3.0577ms       112  27.300us  19.297us  88.805us  PWN(PWN(Sigmoid_185, Mul_186), Add_187)
 GPU activities:  100.00%  14.793ms       112  132.08us  126.99us  186.10us  generatedNativePointwise
      API calls:  100.00%  2.3236ms       112  20.746us  14.881us  38.914us  cuLaunchKernel

==25751==       Range "PWN(PWN(Sigmoid_283, Mul_284), Add_285)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  3.0431ms       112  27.170us  19.361us  61.283us  PWN(PWN(Sigmoid_283, Mul_284), Add_285)
 GPU activities:  100.00%  9.8602ms       112  88.037us  81.543us  124.43us  generatedNativePointwise
      API calls:  100.00%  2.3777ms       112  21.229us  15.233us  55.587us  cuLaunchKernel

==25751==       Range "PWN(PWN(Sigmoid_316, Mul_317), Add_318)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  3.0797ms       112  27.497us  19.457us  125.86us  PWN(PWN(Sigmoid_316, Mul_317), Add_318)
 GPU activities:  100.00%  9.8580ms       112  88.017us  82.278us  124.75us  generatedNativePointwise
      API calls:  100.00%  2.4505ms       112  21.879us  15.297us  120.36us  cuLaunchKernel

==25751==       Range "PWN(PWN(Sigmoid_349, Mul_350), Add_351)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  3.1087ms       112  27.755us  19.201us  127.69us  PWN(PWN(Sigmoid_349, Mul_350), Add_351)
 GPU activities:  100.00%  9.8823ms       112  88.234us  82.183us  124.75us  generatedNativePointwise
      API calls:  100.00%  2.3817ms       112  21.265us  15.137us  63.620us  cuLaunchKernel

==25751==       Range "PWN(PWN(Sigmoid_382, Mul_383), Add_384)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  3.1199ms       112  27.856us  18.721us  120.52us  PWN(PWN(Sigmoid_382, Mul_383), Add_384)
 GPU activities:  100.00%  9.9127ms       112  88.506us  81.830us  124.71us  generatedNativePointwise
      API calls:  100.00%  2.4756ms       112  22.103us  14.656us  114.70us  cuLaunchKernel

==25751==       Range "PWN(PWN(Sigmoid_415, Mul_416), Add_417)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.8618ms       112  25.551us  17.793us  68.004us  PWN(PWN(Sigmoid_415, Mul_416), Add_417)
 GPU activities:  100.00%  9.9025ms       112  88.415us  82.343us  129.71us  generatedNativePointwise
      API calls:  100.00%  2.2445ms       112  20.040us  13.985us  61.507us  cuLaunchKernel

==25751==       Range "PWN(PWN(Sigmoid_448, Mul_449), Add_450)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.9424ms       112  26.271us  18.977us  63.171us  PWN(PWN(Sigmoid_448, Mul_449), Add_450)
 GPU activities:  100.00%  9.8550ms       112  87.991us  81.991us  126.51us  generatedNativePointwise
      API calls:  100.00%  2.3291ms       112  20.795us  15.105us  58.499us  cuLaunchKernel

==25751==       Range "PWN(PWN(Sigmoid_481, Mul_482), Add_483)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.9598ms       112  26.426us  18.145us  59.395us  PWN(PWN(Sigmoid_481, Mul_482), Add_483)
 GPU activities:  100.00%  9.8630ms       112  88.062us  81.830us  124.91us  generatedNativePointwise
      API calls:  100.00%  2.2935ms       112  20.477us  14.368us  53.603us  cuLaunchKernel

==25751==       Range "PWN(PWN(Sigmoid_514, Mul_515), Add_516)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  3.9011ms       112  34.830us  18.465us  952.28us  PWN(PWN(Sigmoid_514, Mul_515), Add_516)
 GPU activities:  100.00%  9.8997ms       112  88.389us  83.206us  125.58us  generatedNativePointwise
      API calls:  100.00%  3.2479ms       112  28.999us  14.881us  924.82us  cuLaunchKernel

==25751==       Range "PWN(PWN(Sigmoid_547, Mul_548), Add_549)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.9464ms       112  26.306us  18.817us  61.891us  PWN(PWN(Sigmoid_547, Mul_548), Add_549)
 GPU activities:  100.00%  7.8489ms       112  70.079us  65.989us  95.464us  generatedNativePointwise
      API calls:  100.00%  2.2872ms       112  20.421us  14.593us  55.491us  cuLaunchKernel

==25751==       Range "PWN(PWN(Sigmoid_645, Mul_646), Add_647)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.7627ms       112  24.666us  18.241us  42.019us  PWN(PWN(Sigmoid_645, Mul_646), Add_647)
 GPU activities:  100.00%  5.0951ms       112  45.491us  42.435us  64.965us  generatedNativePointwise
      API calls:  100.00%  2.1802ms       112  19.465us  14.560us  33.282us  cuLaunchKernel

==25751==       Range "PWN(PWN(Sigmoid_678, Mul_679), Add_680)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.8171ms       112  25.152us  18.753us  78.149us  PWN(PWN(Sigmoid_678, Mul_679), Add_680)
 GPU activities:  100.00%  5.1145ms       112  45.665us  41.988us  66.149us  generatedNativePointwise
      API calls:  100.00%  2.2136ms       112  19.764us  14.849us  72.932us  cuLaunchKernel

==25751==       Range "PWN(PWN(Sigmoid_711, Mul_712), Add_713)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.7274ms       112  24.351us  17.793us  88.869us  PWN(PWN(Sigmoid_711, Mul_712), Add_713)
 GPU activities:  100.00%  5.0850ms       112  45.401us  42.147us  65.477us  generatedNativePointwise
      API calls:  100.00%  2.1339ms       112  19.052us  13.953us  78.788us  cuLaunchKernel

==25751==       Range "PWN(PWN(Sigmoid_744, Mul_745), Add_746)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.5554ms       112  22.816us  17.313us  47.203us  PWN(PWN(Sigmoid_744, Mul_745), Add_746)
 GPU activities:  100.00%  5.0760ms       112  45.321us  41.859us  65.413us  generatedNativePointwise
      API calls:  100.00%  1.9653ms       112  17.547us  13.440us  34.690us  cuLaunchKernel

==25751==       Range "PWN(PWN(Sigmoid_777, Mul_778), Add_779)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.7840ms       112  24.856us  17.857us  96.965us  PWN(PWN(Sigmoid_777, Mul_778), Add_779)
 GPU activities:  100.00%  5.1228ms       112  45.739us  42.019us  65.157us  generatedNativePointwise
      API calls:  100.00%  2.1097ms       112  18.837us  13.920us  73.348us  cuLaunchKernel

==25751==       Range "PWN(PWN(Sigmoid_810, Mul_811), Add_812)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.5069ms       112  22.383us  17.985us  76.356us  PWN(PWN(Sigmoid_810, Mul_811), Add_812)
 GPU activities:  100.00%  5.1196ms       112  45.710us  42.083us  66.021us  generatedNativePointwise
      API calls:  100.00%  1.9260ms       112  17.196us  14.144us  31.362us  cuLaunchKernel

==25751==       Range "PWN(PWN(Sigmoid_843, Mul_844), Add_845)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.5781ms       112  23.018us  17.281us  74.372us  PWN(PWN(Sigmoid_843, Mul_844), Add_845)
 GPU activities:  100.00%  5.0951ms       112  45.492us  42.083us  64.869us  generatedNativePointwise
      API calls:  100.00%  2.0099ms       112  17.945us  13.697us  69.220us  cuLaunchKernel

==25751==       Range "PWN(PWN(Sigmoid_876, Mul_877), Add_878)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.5301ms       112  22.590us  18.113us  60.676us  PWN(PWN(Sigmoid_876, Mul_877), Add_878)
 GPU activities:  100.00%  5.0818ms       112  45.372us  42.307us  64.997us  generatedNativePointwise
      API calls:  100.00%  1.9933ms       112  17.797us  14.208us  55.971us  cuLaunchKernel

==25751==       Range "PWN(PWN(Sigmoid_909, Mul_910), Add_911)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.5901ms       112  23.126us  17.793us  112.17us  PWN(PWN(Sigmoid_909, Mul_910), Add_911)
 GPU activities:  100.00%  4.4356ms       112  39.603us  36.707us  53.413us  generatedNativePointwise
      API calls:  100.00%  2.0148ms       112  17.989us  14.017us  105.73us  cuLaunchKernel

==25751==       Range "PWN(Sigmoid_1011, Mul_1012)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.4471ms       112  21.849us  16.961us  65.028us  PWN(Sigmoid_1011, Mul_1012)
 GPU activities:  100.00%  2.4094ms       112  21.512us  19.394us  28.290us  generatedNativePointwise
      API calls:  100.00%  1.9405ms       112  17.325us  13.601us  61.316us  cuLaunchKernel

==25751==       Range "PWN(Sigmoid_1027, Mul_1028)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.3552ms       112  21.028us  16.449us  74.276us  PWN(Sigmoid_1027, Mul_1028)
 GPU activities:  100.00%  2.3022ms       112  20.555us  19.041us  27.842us  generatedNativePointwise
      API calls:  100.00%  1.8929ms       112  16.900us  13.313us  70.852us  cuLaunchKernel

==25751==       Range "PWN(Sigmoid_103, Mul_104)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  3.1343ms       112  27.985us  16.929us  121.26us  PWN(Sigmoid_103, Mul_104)
 GPU activities:  100.00%  9.3095ms       112  83.120us  78.054us  246.93us  generatedNativePointwise
      API calls:  100.00%  2.5002ms       112  22.323us  13.537us  115.66us  cuLaunchKernel

==25751==       Range "PWN(Sigmoid_1043, Mul_1044)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.3977ms       112  21.408us  16.609us  46.499us  PWN(Sigmoid_1043, Mul_1044)
 GPU activities:  100.00%  2.3634ms       112  21.101us  18.817us  29.090us  generatedNativePointwise
      API calls:  100.00%  1.9389ms       112  17.311us  13.472us  42.594us  cuLaunchKernel

==25751==       Range "PWN(Sigmoid_1059, Mul_1060)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.4371ms       112  21.759us  16.897us  68.868us  PWN(Sigmoid_1059, Mul_1060)
 GPU activities:  100.00%  2.3175ms       112  20.692us  18.977us  28.802us  generatedNativePointwise
      API calls:  100.00%  1.9882ms       112  17.751us  13.889us  64.420us  cuLaunchKernel

==25751==       Range "PWN(Sigmoid_1075, Mul_1076)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.4573ms       112  21.940us  16.737us  51.171us  PWN(Sigmoid_1075, Mul_1076)
 GPU activities:  100.00%  2.2842ms       112  20.395us  18.466us  27.618us  generatedNativePointwise
      API calls:  100.00%  1.9534ms       112  17.441us  13.185us  47.139us  cuLaunchKernel

==25751==       Range "PWN(Sigmoid_1091, Mul_1092)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.3845ms       112  21.290us  16.481us  82.629us  PWN(Sigmoid_1091, Mul_1092)
 GPU activities:  100.00%  2.2573ms       112  20.154us  18.337us  26.114us  generatedNativePointwise
      API calls:  100.00%  1.8821ms       112  16.804us  13.249us  77.893us  cuLaunchKernel

==25751==       Range "PWN(Sigmoid_1107, Mul_1108)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.2946ms       112  20.487us  15.617us  51.363us  PWN(Sigmoid_1107, Mul_1108)
 GPU activities:  100.00%  2.2732ms       112  20.296us  18.434us  30.178us  generatedNativePointwise
      API calls:  100.00%  1.8577ms       112  16.586us  12.961us  43.618us  cuLaunchKernel

==25751==       Range "PWN(Sigmoid_1123, Mul_1124)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.2736ms       112  20.299us  16.289us  43.811us  PWN(Sigmoid_1123, Mul_1124)
 GPU activities:  100.00%  2.3503ms       112  20.984us  19.041us  27.234us  generatedNativePointwise
      API calls:  100.00%  1.8251ms       112  16.295us  12.961us  39.171us  cuLaunchKernel

==25751==       Range "PWN(Sigmoid_1140, Mul_1141)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.3841ms       112  21.286us  17.217us  69.380us  PWN(Sigmoid_1140, Mul_1141)
 GPU activities:  100.00%  4.2075ms       112  37.566us  34.626us  51.460us  generatedNativePointwise
      API calls:  100.00%  1.8951ms       112  16.920us  13.761us  31.138us  cuLaunchKernel

==25751==       Range "PWN(Sigmoid_1156, Mul_1157)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.3124ms       112  20.646us  15.744us  67.652us  PWN(Sigmoid_1156, Mul_1157)
 GPU activities:  100.00%  2.0092ms       112  17.939us  16.257us  24.610us  generatedNativePointwise
      API calls:  100.00%  1.8061ms       112  16.125us  12.640us  33.921us  cuLaunchKernel

==25751==       Range "PWN(Sigmoid_1175, Mul_1176)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.4494ms       112  21.869us  16.513us  98.597us  PWN(Sigmoid_1175, Mul_1176)
 GPU activities:  100.00%  4.1806ms       112  37.326us  34.627us  50.564us  generatedNativePointwise
      API calls:  100.00%  1.9774ms       112  17.655us  13.089us  93.061us  cuLaunchKernel

==25751==       Range "PWN(Sigmoid_1191, Mul_1192)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.3829ms       112  21.276us  16.673us  75.588us  PWN(Sigmoid_1191, Mul_1192)
 GPU activities:  100.00%  4.2148ms       112  37.632us  34.563us  50.308us  generatedNativePointwise
      API calls:  100.00%  1.9154ms       112  17.101us  13.568us  70.116us  cuLaunchKernel

==25751==       Range "PWN(Sigmoid_1207, Mul_1208)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.2439ms       112  20.035us  15.808us  38.467us  PWN(Sigmoid_1207, Mul_1208)
 GPU activities:  100.00%  4.1930ms       112  37.437us  34.723us  52.580us  generatedNativePointwise
      API calls:  100.00%  1.8094ms       112  16.155us  12.768us  33.378us  cuLaunchKernel

==25751==       Range "PWN(Sigmoid_1223, Mul_1224)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.9408ms       112  26.257us  15.264us  767.95us  PWN(Sigmoid_1223, Mul_1224)
 GPU activities:  100.00%  4.1999ms       112  37.499us  34.467us  51.332us  generatedNativePointwise
      API calls:  100.00%  2.5013ms       112  22.332us  12.512us  762.51us  cuLaunchKernel

==25751==       Range "PWN(Sigmoid_1239, Mul_1240)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.2768ms       112  20.328us  15.201us  64.676us  PWN(Sigmoid_1239, Mul_1240)
 GPU activities:  100.00%  4.1765ms       112  37.290us  34.530us  50.084us  generatedNativePointwise
      API calls:  100.00%  1.8128ms       112  16.185us  12.321us  36.162us  cuLaunchKernel

==25751==       Range "PWN(Sigmoid_1255, Mul_1256)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.4076ms       112  21.496us  15.681us  76.132us  PWN(Sigmoid_1255, Mul_1256)
 GPU activities:  100.00%  4.2066ms       112  37.559us  34.499us  50.116us  generatedNativePointwise
      API calls:  100.00%  1.9652ms       112  17.546us  12.769us  72.100us  cuLaunchKernel

==25751==       Range "PWN(Sigmoid_1271, Mul_1272)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.2748ms       112  20.310us  15.393us  63.908us  PWN(Sigmoid_1271, Mul_1272)
 GPU activities:  100.00%  4.2371ms       112  37.831us  34.531us  52.196us  generatedNativePointwise
      API calls:  100.00%  1.8388ms       112  16.417us  12.609us  59.140us  cuLaunchKernel

==25751==       Range "PWN(Sigmoid_1287, Mul_1288)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.2842ms       112  20.394us  15.969us  67.204us  PWN(Sigmoid_1287, Mul_1288)
 GPU activities:  100.00%  4.1887ms       112  37.399us  34.786us  50.884us  generatedNativePointwise
      API calls:  100.00%  1.8220ms       112  16.267us  12.961us  61.859us  cuLaunchKernel

==25751==       Range "PWN(Sigmoid_1304, Mul_1305)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.2618ms       112  20.194us  16.289us  58.980us  PWN(Sigmoid_1304, Mul_1305)
 GPU activities:  100.00%  7.8064ms       112  69.700us  65.733us  95.175us  generatedNativePointwise
      API calls:  100.00%  1.8123ms       112  16.180us  13.057us  54.691us  cuLaunchKernel

==25751==       Range "PWN(Sigmoid_1320, Mul_1321)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.4774ms       112  22.119us  16.385us  71.108us  PWN(Sigmoid_1320, Mul_1321)
 GPU activities:  100.00%  3.6236ms       112  32.354us  29.763us  44.228us  generatedNativePointwise
      API calls:  100.00%  2.0181ms       112  18.018us  13.441us  67.076us  cuLaunchKernel

==25751==       Range "PWN(Sigmoid_1339, Mul_1340)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.4734ms       112  22.084us  16.769us  68.964us  PWN(Sigmoid_1339, Mul_1340)
 GPU activities:  100.00%  7.9530ms       112  71.009us  66.661us  95.368us  generatedNativePointwise
      API calls:  100.00%  1.9765ms       112  17.647us  13.632us  59.556us  cuLaunchKernel

==25751==       Range "PWN(Sigmoid_1355, Mul_1356)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.2391ms       112  19.992us  15.521us  47.779us  PWN(Sigmoid_1355, Mul_1356)
 GPU activities:  100.00%  7.9377ms       112  70.872us  66.373us  95.399us  generatedNativePointwise
      API calls:  100.00%  1.7754ms       112  15.851us  12.544us  38.434us  cuLaunchKernel

==25751==       Range "PWN(Sigmoid_136, Mul_137)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  3.1251ms       112  27.902us  18.177us  117.57us  PWN(Sigmoid_136, Mul_137)
 GPU activities:  100.00%  9.3439ms       112  83.427us  79.879us  119.21us  generatedNativePointwise
      API calls:  100.00%  2.4946ms       112  22.273us  13.857us  111.59us  cuLaunchKernel

==25751==       Range "PWN(Sigmoid_1371, Mul_1372)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.3261ms       112  20.768us  15.585us  64.100us  PWN(Sigmoid_1371, Mul_1372)
 GPU activities:  100.00%  7.9843ms       112  71.288us  66.725us  95.496us  generatedNativePointwise
      API calls:  100.00%  1.7857ms       112  15.943us  12.641us  43.331us  cuLaunchKernel

==25751==       Range "PWN(Sigmoid_1387, Mul_1388)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.2680ms       112  20.250us  15.585us  56.739us  PWN(Sigmoid_1387, Mul_1388)
 GPU activities:  100.00%  7.9066ms       112  70.594us  65.349us  97.352us  generatedNativePointwise
      API calls:  100.00%  1.7921ms       112  16.001us  12.641us  53.091us  cuLaunchKernel

==25751==       Range "PWN(Sigmoid_1403, Mul_1404)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.2868ms       112  20.417us  16.033us  40.354us  PWN(Sigmoid_1403, Mul_1404)
 GPU activities:  100.00%  7.9184ms       112  70.700us  66.757us  96.488us  generatedNativePointwise
      API calls:  100.00%  1.8416ms       112  16.442us  12.801us  35.714us  cuLaunchKernel

==25751==       Range "PWN(Sigmoid_1419, Mul_1420)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.3868ms       112  21.310us  16.289us  73.829us  PWN(Sigmoid_1419, Mul_1420)
 GPU activities:  100.00%  7.8956ms       112  70.496us  65.893us  98.856us  generatedNativePointwise
      API calls:  100.00%  1.9029ms       112  16.990us  13.313us  58.660us  cuLaunchKernel

==25751==       Range "PWN(Sigmoid_1435, Mul_1436)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.1616ms       112  19.299us  15.201us  37.570us  PWN(Sigmoid_1435, Mul_1436)
 GPU activities:  100.00%  7.7950ms       112  69.598us  65.669us  92.615us  generatedNativePointwise
      API calls:  100.00%  1.7097ms       112  15.265us  12.288us  31.362us  cuLaunchKernel

==25751==       Range "PWN(Sigmoid_1451, Mul_1452)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.1952ms       112  19.600us  14.880us  44.707us  PWN(Sigmoid_1451, Mul_1452)
 GPU activities:  100.00%  7.7327ms       112  69.042us  64.581us  91.207us  generatedNativePointwise
      API calls:  100.00%  1.7526ms       112  15.647us  11.872us  40.194us  cuLaunchKernel

==25751==       Range "PWN(Sigmoid_1468, Mul_1469)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.3566ms       112  21.041us  16.129us  81.477us  PWN(Sigmoid_1468, Mul_1469)
 GPU activities:  100.00%  13.071ms       112  116.71us  110.09us  160.43us  generatedNativePointwise
      API calls:  100.00%  1.8450ms       112  16.473us  13.025us  42.915us  cuLaunchKernel

==25751==       Range "PWN(Sigmoid_1484, Mul_1485)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.3611ms       112  21.081us  16.961us  40.322us  PWN(Sigmoid_1484, Mul_1485)
 GPU activities:  100.00%  4.1954ms       112  37.459us  34.211us  50.820us  generatedNativePointwise
      API calls:  100.00%  1.8259ms       112  16.302us  13.664us  27.874us  cuLaunchKernel

==25751==       Range "PWN(Sigmoid_1501, Mul_1502)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.3013ms       112  20.547us  15.873us  69.732us  PWN(Sigmoid_1501, Mul_1502)
 GPU activities:  100.00%  4.2156ms       112  37.639us  34.755us  50.084us  generatedNativePointwise
      API calls:  100.00%  1.7700ms       112  15.803us  12.416us  36.386us  cuLaunchKernel

==25751==       Range "PWN(Sigmoid_1517, Mul_1518)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.3901ms       112  21.339us  16.513us  132.65us  PWN(Sigmoid_1517, Mul_1518)
 GPU activities:  100.00%  4.1755ms       112  37.281us  34.498us  52.868us  generatedNativePointwise
      API calls:  100.00%  1.8856ms       112  16.835us  13.121us  102.66us  cuLaunchKernel

==25751==       Range "PWN(Sigmoid_1533, Mul_1534)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.4043ms       112  21.466us  15.617us  73.572us  PWN(Sigmoid_1533, Mul_1534)
 GPU activities:  100.00%  4.2899ms       112  38.302us  35.683us  50.404us  generatedNativePointwise
      API calls:  100.00%  1.9047ms       112  17.006us  12.289us  68.356us  cuLaunchKernel

==25751==       Range "PWN(Sigmoid_1549, Mul_1550)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.2077ms       112  19.711us  15.713us  50.179us  PWN(Sigmoid_1549, Mul_1550)
 GPU activities:  100.00%  4.1806ms       112  37.326us  34.562us  50.404us  generatedNativePointwise
      API calls:  100.00%  1.7281ms       112  15.429us  12.513us  23.522us  cuLaunchKernel

==25751==       Range "PWN(Sigmoid_1565, Mul_1566)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.3077ms       112  20.604us  16.289us  73.476us  PWN(Sigmoid_1565, Mul_1566)
 GPU activities:  100.00%  4.2354ms       112  37.816us  34.819us  50.244us  generatedNativePointwise
      API calls:  100.00%  1.7958ms       112  16.033us  13.024us  37.858us  cuLaunchKernel

==25751==       Range "PWN(Sigmoid_1581, Mul_1582)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.2829ms       112  20.383us  14.849us  61.667us  PWN(Sigmoid_1581, Mul_1582)
 GPU activities:  100.00%  4.1510ms       112  37.062us  34.178us  50.404us  generatedNativePointwise
      API calls:  100.00%  1.8292ms       112  16.332us  11.904us  55.715us  cuLaunchKernel

==25751==       Range "PWN(Sigmoid_1597, Mul_1598)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.1816ms       112  19.478us  15.008us  33.922us  PWN(Sigmoid_1597, Mul_1598)
 GPU activities:  100.00%  4.1815ms       112  37.334us  34.627us  50.788us  generatedNativePointwise
      API calls:  100.00%  1.7694ms       112  15.798us  12.160us  30.434us  cuLaunchKernel

==25751==       Range "PWN(Sigmoid_1613, Mul_1614)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.2023ms       112  19.663us  15.745us  66.692us  PWN(Sigmoid_1613, Mul_1614)
 GPU activities:  100.00%  4.1802ms       112  37.323us  34.275us  51.684us  generatedNativePointwise
      API calls:  100.00%  1.7154ms       112  15.315us  12.609us  45.315us  cuLaunchKernel

==25751==       Range "PWN(Sigmoid_1630, Mul_1631)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.4484ms       112  21.860us  16.033us  63.332us  PWN(Sigmoid_1630, Mul_1631)
 GPU activities:  100.00%  6.9156ms       112  61.746us  57.573us  83.142us  generatedNativePointwise
      API calls:  100.00%  1.9747ms       112  17.631us  12.961us  57.188us  cuLaunchKernel

==25751==       Range "PWN(Sigmoid_1646, Mul_1647)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.4300ms       112  21.696us  16.097us  73.093us  PWN(Sigmoid_1646, Mul_1647)
 GPU activities:  100.00%  2.3192ms       112  20.707us  18.401us  28.290us  generatedNativePointwise
      API calls:  100.00%  1.9432ms       112  17.350us  12.929us  67.460us  cuLaunchKernel

==25751==       Range "PWN(Sigmoid_1663, Mul_1664)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.2771ms       112  20.330us  16.321us  40.835us  PWN(Sigmoid_1663, Mul_1664)
 GPU activities:  100.00%  2.3832ms       112  21.278us  18.849us  29.187us  generatedNativePointwise
      API calls:  100.00%  1.8139ms       112  16.195us  13.216us  37.666us  cuLaunchKernel

==25751==       Range "PWN(Sigmoid_1679, Mul_1680)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.4206ms       112  21.612us  16.353us  78.468us  PWN(Sigmoid_1679, Mul_1680)
 GPU activities:  100.00%  2.2863ms       112  20.412us  18.817us  26.882us  generatedNativePointwise
      API calls:  100.00%  1.9655ms       112  17.548us  13.121us  75.076us  cuLaunchKernel

==25751==       Range "PWN(Sigmoid_169, Mul_170)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.9620ms       112  26.446us  18.977us  102.02us  PWN(Sigmoid_169, Mul_170)
 GPU activities:  100.00%  9.4403ms       112  84.288us  80.422us  119.82us  generatedNativePointwise
      API calls:  100.00%  2.3464ms       112  20.950us  15.073us  96.549us  cuLaunchKernel

==25751==       Range "PWN(Sigmoid_1695, Mul_1696)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.6065ms       112  23.271us  15.489us  138.31us  PWN(Sigmoid_1695, Mul_1696)
 GPU activities:  100.00%  2.4507ms       112  21.881us  19.970us  27.906us  generatedNativePointwise
      API calls:  100.00%  2.0415ms       112  18.227us  12.449us  118.92us  cuLaunchKernel

==25751==       Range "PWN(Sigmoid_1711, Mul_1712)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.3310ms       112  20.812us  16.001us  113.38us  PWN(Sigmoid_1711, Mul_1712)
 GPU activities:  100.00%  2.3348ms       112  20.846us  19.073us  26.626us  generatedNativePointwise
      API calls:  100.00%  1.8751ms       112  16.742us  12.449us  107.97us  cuLaunchKernel

==25751==       Range "PWN(Sigmoid_1727, Mul_1728)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  3.2753ms       112  29.243us  15.424us  925.59us  PWN(Sigmoid_1727, Mul_1728)
 GPU activities:  100.00%  2.3254ms       112  20.762us  18.785us  27.778us  generatedNativePointwise
      API calls:  100.00%  2.8474ms       112  25.422us  12.320us  919.73us  cuLaunchKernel

==25751==       Range "PWN(Sigmoid_1743, Mul_1744)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.6926ms       112  24.041us  14.817us  253.10us  PWN(Sigmoid_1743, Mul_1744)
 GPU activities:  100.00%  2.2959ms       112  20.499us  18.561us  30.722us  generatedNativePointwise
      API calls:  100.00%  2.2469ms       112  20.061us  11.873us  246.70us  cuLaunchKernel

==25751==       Range "PWN(Sigmoid_1759, Mul_1760)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.3577ms       112  21.050us  15.905us  122.22us  PWN(Sigmoid_1759, Mul_1760)
 GPU activities:  100.00%  2.2961ms       112  20.500us  18.625us  28.450us  generatedNativePointwise
      API calls:  100.00%  1.9523ms       112  17.431us  12.992us  116.23us  cuLaunchKernel

==25751==       Range "PWN(Sigmoid_1775, Mul_1776)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.3803ms       112  21.252us  15.777us  103.08us  PWN(Sigmoid_1775, Mul_1776)
 GPU activities:  100.00%  2.3182ms       112  20.698us  18.850us  27.171us  generatedNativePointwise
      API calls:  100.00%  1.9114ms       112  17.066us  12.961us  97.926us  cuLaunchKernel

==25751==       Range "PWN(Sigmoid_1792, Mul_1793)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  4.2792ms       112  38.207us  15.745us  1.1763ms  PWN(Sigmoid_1792, Mul_1793)
 GPU activities:  100.00%  3.6257ms       112  32.371us  29.730us  44.324us  generatedNativePointwise
      API calls:  100.00%  3.6869ms       112  32.918us  12.513us  1.1696ms  cuLaunchKernel

==25751==       Range "PWN(Sigmoid_1862)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.4260ms       112  21.660us  16.641us  47.555us  PWN(Sigmoid_1862)
 GPU activities:  100.00%  1.9956ms       112  17.818us  16.129us  23.202us  generatedNativePointwise
      API calls:  100.00%  1.9282ms       112  17.215us  13.089us  41.955us  cuLaunchKernel

==25751==       Range "PWN(Sigmoid_1961)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.5493ms       112  22.761us  16.449us  222.22us  PWN(Sigmoid_1961)
 GPU activities:  100.00%  888.49us       112  7.9320us  6.6890us  11.937us  generatedNativePointwise
      API calls:  100.00%  1.9944ms       112  17.807us  12.928us  215.60us  cuLaunchKernel

==25751==       Range "PWN(Sigmoid_202, Mul_203)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  3.0428ms       112  27.167us  17.473us  146.99us  PWN(Sigmoid_202, Mul_203)
 GPU activities:  100.00%  9.3763ms       112  83.716us  78.502us  247.41us  generatedNativePointwise
      API calls:  100.00%  2.4392ms       112  21.778us  14.049us  138.98us  cuLaunchKernel

==25751==       Range "PWN(Sigmoid_2060)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  3.1865ms       112  28.450us  15.265us  797.52us  PWN(Sigmoid_2060)
 GPU activities:  100.00%  388.16us       112  3.4650us  3.1370us  5.4410us  generatedNativePointwise
      API calls:  100.00%  2.7515ms       112  24.567us  12.289us  792.85us  cuLaunchKernel

==25751==       Range "PWN(Sigmoid_219, Mul_220)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  3.0408ms       112  27.149us  18.593us  88.005us  PWN(Sigmoid_219, Mul_220)
 GPU activities:  100.00%  18.114ms       112  161.73us  155.69us  231.79us  generatedNativePointwise
      API calls:  100.00%  2.4317ms       112  21.711us  14.849us  84.197us  cuLaunchKernel

==25751==       Range "PWN(Sigmoid_235, Mul_236)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.9529ms       112  26.365us  18.113us  44.930us  PWN(Sigmoid_235, Mul_236)
 GPU activities:  100.00%  14.867ms       112  132.74us  124.14us  181.33us  generatedNativePointwise
      API calls:  100.00%  2.3125ms       112  20.647us  14.401us  37.538us  cuLaunchKernel

==25751==       Range "PWN(Sigmoid_251, Mul_252)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  3.0702ms       112  27.412us  18.081us  105.67us  PWN(Sigmoid_251, Mul_252)
 GPU activities:  100.00%  6.7999ms       112  60.713us  55.237us  83.079us  generatedNativePointwise
      API calls:  100.00%  2.4698ms       112  22.051us  14.657us  100.84us  cuLaunchKernel

==25751==       Range "PWN(Sigmoid_267, Mul_268)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.9745ms       112  26.558us  17.345us  62.563us  PWN(Sigmoid_267, Mul_268)
 GPU activities:  100.00%  7.9719ms       112  71.177us  65.638us  95.336us  generatedNativePointwise
      API calls:  100.00%  2.3725ms       112  21.183us  13.985us  56.227us  cuLaunchKernel

==25751==       Range "PWN(Sigmoid_300, Mul_301)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.9647ms       112  26.470us  18.017us  139.34us  PWN(Sigmoid_300, Mul_301)
 GPU activities:  100.00%  7.9769ms       112  71.221us  66.662us  95.336us  generatedNativePointwise
      API calls:  100.00%  2.3507ms       112  20.988us  14.145us  133.32us  cuLaunchKernel

==25751==       Range "PWN(Sigmoid_333, Mul_334)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.9627ms       112  26.452us  18.593us  93.221us  PWN(Sigmoid_333, Mul_334)
 GPU activities:  100.00%  7.9657ms       112  71.122us  66.661us  95.879us  generatedNativePointwise
      API calls:  100.00%  2.3399ms       112  20.892us  14.657us  85.541us  cuLaunchKernel

==25751==       Range "PWN(Sigmoid_366, Mul_367)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.7661ms       112  24.697us  17.857us  38.659us  PWN(Sigmoid_366, Mul_367)
 GPU activities:  100.00%  8.0011ms       112  71.438us  66.693us  96.039us  generatedNativePointwise
      API calls:  100.00%  2.1777ms       112  19.443us  14.337us  33.218us  cuLaunchKernel

==25751==       Range "PWN(Sigmoid_399, Mul_400)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.9593ms       112  26.422us  19.169us  104.55us  PWN(Sigmoid_399, Mul_400)
 GPU activities:  100.00%  8.0116ms       112  71.532us  66.885us  96.743us  generatedNativePointwise
      API calls:  100.00%  2.3784ms       112  21.235us  15.073us  98.949us  cuLaunchKernel

==25751==       Range "PWN(Sigmoid_432, Mul_433)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.8750ms       112  25.669us  19.041us  87.301us  PWN(Sigmoid_432, Mul_433)
 GPU activities:  100.00%  7.9708ms       112  71.167us  66.566us  93.447us  generatedNativePointwise
      API calls:  100.00%  2.2890ms       112  20.437us  15.041us  81.253us  cuLaunchKernel

==25751==       Range "PWN(Sigmoid_465, Mul_466)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.9263ms       112  26.127us  17.665us  76.068us  PWN(Sigmoid_465, Mul_466)
 GPU activities:  100.00%  7.9438ms       112  70.926us  65.925us  99.240us  generatedNativePointwise
      API calls:  100.00%  2.2867ms       112  20.417us  13.985us  62.115us  cuLaunchKernel

==25751==       Range "PWN(Sigmoid_498, Mul_499)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.8317ms       112  25.282us  17.121us  58.627us  PWN(Sigmoid_498, Mul_499)
 GPU activities:  100.00%  8.0164ms       112  71.575us  66.501us  95.111us  generatedNativePointwise
      API calls:  100.00%  2.2024ms       112  19.663us  13.825us  43.586us  cuLaunchKernel

==25751==       Range "PWN(Sigmoid_531, Mul_532)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.8654ms       112  25.584us  17.185us  116.23us  PWN(Sigmoid_531, Mul_532)
 GPU activities:  100.00%  7.9922ms       112  71.359us  64.645us  96.487us  generatedNativePointwise
      API calls:  100.00%  2.2499ms       112  20.088us  13.057us  111.08us  cuLaunchKernel

==25751==       Range "PWN(Sigmoid_55, Mul_56)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  3.5663ms       112  31.842us  20.289us  102.63us  PWN(Sigmoid_55, Mul_56)
 GPU activities:  100.00%  74.829ms       112  668.12us  605.87us  5.4506ms  generatedNativePointwise
      API calls:  100.00%  2.7098ms       112  24.194us  14.625us  94.694us  cuLaunchKernel

==25751==       Range "PWN(Sigmoid_564, Mul_565)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  3.0069ms       112  26.847us  17.857us  116.97us  PWN(Sigmoid_564, Mul_565)
 GPU activities:  100.00%  7.8225ms       112  69.843us  64.997us  96.839us  generatedNativePointwise
      API calls:  100.00%  2.3917ms       112  21.354us  14.368us  111.88us  cuLaunchKernel

==25751==       Range "PWN(Sigmoid_581, Mul_582)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.8175ms       112  25.155us  17.025us  57.251us  PWN(Sigmoid_581, Mul_582)
 GPU activities:  100.00%  13.093ms       112  116.90us  110.28us  160.85us  generatedNativePointwise
      API calls:  100.00%  2.2732ms       112  20.296us  13.729us  52.259us  cuLaunchKernel

==25751==       Range "PWN(Sigmoid_597, Mul_598)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.7596ms       112  24.639us  18.465us  47.522us  PWN(Sigmoid_597, Mul_598)
 GPU activities:  100.00%  7.8207ms       112  69.827us  65.989us  95.655us  generatedNativePointwise
      API calls:  100.00%  2.1980ms       112  19.625us  14.785us  42.050us  cuLaunchKernel

==25751==       Range "PWN(Sigmoid_613, Mul_614)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.8027ms       112  25.023us  17.377us  68.132us  PWN(Sigmoid_613, Mul_614)
 GPU activities:  100.00%  3.6621ms       112  32.697us  29.922us  44.355us  generatedNativePointwise
      API calls:  100.00%  2.2483ms       112  20.074us  14.209us  63.076us  cuLaunchKernel

==25751==       Range "PWN(Sigmoid_629, Mul_630)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.7365ms       112  24.432us  17.569us  85.765us  PWN(Sigmoid_629, Mul_630)
 GPU activities:  100.00%  4.2289ms       112  37.758us  34.594us  50.596us  generatedNativePointwise
      API calls:  100.00%  2.1783ms       112  19.448us  14.048us  80.324us  cuLaunchKernel

==25751==       Range "PWN(Sigmoid_662, Mul_663)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.7743ms       112  24.770us  17.537us  90.662us  PWN(Sigmoid_662, Mul_663)
 GPU activities:  100.00%  4.2239ms       112  37.713us  34.691us  49.700us  generatedNativePointwise
      API calls:  100.00%  2.2381ms       112  19.982us  14.241us  86.853us  cuLaunchKernel

==25751==       Range "PWN(Sigmoid_695, Mul_696)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.7420ms       112  24.482us  18.305us  105.45us  PWN(Sigmoid_695, Mul_696)
 GPU activities:  100.00%  4.1802ms       112  37.323us  34.818us  50.724us  generatedNativePointwise
      API calls:  100.00%  2.2086ms       112  19.719us  14.337us  100.13us  cuLaunchKernel

==25751==       Range "PWN(Sigmoid_71, Mul_72)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  3.1702ms       112  28.305us  18.209us  101.89us  PWN(Sigmoid_71, Mul_72)
 GPU activities:  100.00%  32.284ms       112  288.25us  256.63us  2.4834ms  generatedNativePointwise
      API calls:  100.00%  2.5189ms       112  22.490us  14.496us  96.006us  cuLaunchKernel

==25751==       Range "PWN(Sigmoid_728, Mul_729)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.7817ms       112  24.836us  18.721us  74.052us  PWN(Sigmoid_728, Mul_729)
 GPU activities:  100.00%  4.1976ms       112  37.478us  34.595us  51.812us  generatedNativePointwise
      API calls:  100.00%  2.2268ms       112  19.881us  14.848us  68.164us  cuLaunchKernel

==25751==       Range "PWN(Sigmoid_761, Mul_762)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.8463ms       112  25.413us  17.409us  115.21us  PWN(Sigmoid_761, Mul_762)
 GPU activities:  100.00%  4.2047ms       112  37.541us  34.819us  53.220us  generatedNativePointwise
      API calls:  100.00%  2.1631ms       112  19.313us  13.857us  99.398us  cuLaunchKernel

==25751==       Range "PWN(Sigmoid_794, Mul_795)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.5371ms       112  22.652us  16.673us  69.956us  PWN(Sigmoid_794, Mul_795)
 GPU activities:  100.00%  4.1965ms       112  37.469us  34.435us  51.044us  generatedNativePointwise
      API calls:  100.00%  1.9896ms       112  17.763us  13.121us  64.548us  cuLaunchKernel

==25751==       Range "PWN(Sigmoid_827, Mul_828)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.5872ms       112  23.100us  17.889us  44.771us  PWN(Sigmoid_827, Mul_828)
 GPU activities:  100.00%  4.1957ms       112  37.461us  34.915us  49.828us  generatedNativePointwise
      API calls:  100.00%  2.0587ms       112  18.381us  14.401us  39.746us  cuLaunchKernel

==25751==       Range "PWN(Sigmoid_860, Mul_861)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.6205ms       112  23.397us  18.145us  80.005us  PWN(Sigmoid_860, Mul_861)
 GPU activities:  100.00%  4.2014ms       112  37.512us  35.011us  50.852us  generatedNativePointwise
      API calls:  100.00%  2.0585ms       112  18.379us  14.305us  36.258us  cuLaunchKernel

==25751==       Range "PWN(Sigmoid_87, Mul_88)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  3.1723ms       112  28.324us  18.209us  87.365us  PWN(Sigmoid_87, Mul_88)
 GPU activities:  100.00%  16.001ms       112  142.87us  110.02us  2.2492ms  generatedNativePointwise
      API calls:  100.00%  2.5251ms       112  22.545us  14.529us  81.796us  cuLaunchKernel

==25751==       Range "PWN(Sigmoid_893, Mul_894)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.5166ms       112  22.469us  16.897us  77.892us  PWN(Sigmoid_893, Mul_894)
 GPU activities:  100.00%  4.2039ms       112  37.534us  35.010us  49.796us  generatedNativePointwise
      API calls:  100.00%  2.0153ms       112  17.994us  13.665us  71.844us  cuLaunchKernel

==25751==       Range "PWN(Sigmoid_926, Mul_927)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.7041ms       112  24.143us  17.057us  55.203us  PWN(Sigmoid_926, Mul_927)
 GPU activities:  100.00%  4.2209ms       112  37.686us  34.691us  51.364us  generatedNativePointwise
      API calls:  100.00%  2.1232ms       112  18.957us  13.697us  44.130us  cuLaunchKernel

==25751==       Range "PWN(Sigmoid_943, Mul_944)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.3279ms       112  20.784us  16.897us  26.722us  PWN(Sigmoid_943, Mul_944)
 GPU activities:  100.00%  6.8948ms       112  61.561us  57.477us  84.071us  generatedNativePointwise
      API calls:  100.00%  1.8574ms       112  16.584us  13.472us  21.569us  cuLaunchKernel

==25751==       Range "PWN(Sigmoid_959, Mul_960)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.6318ms       112  23.498us  17.089us  108.20us  PWN(Sigmoid_959, Mul_960)
 GPU activities:  100.00%  4.2081ms       112  37.572us  34.339us  50.948us  generatedNativePointwise
      API calls:  100.00%  2.0740ms       112  18.518us  13.377us  103.24us  cuLaunchKernel

==25751==       Range "PWN(Sigmoid_975, Mul_976)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.5247ms       112  22.542us  17.120us  96.454us  PWN(Sigmoid_975, Mul_976)
 GPU activities:  100.00%  2.0064ms       112  17.914us  16.289us  24.834us  generatedNativePointwise
      API calls:  100.00%  1.9463ms       112  17.378us  13.632us  39.042us  cuLaunchKernel

==25751==       Range "PWN(Sigmoid_995, Mul_996)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.7350ms       112  24.419us  18.081us  115.17us  PWN(Sigmoid_995, Mul_996)
 GPU activities:  100.00%  4.2091ms       112  37.581us  34.979us  51.941us  generatedNativePointwise
      API calls:  100.00%  2.1748ms       112  19.418us  14.561us  105.77us  cuLaunchKernel

==25751==       Range "QuantizeLinear_1163_quantize_scale_node_clone_0"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  4.6610ms       112  41.615us  34.914us  95.078us  QuantizeLinear_1163_quantize_scale_node_clone_0
 GPU activities:  100.00%  6.3493ms       112  56.689us  52.804us  69.925us  void CUTENSOR_NAMESPACE::permutationKernelPLC3<CUTENSOR_NAMESPACE::VectorWrite2DTensorView<unsigned char=0, unsigned char=1, bool=0, unsigned int=4, float, char, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::VectorRead2DTensorView<unsigned char=1, unsigned char=0, bool=0, unsigned int=4, float, float, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::ThreadLevelElementwise<CUTENSOR_NAMESPACE::ElementwiseConfig2DCommonCase<CUTENSOR_NAMESPACE::GeneralUnarySmall<float>, CUTENSOR_NAMESPACE::GeneralBinary<float>, int=2, int=32, int=64, int=128, char=4, bool=0, bool=0, bool=1, bool=0, bool=0>, float>, CUTENSOR_NAMESPACE::ElementwiseRuntimePLC3<float, float, char, float>::Params>(unsigned int=4)
      API calls:  100.00%  2.0184ms       112  18.021us  14.369us  54.211us  cudaLaunchKernel

==25751==       Range "QuantizeLinear_1163_quantize_scale_node_clone_1"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  5.0980ms       112  45.517us  35.746us  136.74us  QuantizeLinear_1163_quantize_scale_node_clone_1
 GPU activities:  100.00%  7.0861ms       112  63.268us  59.909us  77.894us  void CUTENSOR_NAMESPACE::permutationKernelPLC3<CUTENSOR_NAMESPACE::VectorWrite2DTensorView<unsigned char=0, unsigned char=1, bool=0, unsigned int=4, float, char, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::VectorRead2DTensorView<unsigned char=1, unsigned char=0, bool=0, unsigned int=4, float, float, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::ThreadLevelElementwise<CUTENSOR_NAMESPACE::ElementwiseConfig2DCommonCase<CUTENSOR_NAMESPACE::GeneralUnarySmall<float>, CUTENSOR_NAMESPACE::GeneralBinary<float>, int=2, int=32, int=64, int=128, char=4, bool=0, bool=0, bool=1, bool=0, bool=0>, float>, CUTENSOR_NAMESPACE::ElementwiseRuntimePLC3<float, float, char, float>::Params>(unsigned int=4)
      API calls:  100.00%  1.9908ms       112  17.774us  14.401us  29.634us  cudaLaunchKernel

==25751==       Range "QuantizeLinear_124_quantize_scale_node"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  7.7887ms       112  69.541us  38.659us  143.59us  QuantizeLinear_124_quantize_scale_node
 GPU activities:  100.00%  13.226ms       112  118.09us  98.472us  1.7740ms  void CUTENSOR_NAMESPACE::permutationKernelPLC3<CUTENSOR_NAMESPACE::VectorWrite2DTensorView<unsigned char=0, unsigned char=1, bool=0, unsigned int=4, float, char, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::VectorRead2DTensorView<unsigned char=1, unsigned char=0, bool=0, unsigned int=4, float, float, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::ThreadLevelElementwise<CUTENSOR_NAMESPACE::ElementwiseConfig2DCommonCase<CUTENSOR_NAMESPACE::GeneralUnarySmall<float>, CUTENSOR_NAMESPACE::GeneralBinary<float>, int=2, int=32, int=64, int=128, char=4, bool=0, bool=0, bool=1, bool=0, bool=0>, float>, CUTENSOR_NAMESPACE::ElementwiseRuntimePLC3<float, float, char, float>::Params>(unsigned int=4)
      API calls:  100.00%  2.8063ms       112  25.056us  15.585us  85.317us  cudaLaunchKernel

==25751==       Range "QuantizeLinear_1327_quantize_scale_node_clone_0"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  4.7946ms       112  42.809us  34.434us  221.96us  QuantizeLinear_1327_quantize_scale_node_clone_0
 GPU activities:  100.00%  10.982ms       112  98.054us  93.351us  129.23us  void CUTENSOR_NAMESPACE::permutationKernelPLC3<CUTENSOR_NAMESPACE::VectorWrite2DTensorView<unsigned char=0, unsigned char=1, bool=0, unsigned int=4, float, char, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::VectorRead2DTensorView<unsigned char=1, unsigned char=0, bool=0, unsigned int=4, float, float, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::ThreadLevelElementwise<CUTENSOR_NAMESPACE::ElementwiseConfig2DCommonCase<CUTENSOR_NAMESPACE::GeneralUnarySmall<float>, CUTENSOR_NAMESPACE::GeneralBinary<float>, int=2, int=32, int=64, int=128, char=4, bool=0, bool=0, bool=1, bool=0, bool=0>, float>, CUTENSOR_NAMESPACE::ElementwiseRuntimePLC3<float, float, char, float>::Params>(unsigned int=4)
      API calls:  100.00%  2.0194ms       112  18.030us  14.369us  74.116us  cudaLaunchKernel

==25751==       Range "QuantizeLinear_1327_quantize_scale_node_clone_1"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  4.9860ms       112  44.518us  35.394us  106.85us  QuantizeLinear_1327_quantize_scale_node_clone_1
 GPU activities:  100.00%  12.462ms       112  111.27us  102.50us  139.63us  void CUTENSOR_NAMESPACE::permutationKernelPLC3<CUTENSOR_NAMESPACE::VectorWrite2DTensorView<unsigned char=0, unsigned char=1, bool=0, unsigned int=4, float, char, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::VectorRead2DTensorView<unsigned char=1, unsigned char=0, bool=0, unsigned int=4, float, float, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::ThreadLevelElementwise<CUTENSOR_NAMESPACE::ElementwiseConfig2DCommonCase<CUTENSOR_NAMESPACE::GeneralUnarySmall<float>, CUTENSOR_NAMESPACE::GeneralBinary<float>, int=2, int=32, int=64, int=128, char=4, bool=0, bool=0, bool=1, bool=0, bool=0>, float>, CUTENSOR_NAMESPACE::ElementwiseRuntimePLC3<float, float, char, float>::Params>(unsigned int=4)
      API calls:  100.00%  2.0725ms       112  18.504us  14.529us  84.901us  cudaLaunchKernel

==25751==       Range "QuantizeLinear_1472_quantize_scale_node"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  5.3388ms       112  47.667us  36.034us  96.518us  QuantizeLinear_1472_quantize_scale_node
 GPU activities:  100.00%  12.234ms       112  109.23us  104.27us  137.10us  void CUTENSOR_NAMESPACE::permutationKernelPLC3<CUTENSOR_NAMESPACE::VectorWrite2DTensorView<unsigned char=0, unsigned char=1, bool=0, unsigned int=4, float, char, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::VectorRead2DTensorView<unsigned char=1, unsigned char=0, bool=0, unsigned int=4, float, float, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::ThreadLevelElementwise<CUTENSOR_NAMESPACE::ElementwiseConfig2DCommonCase<CUTENSOR_NAMESPACE::GeneralUnarySmall<float>, CUTENSOR_NAMESPACE::GeneralBinary<float>, int=2, int=32, int=64, int=128, char=4, bool=0, bool=0, bool=1, bool=0, bool=0>, float>, CUTENSOR_NAMESPACE::ElementwiseRuntimePLC3<float, float, char, float>::Params>(unsigned int=4)
      API calls:  100.00%  2.0443ms       112  18.252us  14.401us  46.050us  cudaLaunchKernel

==25751==       Range "QuantizeLinear_1489_quantize_scale_node_clone_1"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  5.1401ms       112  45.893us  35.266us  137.48us  QuantizeLinear_1489_quantize_scale_node_clone_1
 GPU activities:  100.00%  4.2224ms       112  37.700us  35.459us  46.691us  void CUTENSOR_NAMESPACE::permutationKernelPLC3<CUTENSOR_NAMESPACE::VectorWrite2DTensorView<unsigned char=0, unsigned char=1, bool=0, unsigned int=4, float, char, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::VectorRead2DTensorView<unsigned char=1, unsigned char=0, bool=0, unsigned int=4, float, float, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::ThreadLevelElementwise<CUTENSOR_NAMESPACE::ElementwiseConfig2DCommonCase<CUTENSOR_NAMESPACE::GeneralUnarySmall<float>, CUTENSOR_NAMESPACE::GeneralBinary<float>, int=2, int=32, int=64, int=128, char=4, bool=0, bool=0, bool=1, bool=0, bool=0>, float>, CUTENSOR_NAMESPACE::ElementwiseRuntimePLC3<float, float, char, float>::Params>(unsigned int=4)
      API calls:  100.00%  2.0957ms       112  18.711us  14.400us  67.363us  cudaLaunchKernel

==25751==       Range "QuantizeLinear_157_quantize_scale_node"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  6.4184ms       112  57.307us  37.154us  90.533us  QuantizeLinear_157_quantize_scale_node
 GPU activities:  100.00%  11.835ms       112  105.67us  101.48us  133.26us  void CUTENSOR_NAMESPACE::permutationKernelPLC3<CUTENSOR_NAMESPACE::VectorWrite2DTensorView<unsigned char=0, unsigned char=1, bool=0, unsigned int=4, float, char, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::VectorRead2DTensorView<unsigned char=1, unsigned char=0, bool=0, unsigned int=4, float, float, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::ThreadLevelElementwise<CUTENSOR_NAMESPACE::ElementwiseConfig2DCommonCase<CUTENSOR_NAMESPACE::GeneralUnarySmall<float>, CUTENSOR_NAMESPACE::GeneralBinary<float>, int=2, int=32, int=64, int=128, char=4, bool=0, bool=0, bool=1, bool=0, bool=0>, float>, CUTENSOR_NAMESPACE::ElementwiseRuntimePLC3<float, float, char, float>::Params>(unsigned int=4)
      API calls:  100.00%  2.5503ms       112  22.770us  14.753us  44.035us  cudaLaunchKernel

==25751==       Range "QuantizeLinear_1634_quantize_scale_node"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  6.1352ms       112  54.778us  38.242us  784.30us  QuantizeLinear_1634_quantize_scale_node
 GPU activities:  100.00%  6.7508ms       112  60.274us  57.061us  75.718us  void CUTENSOR_NAMESPACE::permutationKernelPLC3<CUTENSOR_NAMESPACE::VectorWrite2DTensorView<unsigned char=0, unsigned char=1, bool=0, unsigned int=4, float, char, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::VectorRead2DTensorView<unsigned char=1, unsigned char=0, bool=0, unsigned int=4, float, float, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::ThreadLevelElementwise<CUTENSOR_NAMESPACE::ElementwiseConfig2DCommonCase<CUTENSOR_NAMESPACE::GeneralUnarySmall<float>, CUTENSOR_NAMESPACE::GeneralBinary<float>, int=2, int=32, int=64, int=128, char=4, bool=0, bool=0, bool=1, bool=0, bool=0>, float>, CUTENSOR_NAMESPACE::ElementwiseRuntimePLC3<float, float, char, float>::Params>(unsigned int=4)
      API calls:  100.00%  2.0515ms       112  18.316us  15.041us  41.410us  cudaLaunchKernel

==25751==       Range "QuantizeLinear_1651_quantize_scale_node_clone_1"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  6.1196ms       112  54.639us  36.450us  1.0081ms  QuantizeLinear_1651_quantize_scale_node_clone_1
 GPU activities:  100.00%  2.8572ms       112  25.511us  23.522us  32.514us  void CUTENSOR_NAMESPACE::permutationKernelPLC3<CUTENSOR_NAMESPACE::VectorWrite2DTensorView<unsigned char=0, unsigned char=1, bool=0, unsigned int=4, float, char, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::VectorRead2DTensorView<unsigned char=1, unsigned char=0, bool=0, unsigned int=4, float, float, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::ThreadLevelElementwise<CUTENSOR_NAMESPACE::ElementwiseConfig2DCommonCase<CUTENSOR_NAMESPACE::GeneralUnarySmall<float>, CUTENSOR_NAMESPACE::GeneralBinary<float>, int=2, int=32, int=64, int=128, char=4, bool=0, bool=0, bool=1, bool=0, bool=0>, float>, CUTENSOR_NAMESPACE::ElementwiseRuntimePLC3<float, float, char, float>::Params>(unsigned int=4)
      API calls:  100.00%  2.8701ms       112  25.626us  14.560us  979.77us  cudaLaunchKernel

==25751==       Range "QuantizeLinear_255_quantize_scale_node"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  7.4511ms       112  66.528us  37.282us  157.80us  QuantizeLinear_255_quantize_scale_node
 GPU activities:  100.00%  6.7892ms       112  60.618us  57.573us  76.486us  void CUTENSOR_NAMESPACE::permutationKernelPLC3<CUTENSOR_NAMESPACE::VectorWrite2DTensorView<unsigned char=0, unsigned char=1, bool=0, unsigned int=4, float, char, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::VectorRead2DTensorView<unsigned char=1, unsigned char=0, bool=0, unsigned int=4, float, float, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::ThreadLevelElementwise<CUTENSOR_NAMESPACE::ElementwiseConfig2DCommonCase<CUTENSOR_NAMESPACE::GeneralUnarySmall<float>, CUTENSOR_NAMESPACE::GeneralBinary<float>, int=2, int=32, int=64, int=128, char=4, bool=0, bool=0, bool=1, bool=0, bool=0>, float>, CUTENSOR_NAMESPACE::ElementwiseRuntimePLC3<float, float, char, float>::Params>(unsigned int=4)
      API calls:  100.00%  2.7164ms       112  24.253us  14.721us  60.579us  cudaLaunchKernel

==25751==       Range "QuantizeLinear_288_quantize_scale_node"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  6.5828ms       112  58.775us  37.442us  154.28us  QuantizeLinear_288_quantize_scale_node
 GPU activities:  100.00%  6.6259ms       112  59.160us  55.301us  75.814us  void CUTENSOR_NAMESPACE::permutationKernelPLC3<CUTENSOR_NAMESPACE::VectorWrite2DTensorView<unsigned char=0, unsigned char=1, bool=0, unsigned int=4, float, char, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::VectorRead2DTensorView<unsigned char=1, unsigned char=0, bool=0, unsigned int=4, float, float, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::ThreadLevelElementwise<CUTENSOR_NAMESPACE::ElementwiseConfig2DCommonCase<CUTENSOR_NAMESPACE::GeneralUnarySmall<float>, CUTENSOR_NAMESPACE::GeneralBinary<float>, int=2, int=32, int=64, int=128, char=4, bool=0, bool=0, bool=1, bool=0, bool=0>, float>, CUTENSOR_NAMESPACE::ElementwiseRuntimePLC3<float, float, char, float>::Params>(unsigned int=4)
      API calls:  100.00%  2.7334ms       112  24.405us  15.489us  123.46us  cudaLaunchKernel

==25751==       Range "QuantizeLinear_321_quantize_scale_node"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  6.5406ms       112  58.398us  36.706us  108.65us  QuantizeLinear_321_quantize_scale_node
 GPU activities:  100.00%  6.5211ms       112  58.223us  54.084us  74.149us  void CUTENSOR_NAMESPACE::permutationKernelPLC3<CUTENSOR_NAMESPACE::VectorWrite2DTensorView<unsigned char=0, unsigned char=1, bool=0, unsigned int=4, float, char, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::VectorRead2DTensorView<unsigned char=1, unsigned char=0, bool=0, unsigned int=4, float, float, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::ThreadLevelElementwise<CUTENSOR_NAMESPACE::ElementwiseConfig2DCommonCase<CUTENSOR_NAMESPACE::GeneralUnarySmall<float>, CUTENSOR_NAMESPACE::GeneralBinary<float>, int=2, int=32, int=64, int=128, char=4, bool=0, bool=0, bool=1, bool=0, bool=0>, float>, CUTENSOR_NAMESPACE::ElementwiseRuntimePLC3<float, float, char, float>::Params>(unsigned int=4)
      API calls:  100.00%  2.5763ms       112  23.002us  15.265us  47.042us  cudaLaunchKernel

==25751==       Range "QuantizeLinear_354_quantize_scale_node"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  6.2808ms       112  56.078us  35.938us  137.29us  QuantizeLinear_354_quantize_scale_node
 GPU activities:  100.00%  6.5032ms       112  58.064us  54.852us  77.318us  void CUTENSOR_NAMESPACE::permutationKernelPLC3<CUTENSOR_NAMESPACE::VectorWrite2DTensorView<unsigned char=0, unsigned char=1, bool=0, unsigned int=4, float, char, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::VectorRead2DTensorView<unsigned char=1, unsigned char=0, bool=0, unsigned int=4, float, float, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::ThreadLevelElementwise<CUTENSOR_NAMESPACE::ElementwiseConfig2DCommonCase<CUTENSOR_NAMESPACE::GeneralUnarySmall<float>, CUTENSOR_NAMESPACE::GeneralBinary<float>, int=2, int=32, int=64, int=128, char=4, bool=0, bool=0, bool=1, bool=0, bool=0>, float>, CUTENSOR_NAMESPACE::ElementwiseRuntimePLC3<float, float, char, float>::Params>(unsigned int=4)
      API calls:  100.00%  2.6070ms       112  23.276us  14.561us  102.28us  cudaLaunchKernel

==25751==       Range "QuantizeLinear_387_quantize_scale_node"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  6.6024ms       112  58.949us  36.866us  104.07us  QuantizeLinear_387_quantize_scale_node
 GPU activities:  100.00%  6.6100ms       112  59.017us  55.652us  76.582us  void CUTENSOR_NAMESPACE::permutationKernelPLC3<CUTENSOR_NAMESPACE::VectorWrite2DTensorView<unsigned char=0, unsigned char=1, bool=0, unsigned int=4, float, char, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::VectorRead2DTensorView<unsigned char=1, unsigned char=0, bool=0, unsigned int=4, float, float, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::ThreadLevelElementwise<CUTENSOR_NAMESPACE::ElementwiseConfig2DCommonCase<CUTENSOR_NAMESPACE::GeneralUnarySmall<float>, CUTENSOR_NAMESPACE::GeneralBinary<float>, int=2, int=32, int=64, int=128, char=4, bool=0, bool=0, bool=1, bool=0, bool=0>, float>, CUTENSOR_NAMESPACE::ElementwiseRuntimePLC3<float, float, char, float>::Params>(unsigned int=4)
      API calls:  100.00%  2.5963ms       112  23.181us  15.457us  69.572us  cudaLaunchKernel

==25751==       Range "QuantizeLinear_420_quantize_scale_node"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  6.2444ms       112  55.753us  36.866us  140.33us  QuantizeLinear_420_quantize_scale_node
 GPU activities:  100.00%  6.5736ms       112  58.693us  54.724us  75.078us  void CUTENSOR_NAMESPACE::permutationKernelPLC3<CUTENSOR_NAMESPACE::VectorWrite2DTensorView<unsigned char=0, unsigned char=1, bool=0, unsigned int=4, float, char, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::VectorRead2DTensorView<unsigned char=1, unsigned char=0, bool=0, unsigned int=4, float, float, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::ThreadLevelElementwise<CUTENSOR_NAMESPACE::ElementwiseConfig2DCommonCase<CUTENSOR_NAMESPACE::GeneralUnarySmall<float>, CUTENSOR_NAMESPACE::GeneralBinary<float>, int=2, int=32, int=64, int=128, char=4, bool=0, bool=0, bool=1, bool=0, bool=0>, float>, CUTENSOR_NAMESPACE::ElementwiseRuntimePLC3<float, float, char, float>::Params>(unsigned int=4)
      API calls:  100.00%  2.5358ms       112  22.641us  14.817us  53.251us  cudaLaunchKernel

==25751==       Range "QuantizeLinear_43_quantize_scale_node_clone_0"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  6.5699ms       112  58.659us  35.426us  112.93us  QuantizeLinear_43_quantize_scale_node_clone_0
 GPU activities:  100.00%  2.1699ms       112  19.373us  17.474us  51.140us  void CUTENSOR_NAMESPACE::permutationKernelPLC3<CUTENSOR_NAMESPACE::VectorWrite2DTensorView<unsigned char=0, unsigned char=1, bool=1, unsigned int=4, float, char, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::VectorRead2DTensorView<unsigned char=0, unsigned char=1, bool=1, unsigned int=4, float, float, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::ThreadLevelElementwise<CUTENSOR_NAMESPACE::ElementwiseConfig1DCommonCase<CUTENSOR_NAMESPACE::GeneralUnarySmall<float>, CUTENSOR_NAMESPACE::GeneralBinary<float>, int=1, int=2048, int=1, int=128, char=4, bool=1, bool=0, bool=1, bool=0, bool=0>, float>, CUTENSOR_NAMESPACE::ElementwiseRuntimePLC3<float, float, char, float>::Params>(unsigned int=4)
      API calls:  100.00%  2.6678ms       112  23.819us  14.625us  74.852us  cudaLaunchKernel

==25751==       Range "QuantizeLinear_43_quantize_scale_node_clone_1"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  6.7246ms       112  60.041us  36.450us  111.56us  QuantizeLinear_43_quantize_scale_node_clone_1
 GPU activities:  100.00%  2.1681ms       112  19.358us  17.441us  50.756us  void CUTENSOR_NAMESPACE::permutationKernelPLC3<CUTENSOR_NAMESPACE::VectorWrite2DTensorView<unsigned char=0, unsigned char=1, bool=1, unsigned int=4, float, char, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::VectorRead2DTensorView<unsigned char=0, unsigned char=1, bool=1, unsigned int=4, float, float, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::ThreadLevelElementwise<CUTENSOR_NAMESPACE::ElementwiseConfig1DCommonCase<CUTENSOR_NAMESPACE::GeneralUnarySmall<float>, CUTENSOR_NAMESPACE::GeneralBinary<float>, int=1, int=2048, int=1, int=128, char=4, bool=1, bool=0, bool=1, bool=0, bool=0>, float>, CUTENSOR_NAMESPACE::ElementwiseRuntimePLC3<float, float, char, float>::Params>(unsigned int=4)
      API calls:  100.00%  2.7965ms       112  24.968us  14.689us  71.780us  cudaLaunchKernel

==25751==       Range "QuantizeLinear_43_quantize_scale_node_clone_2"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  8.1308ms       112  72.596us  46.083us  132.26us  QuantizeLinear_43_quantize_scale_node_clone_2
 GPU activities:  100.00%  2.7980ms       112  24.981us  22.785us  57.540us  void CUTENSOR_NAMESPACE::permutationKernelPLC3<CUTENSOR_NAMESPACE::VectorWrite2DTensorView<unsigned char=0, unsigned char=1, bool=1, unsigned int=4, float, char, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::VectorRead2DTensorView<unsigned char=0, unsigned char=1, bool=1, unsigned int=4, float, float, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::ThreadLevelElementwise<CUTENSOR_NAMESPACE::ElementwiseConfig1DCommonCase<CUTENSOR_NAMESPACE::GeneralUnarySmall<float>, CUTENSOR_NAMESPACE::GeneralBinary<float>, int=1, int=2048, int=1, int=128, char=4, bool=1, bool=0, bool=1, bool=0, bool=0>, float>, CUTENSOR_NAMESPACE::ElementwiseRuntimePLC3<float, float, char, float>::Params>(unsigned int=4)
      API calls:  100.00%  3.2082ms       112  28.645us  18.689us  82.533us  cudaLaunchKernel

==25751==       Range "QuantizeLinear_43_quantize_scale_node_clone_3"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  11.453ms       112  102.26us  61.060us  215.85us  QuantizeLinear_43_quantize_scale_node_clone_3
 GPU activities:  100.00%  4.1140ms       112  36.732us  34.659us  101.96us  void CUTENSOR_NAMESPACE::permutationKernelPLC3<CUTENSOR_NAMESPACE::VectorWrite2DTensorView<unsigned char=0, unsigned char=1, bool=0, unsigned int=4, float, char, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::VectorRead2DTensorView<unsigned char=1, unsigned char=0, bool=0, unsigned int=4, float, float, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::ThreadLevelElementwise<CUTENSOR_NAMESPACE::ElementwiseConfig2DSmallMode0<CUTENSOR_NAMESPACE::GeneralUnarySmall<float>, CUTENSOR_NAMESPACE::GeneralBinary<float>, int=2, int=4, int=256, int=128, char=4, bool=0, bool=0, bool=1, bool=0, bool=1>, float>, CUTENSOR_NAMESPACE::ElementwiseRuntimePLC3<float, float, char, float>::Params>(unsigned int=4)
      API calls:  100.00%  3.5573ms       112  31.761us  19.137us  150.60us  cudaLaunchKernel

==25751==       Range "QuantizeLinear_453_quantize_scale_node"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  6.6674ms       112  59.530us  35.906us  152.01us  QuantizeLinear_453_quantize_scale_node
 GPU activities:  100.00%  6.6354ms       112  59.244us  55.876us  76.006us  void CUTENSOR_NAMESPACE::permutationKernelPLC3<CUTENSOR_NAMESPACE::VectorWrite2DTensorView<unsigned char=0, unsigned char=1, bool=0, unsigned int=4, float, char, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::VectorRead2DTensorView<unsigned char=1, unsigned char=0, bool=0, unsigned int=4, float, float, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::ThreadLevelElementwise<CUTENSOR_NAMESPACE::ElementwiseConfig2DCommonCase<CUTENSOR_NAMESPACE::GeneralUnarySmall<float>, CUTENSOR_NAMESPACE::GeneralBinary<float>, int=2, int=32, int=64, int=128, char=4, bool=0, bool=0, bool=1, bool=0, bool=0>, float>, CUTENSOR_NAMESPACE::ElementwiseRuntimePLC3<float, float, char, float>::Params>(unsigned int=4)
      API calls:  100.00%  2.5793ms       112  23.029us  14.945us  62.436us  cudaLaunchKernel

==25751==       Range "QuantizeLinear_486_quantize_scale_node"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  6.0822ms       112  54.305us  36.162us  88.197us  QuantizeLinear_486_quantize_scale_node
 GPU activities:  100.00%  6.5775ms       112  58.727us  55.620us  76.518us  void CUTENSOR_NAMESPACE::permutationKernelPLC3<CUTENSOR_NAMESPACE::VectorWrite2DTensorView<unsigned char=0, unsigned char=1, bool=0, unsigned int=4, float, char, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::VectorRead2DTensorView<unsigned char=1, unsigned char=0, bool=0, unsigned int=4, float, float, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::ThreadLevelElementwise<CUTENSOR_NAMESPACE::ElementwiseConfig2DCommonCase<CUTENSOR_NAMESPACE::GeneralUnarySmall<float>, CUTENSOR_NAMESPACE::GeneralBinary<float>, int=2, int=32, int=64, int=128, char=4, bool=0, bool=0, bool=1, bool=0, bool=0>, float>, CUTENSOR_NAMESPACE::ElementwiseRuntimePLC3<float, float, char, float>::Params>(unsigned int=4)
      API calls:  100.00%  2.5316ms       112  22.603us  14.944us  53.699us  cudaLaunchKernel

==25751==       Range "QuantizeLinear_519_quantize_scale_node"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  6.5784ms       112  58.735us  36.738us  139.24us  QuantizeLinear_519_quantize_scale_node
 GPU activities:  100.00%  6.6148ms       112  59.061us  54.949us  74.950us  void CUTENSOR_NAMESPACE::permutationKernelPLC3<CUTENSOR_NAMESPACE::VectorWrite2DTensorView<unsigned char=0, unsigned char=1, bool=0, unsigned int=4, float, char, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::VectorRead2DTensorView<unsigned char=1, unsigned char=0, bool=0, unsigned int=4, float, float, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::ThreadLevelElementwise<CUTENSOR_NAMESPACE::ElementwiseConfig2DCommonCase<CUTENSOR_NAMESPACE::GeneralUnarySmall<float>, CUTENSOR_NAMESPACE::GeneralBinary<float>, int=2, int=32, int=64, int=128, char=4, bool=0, bool=0, bool=1, bool=0, bool=0>, float>, CUTENSOR_NAMESPACE::ElementwiseRuntimePLC3<float, float, char, float>::Params>(unsigned int=4)
      API calls:  100.00%  2.5680ms       112  22.928us  14.593us  55.331us  cudaLaunchKernel

==25751==       Range "QuantizeLinear_585_quantize_scale_node"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  6.3642ms       112  56.823us  36.674us  138.09us  QuantizeLinear_585_quantize_scale_node
 GPU activities:  100.00%  12.422ms       112  110.91us  104.81us  137.74us  void CUTENSOR_NAMESPACE::permutationKernelPLC3<CUTENSOR_NAMESPACE::VectorWrite2DTensorView<unsigned char=0, unsigned char=1, bool=0, unsigned int=4, float, char, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::VectorRead2DTensorView<unsigned char=1, unsigned char=0, bool=0, unsigned int=4, float, float, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::ThreadLevelElementwise<CUTENSOR_NAMESPACE::ElementwiseConfig2DCommonCase<CUTENSOR_NAMESPACE::GeneralUnarySmall<float>, CUTENSOR_NAMESPACE::GeneralBinary<float>, int=2, int=32, int=64, int=128, char=4, bool=0, bool=0, bool=1, bool=0, bool=0>, float>, CUTENSOR_NAMESPACE::ElementwiseRuntimePLC3<float, float, char, float>::Params>(unsigned int=4)
      API calls:  100.00%  2.6029ms       112  23.240us  14.689us  92.101us  cudaLaunchKernel

==25751==       Range "QuantizeLinear_617_quantize_scale_node"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  5.9231ms       112  52.885us  36.642us  123.43us  QuantizeLinear_617_quantize_scale_node
 GPU activities:  100.00%  4.2316ms       112  37.782us  35.651us  47.491us  void CUTENSOR_NAMESPACE::permutationKernelPLC3<CUTENSOR_NAMESPACE::VectorWrite2DTensorView<unsigned char=0, unsigned char=1, bool=0, unsigned int=4, float, char, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::VectorRead2DTensorView<unsigned char=1, unsigned char=0, bool=0, unsigned int=4, float, float, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::ThreadLevelElementwise<CUTENSOR_NAMESPACE::ElementwiseConfig2DCommonCase<CUTENSOR_NAMESPACE::GeneralUnarySmall<float>, CUTENSOR_NAMESPACE::GeneralBinary<float>, int=2, int=32, int=64, int=128, char=4, bool=0, bool=0, bool=1, bool=0, bool=0>, float>, CUTENSOR_NAMESPACE::ElementwiseRuntimePLC3<float, float, char, float>::Params>(unsigned int=4)
      API calls:  100.00%  2.4457ms       112  21.836us  14.785us  67.108us  cudaLaunchKernel

==25751==       Range "QuantizeLinear_650_quantize_scale_node"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  6.2508ms       112  55.810us  36.002us  101.73us  QuantizeLinear_650_quantize_scale_node
 GPU activities:  100.00%  4.2067ms       112  37.559us  35.170us  46.628us  void CUTENSOR_NAMESPACE::permutationKernelPLC3<CUTENSOR_NAMESPACE::VectorWrite2DTensorView<unsigned char=0, unsigned char=1, bool=0, unsigned int=4, float, char, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::VectorRead2DTensorView<unsigned char=1, unsigned char=0, bool=0, unsigned int=4, float, float, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::ThreadLevelElementwise<CUTENSOR_NAMESPACE::ElementwiseConfig2DCommonCase<CUTENSOR_NAMESPACE::GeneralUnarySmall<float>, CUTENSOR_NAMESPACE::GeneralBinary<float>, int=2, int=32, int=64, int=128, char=4, bool=0, bool=0, bool=1, bool=0, bool=0>, float>, CUTENSOR_NAMESPACE::ElementwiseRuntimePLC3<float, float, char, float>::Params>(unsigned int=4)
      API calls:  100.00%  2.5007ms       112  22.327us  15.041us  60.995us  cudaLaunchKernel

==25751==       Range "QuantizeLinear_683_quantize_scale_node"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  5.9892ms       112  53.474us  37.154us  187.98us  QuantizeLinear_683_quantize_scale_node
 GPU activities:  100.00%  4.1138ms       112  36.730us  34.946us  47.172us  void CUTENSOR_NAMESPACE::permutationKernelPLC3<CUTENSOR_NAMESPACE::VectorWrite2DTensorView<unsigned char=0, unsigned char=1, bool=0, unsigned int=4, float, char, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::VectorRead2DTensorView<unsigned char=1, unsigned char=0, bool=0, unsigned int=4, float, float, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::ThreadLevelElementwise<CUTENSOR_NAMESPACE::ElementwiseConfig2DCommonCase<CUTENSOR_NAMESPACE::GeneralUnarySmall<float>, CUTENSOR_NAMESPACE::GeneralBinary<float>, int=2, int=32, int=64, int=128, char=4, bool=0, bool=0, bool=1, bool=0, bool=0>, float>, CUTENSOR_NAMESPACE::ElementwiseRuntimePLC3<float, float, char, float>::Params>(unsigned int=4)
      API calls:  100.00%  2.4331ms       112  21.724us  15.105us  124.10us  cudaLaunchKernel

==25751==       Range "QuantizeLinear_716_quantize_scale_node"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  6.3337ms       112  56.550us  35.970us  145.70us  QuantizeLinear_716_quantize_scale_node
 GPU activities:  100.00%  4.2237ms       112  37.711us  35.618us  46.915us  void CUTENSOR_NAMESPACE::permutationKernelPLC3<CUTENSOR_NAMESPACE::VectorWrite2DTensorView<unsigned char=0, unsigned char=1, bool=0, unsigned int=4, float, char, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::VectorRead2DTensorView<unsigned char=1, unsigned char=0, bool=0, unsigned int=4, float, float, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::ThreadLevelElementwise<CUTENSOR_NAMESPACE::ElementwiseConfig2DCommonCase<CUTENSOR_NAMESPACE::GeneralUnarySmall<float>, CUTENSOR_NAMESPACE::GeneralBinary<float>, int=2, int=32, int=64, int=128, char=4, bool=0, bool=0, bool=1, bool=0, bool=0>, float>, CUTENSOR_NAMESPACE::ElementwiseRuntimePLC3<float, float, char, float>::Params>(unsigned int=4)
      API calls:  100.00%  2.5087ms       112  22.399us  14.817us  121.96us  cudaLaunchKernel

==25751==       Range "QuantizeLinear_749_quantize_scale_node"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  5.6592ms       112  50.528us  36.067us  91.045us  QuantizeLinear_749_quantize_scale_node
 GPU activities:  100.00%  4.1462ms       112  37.019us  34.979us  47.428us  void CUTENSOR_NAMESPACE::permutationKernelPLC3<CUTENSOR_NAMESPACE::VectorWrite2DTensorView<unsigned char=0, unsigned char=1, bool=0, unsigned int=4, float, char, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::VectorRead2DTensorView<unsigned char=1, unsigned char=0, bool=0, unsigned int=4, float, float, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::ThreadLevelElementwise<CUTENSOR_NAMESPACE::ElementwiseConfig2DCommonCase<CUTENSOR_NAMESPACE::GeneralUnarySmall<float>, CUTENSOR_NAMESPACE::GeneralBinary<float>, int=2, int=32, int=64, int=128, char=4, bool=0, bool=0, bool=1, bool=0, bool=0>, float>, CUTENSOR_NAMESPACE::ElementwiseRuntimePLC3<float, float, char, float>::Params>(unsigned int=4)
      API calls:  100.00%  2.3834ms       112  21.280us  14.881us  49.283us  cudaLaunchKernel

==25751==       Range "QuantizeLinear_782_quantize_scale_node"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  5.5730ms       112  49.758us  35.554us  115.78us  QuantizeLinear_782_quantize_scale_node
 GPU activities:  100.00%  4.2015ms       112  37.513us  35.363us  46.563us  void CUTENSOR_NAMESPACE::permutationKernelPLC3<CUTENSOR_NAMESPACE::VectorWrite2DTensorView<unsigned char=0, unsigned char=1, bool=0, unsigned int=4, float, char, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::VectorRead2DTensorView<unsigned char=1, unsigned char=0, bool=0, unsigned int=4, float, float, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::ThreadLevelElementwise<CUTENSOR_NAMESPACE::ElementwiseConfig2DCommonCase<CUTENSOR_NAMESPACE::GeneralUnarySmall<float>, CUTENSOR_NAMESPACE::GeneralBinary<float>, int=2, int=32, int=64, int=128, char=4, bool=0, bool=0, bool=1, bool=0, bool=0>, float>, CUTENSOR_NAMESPACE::ElementwiseRuntimePLC3<float, float, char, float>::Params>(unsigned int=4)
      API calls:  100.00%  2.2178ms       112  19.802us  14.849us  37.154us  cudaLaunchKernel

==25751==       Range "QuantizeLinear_815_quantize_scale_node"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  5.2920ms       112  47.249us  35.682us  101.48us  QuantizeLinear_815_quantize_scale_node
 GPU activities:  100.00%  4.1131ms       112  36.723us  34.403us  48.196us  void CUTENSOR_NAMESPACE::permutationKernelPLC3<CUTENSOR_NAMESPACE::VectorWrite2DTensorView<unsigned char=0, unsigned char=1, bool=0, unsigned int=4, float, char, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::VectorRead2DTensorView<unsigned char=1, unsigned char=0, bool=0, unsigned int=4, float, float, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::ThreadLevelElementwise<CUTENSOR_NAMESPACE::ElementwiseConfig2DCommonCase<CUTENSOR_NAMESPACE::GeneralUnarySmall<float>, CUTENSOR_NAMESPACE::GeneralBinary<float>, int=2, int=32, int=64, int=128, char=4, bool=0, bool=0, bool=1, bool=0, bool=0>, float>, CUTENSOR_NAMESPACE::ElementwiseRuntimePLC3<float, float, char, float>::Params>(unsigned int=4)
      API calls:  100.00%  2.2020ms       112  19.660us  14.752us  42.402us  cudaLaunchKernel

==25751==       Range "QuantizeLinear_848_quantize_scale_node"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  5.4321ms       112  48.501us  36.002us  82.021us  QuantizeLinear_848_quantize_scale_node
 GPU activities:  100.00%  4.1745ms       112  37.272us  34.659us  47.652us  void CUTENSOR_NAMESPACE::permutationKernelPLC3<CUTENSOR_NAMESPACE::VectorWrite2DTensorView<unsigned char=0, unsigned char=1, bool=0, unsigned int=4, float, char, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::VectorRead2DTensorView<unsigned char=1, unsigned char=0, bool=0, unsigned int=4, float, float, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::ThreadLevelElementwise<CUTENSOR_NAMESPACE::ElementwiseConfig2DCommonCase<CUTENSOR_NAMESPACE::GeneralUnarySmall<float>, CUTENSOR_NAMESPACE::GeneralBinary<float>, int=2, int=32, int=64, int=128, char=4, bool=0, bool=0, bool=1, bool=0, bool=0>, float>, CUTENSOR_NAMESPACE::ElementwiseRuntimePLC3<float, float, char, float>::Params>(unsigned int=4)
      API calls:  100.00%  2.1500ms       112  19.196us  14.817us  39.458us  cudaLaunchKernel

==25751==       Range "QuantizeLinear_881_quantize_scale_node"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  5.2753ms       112  47.100us  36.067us  89.574us  QuantizeLinear_881_quantize_scale_node
 GPU activities:  100.00%  4.1357ms       112  36.926us  34.050us  48.643us  void CUTENSOR_NAMESPACE::permutationKernelPLC3<CUTENSOR_NAMESPACE::VectorWrite2DTensorView<unsigned char=0, unsigned char=1, bool=0, unsigned int=4, float, char, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::VectorRead2DTensorView<unsigned char=1, unsigned char=0, bool=0, unsigned int=4, float, float, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::ThreadLevelElementwise<CUTENSOR_NAMESPACE::ElementwiseConfig2DCommonCase<CUTENSOR_NAMESPACE::GeneralUnarySmall<float>, CUTENSOR_NAMESPACE::GeneralBinary<float>, int=2, int=32, int=64, int=128, char=4, bool=0, bool=0, bool=1, bool=0, bool=0>, float>, CUTENSOR_NAMESPACE::ElementwiseRuntimePLC3<float, float, char, float>::Params>(unsigned int=4)
      API calls:  100.00%  2.1525ms       112  19.218us  14.720us  39.971us  cudaLaunchKernel

==25751==       Range "QuantizeLinear_91_quantize_scale_node"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  8.5505ms       112  76.344us  40.802us  184.30us  QuantizeLinear_91_quantize_scale_node
 GPU activities:  100.00%  12.957ms       112  115.69us  100.49us  1.1899ms  void CUTENSOR_NAMESPACE::permutationKernelPLC3<CUTENSOR_NAMESPACE::VectorWrite2DTensorView<unsigned char=0, unsigned char=1, bool=0, unsigned int=4, float, char, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::VectorRead2DTensorView<unsigned char=1, unsigned char=0, bool=0, unsigned int=4, float, float, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::ThreadLevelElementwise<CUTENSOR_NAMESPACE::ElementwiseConfig2DCommonCase<CUTENSOR_NAMESPACE::GeneralUnarySmall<float>, CUTENSOR_NAMESPACE::GeneralBinary<float>, int=2, int=32, int=64, int=128, char=4, bool=0, bool=0, bool=1, bool=0, bool=0>, float>, CUTENSOR_NAMESPACE::ElementwiseRuntimePLC3<float, float, char, float>::Params>(unsigned int=4)
      API calls:  100.00%  3.1705ms       112  28.308us  17.377us  121.03us  cudaLaunchKernel

==25751==       Range "QuantizeLinear_947_quantize_scale_node"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  5.9895ms       112  53.477us  36.802us  180.49us  QuantizeLinear_947_quantize_scale_node
 GPU activities:  100.00%  7.0067ms       112  62.560us  59.621us  78.278us  void CUTENSOR_NAMESPACE::permutationKernelPLC3<CUTENSOR_NAMESPACE::VectorWrite2DTensorView<unsigned char=0, unsigned char=1, bool=0, unsigned int=4, float, char, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::VectorRead2DTensorView<unsigned char=1, unsigned char=0, bool=0, unsigned int=4, float, float, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::ThreadLevelElementwise<CUTENSOR_NAMESPACE::ElementwiseConfig2DCommonCase<CUTENSOR_NAMESPACE::GeneralUnarySmall<float>, CUTENSOR_NAMESPACE::GeneralBinary<float>, int=2, int=32, int=64, int=128, char=4, bool=0, bool=0, bool=1, bool=0, bool=0>, float>, CUTENSOR_NAMESPACE::ElementwiseRuntimePLC3<float, float, char, float>::Params>(unsigned int=4)
      API calls:  100.00%  2.1849ms       112  19.508us  14.529us  95.077us  cudaLaunchKernel

==25751==       Range "QuantizeLinear_983_quantize_scale_node_clone_0"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  4.8191ms       112  43.027us  34.530us  128.33us  QuantizeLinear_983_quantize_scale_node_clone_0
 GPU activities:  100.00%  2.3696ms       112  21.157us  19.490us  28.099us  void CUTENSOR_NAMESPACE::permutationKernelPLC3<CUTENSOR_NAMESPACE::VectorWrite2DTensorView<unsigned char=0, unsigned char=1, bool=0, unsigned int=4, float, char, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::VectorRead2DTensorView<unsigned char=1, unsigned char=0, bool=0, unsigned int=4, float, float, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::ThreadLevelElementwise<CUTENSOR_NAMESPACE::ElementwiseConfig2DCommonCase<CUTENSOR_NAMESPACE::GeneralUnarySmall<float>, CUTENSOR_NAMESPACE::GeneralBinary<float>, int=2, int=32, int=64, int=128, char=4, bool=0, bool=0, bool=1, bool=0, bool=0>, float>, CUTENSOR_NAMESPACE::ElementwiseRuntimePLC3<float, float, char, float>::Params>(unsigned int=4)
      API calls:  100.00%  2.0542ms       112  18.341us  14.241us  48.834us  cudaLaunchKernel

==25751==       Range "QuantizeLinear_983_quantize_scale_node_clone_1"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  4.9587ms       112  44.274us  35.138us  178.03us  QuantizeLinear_983_quantize_scale_node_clone_1
 GPU activities:  100.00%  2.3687ms       112  21.149us  19.490us  27.842us  void CUTENSOR_NAMESPACE::permutationKernelPLC3<CUTENSOR_NAMESPACE::VectorWrite2DTensorView<unsigned char=0, unsigned char=1, bool=0, unsigned int=4, float, char, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::VectorRead2DTensorView<unsigned char=1, unsigned char=0, bool=0, unsigned int=4, float, float, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::ThreadLevelElementwise<CUTENSOR_NAMESPACE::ElementwiseConfig2DCommonCase<CUTENSOR_NAMESPACE::GeneralUnarySmall<float>, CUTENSOR_NAMESPACE::GeneralBinary<float>, int=2, int=32, int=64, int=128, char=4, bool=0, bool=0, bool=1, bool=0, bool=0>, float>, CUTENSOR_NAMESPACE::ElementwiseRuntimePLC3<float, float, char, float>::Params>(unsigned int=4)
      API calls:  100.00%  2.1393ms       112  19.101us  14.337us  102.92us  cudaLaunchKernel

==25751==       Range "QuantizeLinear_983_quantize_scale_node_clone_2"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  5.2317ms       112  46.711us  34.626us  92.485us  QuantizeLinear_983_quantize_scale_node_clone_2
 GPU activities:  100.00%  2.3725ms       112  21.183us  19.778us  28.131us  void CUTENSOR_NAMESPACE::permutationKernelPLC3<CUTENSOR_NAMESPACE::VectorWrite2DTensorView<unsigned char=0, unsigned char=1, bool=0, unsigned int=4, float, char, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::VectorRead2DTensorView<unsigned char=1, unsigned char=0, bool=0, unsigned int=4, float, float, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::ThreadLevelElementwise<CUTENSOR_NAMESPACE::ElementwiseConfig2DCommonCase<CUTENSOR_NAMESPACE::GeneralUnarySmall<float>, CUTENSOR_NAMESPACE::GeneralBinary<float>, int=2, int=32, int=64, int=128, char=4, bool=0, bool=0, bool=1, bool=0, bool=0>, float>, CUTENSOR_NAMESPACE::ElementwiseRuntimePLC3<float, float, char, float>::Params>(unsigned int=4)
      API calls:  100.00%  2.1708ms       112  19.381us  14.241us  64.676us  cudaLaunchKernel

==25751==       Range "QuantizeLinear_983_quantize_scale_node_clone_3"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  5.4574ms       112  48.727us  37.218us  99.014us  QuantizeLinear_983_quantize_scale_node_clone_3
 GPU activities:  100.00%  2.8684ms       112  25.610us  23.906us  32.867us  void CUTENSOR_NAMESPACE::permutationKernelPLC3<CUTENSOR_NAMESPACE::VectorWrite2DTensorView<unsigned char=0, unsigned char=1, bool=0, unsigned int=4, float, char, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::VectorRead2DTensorView<unsigned char=1, unsigned char=0, bool=0, unsigned int=4, float, float, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::ThreadLevelElementwise<CUTENSOR_NAMESPACE::ElementwiseConfig2DCommonCase<CUTENSOR_NAMESPACE::GeneralUnarySmall<float>, CUTENSOR_NAMESPACE::GeneralBinary<float>, int=2, int=32, int=64, int=128, char=4, bool=0, bool=0, bool=1, bool=0, bool=0>, float>, CUTENSOR_NAMESPACE::ElementwiseRuntimePLC3<float, float, char, float>::Params>(unsigned int=4)
      API calls:  100.00%  2.2216ms       112  19.835us  14.689us  51.747us  cudaLaunchKernel

==25751==       Range "Range_1813"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  5.9168ms       112  52.828us  20.417us  2.6134ms  Range_1813
 GPU activities:  100.00%  4.5313ms       112  40.458us  39.043us  51.620us  void cuFillLayer::fill<int>(cuFillLayer::KernelArgs<int>)
      API calls:  100.00%  4.9825ms       112  44.486us  15.169us  2.6013ms  cudaLaunchKernel

==25751==       Range "Range_1817"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.1336ms       112  19.049us  15.393us  60.004us  Range_1817
 GPU activities:  100.00%  1.8370ms       112  16.402us  15.330us  23.042us  void cuFillLayer::fill<int>(cuFillLayer::KernelArgs<int>)
      API calls:  100.00%  1.7449ms       112  15.579us  12.577us  43.587us  cudaLaunchKernel

==25751==       Range "Range_1912"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.8308ms       112  25.274us  17.249us  362.49us  Range_1912
 GPU activities:  100.00%  3.8652ms       112  34.510us  33.378us  44.483us  void cuFillLayer::fill<int>(cuFillLayer::KernelArgs<int>)
      API calls:  100.00%  2.2645ms       112  20.219us  13.728us  337.24us  cudaLaunchKernel

==25751==       Range "Range_1916"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.1625ms       112  19.308us  15.457us  67.812us  Range_1916
 GPU activities:  100.00%  1.7214ms       112  15.369us  14.690us  24.386us  void cuFillLayer::fill<int>(cuFillLayer::KernelArgs<int>)
      API calls:  100.00%  1.7565ms       112  15.683us  12.384us  62.019us  cudaLaunchKernel

==25751==       Range "Range_2011"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.4766ms       112  22.112us  16.865us  99.366us  Range_2011
 GPU activities:  100.00%  3.7726ms       112  33.684us  32.706us  43.044us  void cuFillLayer::fill<int>(cuFillLayer::KernelArgs<int>)
      API calls:  100.00%  1.9783ms       112  17.663us  13.057us  92.710us  cudaLaunchKernel

==25751==       Range "Range_2015"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.4721ms       112  22.072us  15.136us  121.13us  Range_2015
 GPU activities:  100.00%  1.6802ms       112  15.001us  14.657us  22.657us  void cuFillLayer::fill<int>(cuFillLayer::KernelArgs<int>)
      API calls:  100.00%  2.0711ms       112  18.492us  12.256us  117.61us  cudaLaunchKernel

==25751==       Range "Reformatting CopyNode for Input Tensor 0 to PWN(PWN(Sigmoid_283, Mul_284), Add_285)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  7.7907ms       112  69.560us  46.979us  255.69us  Reformatting CopyNode for Input Tensor 0 to PWN(PWN(Sigmoid_283, Mul_284), Add_285)
 GPU activities:  100.00%  12.492ms       112  111.53us  106.95us  160.08us  void CUTENSOR_NAMESPACE::permutationKernelPLC3<CUTENSOR_NAMESPACE::VectorWrite2DTensorView<unsigned char=0, unsigned char=1, bool=0, unsigned int=4, float, float, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::VectorRead2DTensorView<unsigned char=1, unsigned char=0, bool=0, unsigned int=4, float, float, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::ThreadLevelElementwise<CUTENSOR_NAMESPACE::ElementwiseConfig2DCommonCase<CUTENSOR_NAMESPACE::GeneralUnarySmall<float>, CUTENSOR_NAMESPACE::GeneralBinary<float>, int=2, int=64, int=64, int=256, char=4, bool=0, bool=1, bool=1, bool=1, bool=0>, float>, CUTENSOR_NAMESPACE::ElementwiseRuntimePLC3<float, float, float, float>::Params>(unsigned int=4)
      API calls:  100.00%  3.0093ms       112  26.868us  18.209us  69.316us  cudaLaunchKernel

==25751==       Range "Reformatting CopyNode for Input Tensor 0 to PWN(PWN(Sigmoid_316, Mul_317), Add_318)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  7.0893ms       112  63.297us  37.794us  105.86us  Reformatting CopyNode for Input Tensor 0 to PWN(PWN(Sigmoid_316, Mul_317), Add_318)
 GPU activities:  100.00%  11.943ms       112  106.63us  101.13us  156.84us  void CUTENSOR_NAMESPACE::permutationKernelPLC3<CUTENSOR_NAMESPACE::VectorWrite2DTensorView<unsigned char=0, unsigned char=1, bool=0, unsigned int=4, float, float, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::VectorRead2DTensorView<unsigned char=1, unsigned char=0, bool=0, unsigned int=4, float, float, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::ThreadLevelElementwise<CUTENSOR_NAMESPACE::ElementwiseConfig2DCommonCase<CUTENSOR_NAMESPACE::GeneralUnarySmall<float>, CUTENSOR_NAMESPACE::GeneralBinary<float>, int=2, int=64, int=64, int=256, char=4, bool=0, bool=1, bool=1, bool=1, bool=0>, float>, CUTENSOR_NAMESPACE::ElementwiseRuntimePLC3<float, float, float, float>::Params>(unsigned int=4)
      API calls:  100.00%  2.7194ms       112  24.280us  15.137us  54.883us  cudaLaunchKernel

==25751==       Range "Reformatting CopyNode for Input Tensor 0 to PWN(PWN(Sigmoid_349, Mul_350), Add_351)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  6.7278ms       112  60.069us  38.467us  141.16us  Reformatting CopyNode for Input Tensor 0 to PWN(PWN(Sigmoid_349, Mul_350), Add_351)
 GPU activities:  100.00%  11.973ms       112  106.90us  101.32us  160.88us  void CUTENSOR_NAMESPACE::permutationKernelPLC3<CUTENSOR_NAMESPACE::VectorWrite2DTensorView<unsigned char=0, unsigned char=1, bool=0, unsigned int=4, float, float, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::VectorRead2DTensorView<unsigned char=1, unsigned char=0, bool=0, unsigned int=4, float, float, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::ThreadLevelElementwise<CUTENSOR_NAMESPACE::ElementwiseConfig2DCommonCase<CUTENSOR_NAMESPACE::GeneralUnarySmall<float>, CUTENSOR_NAMESPACE::GeneralBinary<float>, int=2, int=64, int=64, int=256, char=4, bool=0, bool=1, bool=1, bool=1, bool=0>, float>, CUTENSOR_NAMESPACE::ElementwiseRuntimePLC3<float, float, float, float>::Params>(unsigned int=4)
      API calls:  100.00%  2.6519ms       112  23.677us  15.393us  52.707us  cudaLaunchKernel

==25751==       Range "Reformatting CopyNode for Input Tensor 0 to PWN(PWN(Sigmoid_382, Mul_383), Add_384)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  6.9786ms       112  62.309us  37.890us  174.19us  Reformatting CopyNode for Input Tensor 0 to PWN(PWN(Sigmoid_382, Mul_383), Add_384)
 GPU activities:  100.00%  12.957ms       112  115.69us  100.04us  1.1712ms  void CUTENSOR_NAMESPACE::permutationKernelPLC3<CUTENSOR_NAMESPACE::VectorWrite2DTensorView<unsigned char=0, unsigned char=1, bool=0, unsigned int=4, float, float, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::VectorRead2DTensorView<unsigned char=1, unsigned char=0, bool=0, unsigned int=4, float, float, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::ThreadLevelElementwise<CUTENSOR_NAMESPACE::ElementwiseConfig2DCommonCase<CUTENSOR_NAMESPACE::GeneralUnarySmall<float>, CUTENSOR_NAMESPACE::GeneralBinary<float>, int=2, int=64, int=64, int=256, char=4, bool=0, bool=1, bool=1, bool=1, bool=0>, float>, CUTENSOR_NAMESPACE::ElementwiseRuntimePLC3<float, float, float, float>::Params>(unsigned int=4)
      API calls:  100.00%  2.7987ms       112  24.988us  15.393us  112.77us  cudaLaunchKernel

==25751==       Range "Reformatting CopyNode for Input Tensor 0 to PWN(PWN(Sigmoid_415, Mul_416), Add_417)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  6.5240ms       112  58.249us  37.538us  132.71us  Reformatting CopyNode for Input Tensor 0 to PWN(PWN(Sigmoid_415, Mul_416), Add_417)
 GPU activities:  100.00%  12.089ms       112  107.94us  101.67us  155.21us  void CUTENSOR_NAMESPACE::permutationKernelPLC3<CUTENSOR_NAMESPACE::VectorWrite2DTensorView<unsigned char=0, unsigned char=1, bool=0, unsigned int=4, float, float, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::VectorRead2DTensorView<unsigned char=1, unsigned char=0, bool=0, unsigned int=4, float, float, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::ThreadLevelElementwise<CUTENSOR_NAMESPACE::ElementwiseConfig2DCommonCase<CUTENSOR_NAMESPACE::GeneralUnarySmall<float>, CUTENSOR_NAMESPACE::GeneralBinary<float>, int=2, int=64, int=64, int=256, char=4, bool=0, bool=1, bool=1, bool=1, bool=0>, float>, CUTENSOR_NAMESPACE::ElementwiseRuntimePLC3<float, float, float, float>::Params>(unsigned int=4)
      API calls:  100.00%  2.6339ms       112  23.517us  14.881us  48.003us  cudaLaunchKernel

==25751==       Range "Reformatting CopyNode for Input Tensor 0 to PWN(PWN(Sigmoid_448, Mul_449), Add_450)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  6.9399ms       112  61.963us  37.698us  114.85us  Reformatting CopyNode for Input Tensor 0 to PWN(PWN(Sigmoid_448, Mul_449), Add_450)
 GPU activities:  100.00%  12.037ms       112  107.47us  102.63us  158.35us  void CUTENSOR_NAMESPACE::permutationKernelPLC3<CUTENSOR_NAMESPACE::VectorWrite2DTensorView<unsigned char=0, unsigned char=1, bool=0, unsigned int=4, float, float, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::VectorRead2DTensorView<unsigned char=1, unsigned char=0, bool=0, unsigned int=4, float, float, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::ThreadLevelElementwise<CUTENSOR_NAMESPACE::ElementwiseConfig2DCommonCase<CUTENSOR_NAMESPACE::GeneralUnarySmall<float>, CUTENSOR_NAMESPACE::GeneralBinary<float>, int=2, int=64, int=64, int=256, char=4, bool=0, bool=1, bool=1, bool=1, bool=0>, float>, CUTENSOR_NAMESPACE::ElementwiseRuntimePLC3<float, float, float, float>::Params>(unsigned int=4)
      API calls:  100.00%  2.8296ms       112  25.264us  15.169us  89.798us  cudaLaunchKernel

==25751==       Range "Reformatting CopyNode for Input Tensor 0 to PWN(PWN(Sigmoid_481, Mul_482), Add_483)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  6.3297ms       112  56.514us  37.250us  118.09us  Reformatting CopyNode for Input Tensor 0 to PWN(PWN(Sigmoid_481, Mul_482), Add_483)
 GPU activities:  100.00%  12.101ms       112  108.05us  102.34us  158.99us  void CUTENSOR_NAMESPACE::permutationKernelPLC3<CUTENSOR_NAMESPACE::VectorWrite2DTensorView<unsigned char=0, unsigned char=1, bool=0, unsigned int=4, float, float, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::VectorRead2DTensorView<unsigned char=1, unsigned char=0, bool=0, unsigned int=4, float, float, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::ThreadLevelElementwise<CUTENSOR_NAMESPACE::ElementwiseConfig2DCommonCase<CUTENSOR_NAMESPACE::GeneralUnarySmall<float>, CUTENSOR_NAMESPACE::GeneralBinary<float>, int=2, int=64, int=64, int=256, char=4, bool=0, bool=1, bool=1, bool=1, bool=0>, float>, CUTENSOR_NAMESPACE::ElementwiseRuntimePLC3<float, float, float, float>::Params>(unsigned int=4)
      API calls:  100.00%  2.6106ms       112  23.308us  14.977us  93.446us  cudaLaunchKernel

==25751==       Range "Reformatting CopyNode for Input Tensor 0 to PWN(PWN(Sigmoid_514, Mul_515), Add_516)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  6.7261ms       112  60.054us  36.802us  143.53us  Reformatting CopyNode for Input Tensor 0 to PWN(PWN(Sigmoid_514, Mul_515), Add_516)
 GPU activities:  100.00%  12.017ms       112  107.30us  99.816us  159.44us  void CUTENSOR_NAMESPACE::permutationKernelPLC3<CUTENSOR_NAMESPACE::VectorWrite2DTensorView<unsigned char=0, unsigned char=1, bool=0, unsigned int=4, float, float, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::VectorRead2DTensorView<unsigned char=1, unsigned char=0, bool=0, unsigned int=4, float, float, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::ThreadLevelElementwise<CUTENSOR_NAMESPACE::ElementwiseConfig2DCommonCase<CUTENSOR_NAMESPACE::GeneralUnarySmall<float>, CUTENSOR_NAMESPACE::GeneralBinary<float>, int=2, int=64, int=64, int=256, char=4, bool=0, bool=1, bool=1, bool=1, bool=0>, float>, CUTENSOR_NAMESPACE::ElementwiseRuntimePLC3<float, float, float, float>::Params>(unsigned int=4)
      API calls:  100.00%  2.6617ms       112  23.764us  15.137us  103.37us  cudaLaunchKernel

==25751==       Range "Reformatting CopyNode for Input Tensor 0 to PWN(PWN(Sigmoid_547, Mul_548), Add_549)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  6.5642ms       112  58.609us  37.218us  148.20us  Reformatting CopyNode for Input Tensor 0 to PWN(PWN(Sigmoid_547, Mul_548), Add_549)
 GPU activities:  100.00%  12.029ms       112  107.40us  102.41us  157.84us  void CUTENSOR_NAMESPACE::permutationKernelPLC3<CUTENSOR_NAMESPACE::VectorWrite2DTensorView<unsigned char=0, unsigned char=1, bool=0, unsigned int=4, float, float, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::VectorRead2DTensorView<unsigned char=1, unsigned char=0, bool=0, unsigned int=4, float, float, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::ThreadLevelElementwise<CUTENSOR_NAMESPACE::ElementwiseConfig2DCommonCase<CUTENSOR_NAMESPACE::GeneralUnarySmall<float>, CUTENSOR_NAMESPACE::GeneralBinary<float>, int=2, int=64, int=64, int=256, char=4, bool=0, bool=1, bool=1, bool=1, bool=0>, float>, CUTENSOR_NAMESPACE::ElementwiseRuntimePLC3<float, float, float, float>::Params>(unsigned int=4)
      API calls:  100.00%  2.6690ms       112  23.830us  15.201us  58.084us  cudaLaunchKernel

==25751==       Range "Reformatting CopyNode for Input Tensor 0 to PWN(PWN(Sigmoid_645, Mul_646), Add_647)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  7.2949ms       112  65.133us  37.602us  1.2999ms  Reformatting CopyNode for Input Tensor 0 to PWN(PWN(Sigmoid_645, Mul_646), Add_647)
 GPU activities:  100.00%  7.6755ms       112  68.530us  65.029us  97.640us  void CUTENSOR_NAMESPACE::permutationKernelPLC3<CUTENSOR_NAMESPACE::VectorWrite2DTensorView<unsigned char=0, unsigned char=1, bool=0, unsigned int=4, float, float, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::VectorRead2DTensorView<unsigned char=1, unsigned char=0, bool=0, unsigned int=4, float, float, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::ThreadLevelElementwise<CUTENSOR_NAMESPACE::ElementwiseConfig2DCommonCase<CUTENSOR_NAMESPACE::GeneralUnarySmall<float>, CUTENSOR_NAMESPACE::GeneralBinary<float>, int=2, int=64, int=64, int=256, char=4, bool=0, bool=1, bool=1, bool=1, bool=0>, float>, CUTENSOR_NAMESPACE::ElementwiseRuntimePLC3<float, float, float, float>::Params>(unsigned int=4)
      API calls:  100.00%  3.6682ms       112  32.752us  15.073us  1.2650ms  cudaLaunchKernel

==25751==       Range "Reformatting CopyNode for Input Tensor 0 to PWN(PWN(Sigmoid_678, Mul_679), Add_680)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  6.1890ms       112  55.258us  36.866us  155.91us  Reformatting CopyNode for Input Tensor 0 to PWN(PWN(Sigmoid_678, Mul_679), Add_680)
 GPU activities:  100.00%  7.7593ms       112  69.279us  65.829us  96.679us  void CUTENSOR_NAMESPACE::permutationKernelPLC3<CUTENSOR_NAMESPACE::VectorWrite2DTensorView<unsigned char=0, unsigned char=1, bool=0, unsigned int=4, float, float, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::VectorRead2DTensorView<unsigned char=1, unsigned char=0, bool=0, unsigned int=4, float, float, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::ThreadLevelElementwise<CUTENSOR_NAMESPACE::ElementwiseConfig2DCommonCase<CUTENSOR_NAMESPACE::GeneralUnarySmall<float>, CUTENSOR_NAMESPACE::GeneralBinary<float>, int=2, int=64, int=64, int=256, char=4, bool=0, bool=1, bool=1, bool=1, bool=0>, float>, CUTENSOR_NAMESPACE::ElementwiseRuntimePLC3<float, float, float, float>::Params>(unsigned int=4)
      API calls:  100.00%  2.5450ms       112  22.723us  14.913us  70.148us  cudaLaunchKernel

==25751==       Range "Reformatting CopyNode for Input Tensor 0 to PWN(PWN(Sigmoid_711, Mul_712), Add_713)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  6.0786ms       112  54.273us  37.474us  161.93us  Reformatting CopyNode for Input Tensor 0 to PWN(PWN(Sigmoid_711, Mul_712), Add_713)
 GPU activities:  100.00%  7.7212ms       112  68.939us  65.254us  96.263us  void CUTENSOR_NAMESPACE::permutationKernelPLC3<CUTENSOR_NAMESPACE::VectorWrite2DTensorView<unsigned char=0, unsigned char=1, bool=0, unsigned int=4, float, float, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::VectorRead2DTensorView<unsigned char=1, unsigned char=0, bool=0, unsigned int=4, float, float, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::ThreadLevelElementwise<CUTENSOR_NAMESPACE::ElementwiseConfig2DCommonCase<CUTENSOR_NAMESPACE::GeneralUnarySmall<float>, CUTENSOR_NAMESPACE::GeneralBinary<float>, int=2, int=64, int=64, int=256, char=4, bool=0, bool=1, bool=1, bool=1, bool=0>, float>, CUTENSOR_NAMESPACE::ElementwiseRuntimePLC3<float, float, float, float>::Params>(unsigned int=4)
      API calls:  100.00%  2.5753ms       112  22.993us  15.265us  133.48us  cudaLaunchKernel

==25751==       Range "Reformatting CopyNode for Input Tensor 0 to PWN(PWN(Sigmoid_744, Mul_745), Add_746)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  6.0934ms       112  54.405us  37.666us  206.54us  Reformatting CopyNode for Input Tensor 0 to PWN(PWN(Sigmoid_744, Mul_745), Add_746)
 GPU activities:  100.00%  7.7432ms       112  69.136us  64.837us  101.26us  void CUTENSOR_NAMESPACE::permutationKernelPLC3<CUTENSOR_NAMESPACE::VectorWrite2DTensorView<unsigned char=0, unsigned char=1, bool=0, unsigned int=4, float, float, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::VectorRead2DTensorView<unsigned char=1, unsigned char=0, bool=0, unsigned int=4, float, float, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::ThreadLevelElementwise<CUTENSOR_NAMESPACE::ElementwiseConfig2DCommonCase<CUTENSOR_NAMESPACE::GeneralUnarySmall<float>, CUTENSOR_NAMESPACE::GeneralBinary<float>, int=2, int=64, int=64, int=256, char=4, bool=0, bool=1, bool=1, bool=1, bool=0>, float>, CUTENSOR_NAMESPACE::ElementwiseRuntimePLC3<float, float, float, float>::Params>(unsigned int=4)
      API calls:  100.00%  2.3970ms       112  21.401us  15.264us  53.379us  cudaLaunchKernel

==25751==       Range "Reformatting CopyNode for Input Tensor 0 to PWN(PWN(Sigmoid_777, Mul_778), Add_779)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  5.6917ms       112  50.818us  36.930us  98.309us  Reformatting CopyNode for Input Tensor 0 to PWN(PWN(Sigmoid_777, Mul_778), Add_779)
 GPU activities:  100.00%  7.7052ms       112  68.795us  64.645us  99.592us  void CUTENSOR_NAMESPACE::permutationKernelPLC3<CUTENSOR_NAMESPACE::VectorWrite2DTensorView<unsigned char=0, unsigned char=1, bool=0, unsigned int=4, float, float, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::VectorRead2DTensorView<unsigned char=1, unsigned char=0, bool=0, unsigned int=4, float, float, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::ThreadLevelElementwise<CUTENSOR_NAMESPACE::ElementwiseConfig2DCommonCase<CUTENSOR_NAMESPACE::GeneralUnarySmall<float>, CUTENSOR_NAMESPACE::GeneralBinary<float>, int=2, int=64, int=64, int=256, char=4, bool=0, bool=1, bool=1, bool=1, bool=0>, float>, CUTENSOR_NAMESPACE::ElementwiseRuntimePLC3<float, float, float, float>::Params>(unsigned int=4)
      API calls:  100.00%  2.2714ms       112  20.280us  15.105us  43.523us  cudaLaunchKernel

==25751==       Range "Reformatting CopyNode for Input Tensor 0 to PWN(PWN(Sigmoid_810, Mul_811), Add_812)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  5.7100ms       112  50.982us  36.994us  135.37us  Reformatting CopyNode for Input Tensor 0 to PWN(PWN(Sigmoid_810, Mul_811), Add_812)
 GPU activities:  100.00%  7.7120ms       112  68.857us  65.093us  103.30us  void CUTENSOR_NAMESPACE::permutationKernelPLC3<CUTENSOR_NAMESPACE::VectorWrite2DTensorView<unsigned char=0, unsigned char=1, bool=0, unsigned int=4, float, float, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::VectorRead2DTensorView<unsigned char=1, unsigned char=0, bool=0, unsigned int=4, float, float, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::ThreadLevelElementwise<CUTENSOR_NAMESPACE::ElementwiseConfig2DCommonCase<CUTENSOR_NAMESPACE::GeneralUnarySmall<float>, CUTENSOR_NAMESPACE::GeneralBinary<float>, int=2, int=64, int=64, int=256, char=4, bool=0, bool=1, bool=1, bool=1, bool=0>, float>, CUTENSOR_NAMESPACE::ElementwiseRuntimePLC3<float, float, float, float>::Params>(unsigned int=4)
      API calls:  100.00%  2.3009ms       112  20.544us  14.945us  75.973us  cudaLaunchKernel

==25751==       Range "Reformatting CopyNode for Input Tensor 0 to PWN(PWN(Sigmoid_843, Mul_844), Add_845)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  5.7196ms       112  51.067us  36.674us  145.58us  Reformatting CopyNode for Input Tensor 0 to PWN(PWN(Sigmoid_843, Mul_844), Add_845)
 GPU activities:  100.00%  7.7329ms       112  69.043us  64.901us  97.672us  void CUTENSOR_NAMESPACE::permutationKernelPLC3<CUTENSOR_NAMESPACE::VectorWrite2DTensorView<unsigned char=0, unsigned char=1, bool=0, unsigned int=4, float, float, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::VectorRead2DTensorView<unsigned char=1, unsigned char=0, bool=0, unsigned int=4, float, float, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::ThreadLevelElementwise<CUTENSOR_NAMESPACE::ElementwiseConfig2DCommonCase<CUTENSOR_NAMESPACE::GeneralUnarySmall<float>, CUTENSOR_NAMESPACE::GeneralBinary<float>, int=2, int=64, int=64, int=256, char=4, bool=0, bool=1, bool=1, bool=1, bool=0>, float>, CUTENSOR_NAMESPACE::ElementwiseRuntimePLC3<float, float, float, float>::Params>(unsigned int=4)
      API calls:  100.00%  2.3322ms       112  20.823us  15.009us  106.09us  cudaLaunchKernel

==25751==       Range "Reformatting CopyNode for Input Tensor 0 to PWN(PWN(Sigmoid_876, Mul_877), Add_878)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  5.5371ms       112  49.438us  37.058us  104.17us  Reformatting CopyNode for Input Tensor 0 to PWN(PWN(Sigmoid_876, Mul_877), Add_878)
 GPU activities:  100.00%  7.7013ms       112  68.761us  64.997us  99.528us  void CUTENSOR_NAMESPACE::permutationKernelPLC3<CUTENSOR_NAMESPACE::VectorWrite2DTensorView<unsigned char=0, unsigned char=1, bool=0, unsigned int=4, float, float, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::VectorRead2DTensorView<unsigned char=1, unsigned char=0, bool=0, unsigned int=4, float, float, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::ThreadLevelElementwise<CUTENSOR_NAMESPACE::ElementwiseConfig2DCommonCase<CUTENSOR_NAMESPACE::GeneralUnarySmall<float>, CUTENSOR_NAMESPACE::GeneralBinary<float>, int=2, int=64, int=64, int=256, char=4, bool=0, bool=1, bool=1, bool=1, bool=0>, float>, CUTENSOR_NAMESPACE::ElementwiseRuntimePLC3<float, float, float, float>::Params>(unsigned int=4)
      API calls:  100.00%  2.1932ms       112  19.582us  15.105us  42.211us  cudaLaunchKernel

==25751==       Range "Reformatting CopyNode for Input Tensor 0 to PWN(PWN(Sigmoid_909, Mul_910), Add_911)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  5.1809ms       112  46.257us  36.898us  75.300us  Reformatting CopyNode for Input Tensor 0 to PWN(PWN(Sigmoid_909, Mul_910), Add_911)
 GPU activities:  100.00%  8.0820ms       112  72.160us  65.190us  404.64us  void CUTENSOR_NAMESPACE::permutationKernelPLC3<CUTENSOR_NAMESPACE::VectorWrite2DTensorView<unsigned char=0, unsigned char=1, bool=0, unsigned int=4, float, float, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::VectorRead2DTensorView<unsigned char=1, unsigned char=0, bool=0, unsigned int=4, float, float, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::ThreadLevelElementwise<CUTENSOR_NAMESPACE::ElementwiseConfig2DCommonCase<CUTENSOR_NAMESPACE::GeneralUnarySmall<float>, CUTENSOR_NAMESPACE::GeneralBinary<float>, int=2, int=64, int=64, int=256, char=4, bool=0, bool=1, bool=1, bool=1, bool=0>, float>, CUTENSOR_NAMESPACE::ElementwiseRuntimePLC3<float, float, float, float>::Params>(unsigned int=4)
      API calls:  100.00%  2.1859ms       112  19.516us  15.137us  45.475us  cudaLaunchKernel

==25751==       Range "Reformatting CopyNode for Input Tensor 0 to PWN(Sigmoid_1156, Mul_1157)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  5.9459ms       112  53.088us  38.562us  109.00us  Reformatting CopyNode for Input Tensor 0 to PWN(Sigmoid_1156, Mul_1157)
 GPU activities:  100.00%  4.9936ms       112  44.585us  42.564us  61.765us  void CUTENSOR_NAMESPACE::permutationKernelPLC3<CUTENSOR_NAMESPACE::VectorWrite2DTensorView<unsigned char=0, unsigned char=1, bool=0, unsigned int=4, float, float, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::VectorRead2DTensorView<unsigned char=1, unsigned char=0, bool=0, unsigned int=4, float, float, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::ThreadLevelElementwise<CUTENSOR_NAMESPACE::ElementwiseConfig2DCommonCase<CUTENSOR_NAMESPACE::GeneralUnarySmall<float>, CUTENSOR_NAMESPACE::GeneralBinary<float>, int=2, int=64, int=64, int=256, char=4, bool=0, bool=1, bool=1, bool=1, bool=0>, float>, CUTENSOR_NAMESPACE::ElementwiseRuntimePLC3<float, float, float, float>::Params>(unsigned int=4)
      API calls:  100.00%  2.2963ms       112  20.502us  14.816us  76.132us  cudaLaunchKernel

==25751==       Range "Reformatting CopyNode for Input Tensor 0 to PWN(Sigmoid_1320, Mul_1321)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  5.8181ms       112  51.947us  38.595us  152.68us  Reformatting CopyNode for Input Tensor 0 to PWN(Sigmoid_1320, Mul_1321)
 GPU activities:  100.00%  7.7218ms       112  68.944us  64.613us  94.343us  void CUTENSOR_NAMESPACE::permutationKernelPLC3<CUTENSOR_NAMESPACE::VectorWrite2DTensorView<unsigned char=0, unsigned char=1, bool=0, unsigned int=4, float, float, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::VectorRead2DTensorView<unsigned char=1, unsigned char=0, bool=0, unsigned int=4, float, float, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::ThreadLevelElementwise<CUTENSOR_NAMESPACE::ElementwiseConfig2DCommonCase<CUTENSOR_NAMESPACE::GeneralUnarySmall<float>, CUTENSOR_NAMESPACE::GeneralBinary<float>, int=2, int=64, int=64, int=256, char=4, bool=0, bool=1, bool=1, bool=1, bool=0>, float>, CUTENSOR_NAMESPACE::ElementwiseRuntimePLC3<float, float, float, float>::Params>(unsigned int=4)
      API calls:  100.00%  2.1812ms       112  19.475us  15.073us  57.027us  cudaLaunchKernel

==25751==       Range "Reformatting CopyNode for Input Tensor 0 to PWN(Sigmoid_1792, Mul_1793)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  6.2500ms       112  55.803us  41.091us  252.30us  Reformatting CopyNode for Input Tensor 0 to PWN(Sigmoid_1792, Mul_1793)
 GPU activities:  100.00%  7.7244ms       112  68.968us  65.477us  95.816us  void CUTENSOR_NAMESPACE::permutationKernelPLC3<CUTENSOR_NAMESPACE::VectorWrite2DTensorView<unsigned char=0, unsigned char=1, bool=0, unsigned int=4, float, float, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::VectorRead2DTensorView<unsigned char=1, unsigned char=0, bool=0, unsigned int=4, float, float, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::ThreadLevelElementwise<CUTENSOR_NAMESPACE::ElementwiseConfig2DCommonCase<CUTENSOR_NAMESPACE::GeneralUnarySmall<float>, CUTENSOR_NAMESPACE::GeneralBinary<float>, int=2, int=64, int=64, int=256, char=4, bool=0, bool=1, bool=1, bool=1, bool=0>, float>, CUTENSOR_NAMESPACE::ElementwiseRuntimePLC3<float, float, float, float>::Params>(unsigned int=4)
      API calls:  100.00%  2.5672ms       112  22.921us  15.680us  190.25us  cudaLaunchKernel

==25751==       Range "Reformatting CopyNode for Input Tensor 0 to PWN(Sigmoid_613, Mul_614)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  6.9375ms       112  61.941us  37.122us  170.22us  Reformatting CopyNode for Input Tensor 0 to PWN(Sigmoid_613, Mul_614)
 GPU activities:  100.00%  7.7362ms       112  69.072us  64.549us  96.776us  void CUTENSOR_NAMESPACE::permutationKernelPLC3<CUTENSOR_NAMESPACE::VectorWrite2DTensorView<unsigned char=0, unsigned char=1, bool=0, unsigned int=4, float, float, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::VectorRead2DTensorView<unsigned char=1, unsigned char=0, bool=0, unsigned int=4, float, float, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::ThreadLevelElementwise<CUTENSOR_NAMESPACE::ElementwiseConfig2DCommonCase<CUTENSOR_NAMESPACE::GeneralUnarySmall<float>, CUTENSOR_NAMESPACE::GeneralBinary<float>, int=2, int=64, int=64, int=256, char=4, bool=0, bool=1, bool=1, bool=1, bool=0>, float>, CUTENSOR_NAMESPACE::ElementwiseRuntimePLC3<float, float, float, float>::Params>(unsigned int=4)
      API calls:  100.00%  2.5229ms       112  22.525us  15.233us  47.907us  cudaLaunchKernel

==25751==       Range "Reformatting CopyNode for Input Tensor 0 to PWN(Sigmoid_975, Mul_976)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  5.2121ms       112  46.536us  36.610us  102.73us  Reformatting CopyNode for Input Tensor 0 to PWN(Sigmoid_975, Mul_976)
 GPU activities:  100.00%  5.0066ms       112  44.701us  41.699us  61.605us  void CUTENSOR_NAMESPACE::permutationKernelPLC3<CUTENSOR_NAMESPACE::VectorWrite2DTensorView<unsigned char=0, unsigned char=1, bool=0, unsigned int=4, float, float, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::VectorRead2DTensorView<unsigned char=1, unsigned char=0, bool=0, unsigned int=4, float, float, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::ThreadLevelElementwise<CUTENSOR_NAMESPACE::ElementwiseConfig2DCommonCase<CUTENSOR_NAMESPACE::GeneralUnarySmall<float>, CUTENSOR_NAMESPACE::GeneralBinary<float>, int=2, int=64, int=64, int=256, char=4, bool=0, bool=1, bool=1, bool=1, bool=0>, float>, CUTENSOR_NAMESPACE::ElementwiseRuntimePLC3<float, float, float, float>::Params>(unsigned int=4)
      API calls:  100.00%  2.0951ms       112  18.706us  14.881us  31.585us  cudaLaunchKernel

==25751==       Range "Reformatting CopyNode for Input Tensor 0 to model.0.conv.conv.module.weight + QuantizeLinear_49_quantize_scale_node + Conv_53"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  4.8659ms       112  43.445us  27.489us  115.69us  Reformatting CopyNode for Input Tensor 0 to model.0.conv.conv.module.weight + QuantizeLinear_49_quantize_scale_node + Conv_53
 GPU activities:  100.00%  6.9970ms       112  62.473us  57.477us  132.94us  cuInt8::ncqhw4ToNc32hw32(char4 const *, char4*, nvinfer1::rt::ReducedDivisor, int, nvinfer1::rt::ReducedDivisor, nvinfer1::rt::ReducedDivisor, int, int, float const *, float const *)
      API calls:  100.00%  3.0822ms       112  27.519us  16.641us  100.97us  cudaLaunchKernel

==25751==       Range "Reformatting CopyNode for Input Tensor 0 to model.17.cv3.conv.module.weight + QuantizeLinear_1462_quantize_scale_node + Conv_1466"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  3.2353ms       112  28.886us  21.569us  74.276us  Reformatting CopyNode for Input Tensor 0 to model.17.cv3.conv.module.weight + QuantizeLinear_1462_quantize_scale_node + Conv_1466
 GPU activities:  100.00%  3.9025ms       112  34.843us  32.291us  45.924us  cuInt8::nc32hw32ToNcqhw4_block(char4 const *, char4*, int, int, int, int, float const *, float const *)
      API calls:  100.00%  1.9232ms       112  17.171us  13.504us  43.587us  cudaLaunchKernel

==25751==       Range "Reformatting CopyNode for Input Tensor 0 to model.2.m.0.cv2.conv.module.weight + QuantizeLinear_113_quantize_scale_node + Conv_117"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  3.8156ms       112  34.068us  20.738us  79.525us  Reformatting CopyNode for Input Tensor 0 to model.2.m.0.cv2.conv.module.weight + QuantizeLinear_113_quantize_scale_node + Conv_117
 GPU activities:  100.00%  6.7287ms       112  60.077us  54.821us  149.71us  cuInt8::nchwToNcqhw4(char const *, unsigned int*, int, int, int, int, int, int, float const *, float const *)
      API calls:  100.00%  2.4924ms       112  22.253us  13.601us  56.867us  cudaLaunchKernel

==25751==       Range "Reformatting CopyNode for Input Tensor 0 to model.2.m.1.cv2.conv.module.weight + QuantizeLinear_146_quantize_scale_node + Conv_150"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  3.7309ms       112  33.311us  21.153us  69.604us  Reformatting CopyNode for Input Tensor 0 to model.2.m.1.cv2.conv.module.weight + QuantizeLinear_146_quantize_scale_node + Conv_150
 GPU activities:  100.00%  6.7729ms       112  60.472us  56.676us  75.494us  cuInt8::nchwToNcqhw4(char const *, unsigned int*, int, int, int, int, int, int, float const *, float const *)
      API calls:  100.00%  2.4313ms       112  21.707us  13.665us  58.563us  cudaLaunchKernel

==25751==       Range "Reformatting CopyNode for Input Tensor 0 to model.2.m.2.cv2.conv.module.weight + QuantizeLinear_179_quantize_scale_node + Conv_183"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  4.0658ms       112  36.301us  20.609us  133.90us  Reformatting CopyNode for Input Tensor 0 to model.2.m.2.cv2.conv.module.weight + QuantizeLinear_179_quantize_scale_node + Conv_183
 GPU activities:  100.00%  6.6979ms       112  59.802us  56.197us  74.822us  cuInt8::nchwToNcqhw4(char const *, unsigned int*, int, int, int, int, int, int, float const *, float const *)
      API calls:  100.00%  2.6901ms       112  24.018us  13.345us  122.70us  cudaLaunchKernel

==25751==       Range "Reformatting CopyNode for Input Tensor 0 to model.20.cv3.conv.module.weight + QuantizeLinear_1624_quantize_scale_node + Conv_1628"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  3.0735ms       112  27.441us  21.538us  85.284us  Reformatting CopyNode for Input Tensor 0 to model.20.cv3.conv.module.weight + QuantizeLinear_1624_quantize_scale_node + Conv_1628
 GPU activities:  100.00%  2.2997ms       112  20.532us  18.594us  27.298us  cuInt8::nc32hw32ToNcqhw4_block(char4 const *, char4*, int, int, int, int, float const *, float const *)
      API calls:  100.00%  1.8959ms       112  16.928us  13.473us  56.899us  cudaLaunchKernel

==25751==       Range "Reformatting CopyNode for Input Tensor 0 to model.3.conv.module.weight + QuantizeLinear_229_quantize_scale_node + Conv_233"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  8.8816ms       112  79.299us  49.571us  172.97us  Reformatting CopyNode for Input Tensor 0 to model.3.conv.module.weight + QuantizeLinear_229_quantize_scale_node + Conv_233
 GPU activities:  100.00%  25.536ms       112  228.00us  221.68us  336.09us  void CUTENSOR_NAMESPACE::permutationKernelPLC3<CUTENSOR_NAMESPACE::VectorWrite2DTensorView<unsigned char=0, unsigned char=1, bool=0, unsigned int=16, float, char, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::VectorRead2DTensorView<unsigned char=1, unsigned char=0, bool=0, unsigned int=16, float, char, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::ThreadLevelElementwise<CUTENSOR_NAMESPACE::ElementwiseConfig2DCommonCase<CUTENSOR_NAMESPACE::GeneralUnarySmall<float>, CUTENSOR_NAMESPACE::GeneralBinary<float>, int=2, int=32, int=64, int=128, char=16, bool=0, bool=0, bool=1, bool=0, bool=1>, float>, CUTENSOR_NAMESPACE::ElementwiseRuntimePLC3<char, float, char, float>::Params>(unsigned int=16)
      API calls:  100.00%  3.0068ms       112  26.846us  19.681us  35.810us  cudaLaunchKernel

==25751==       Range "Reformatting CopyNode for Input Tensor 0 to model.4.cv1.conv.module.weight + QuantizeLinear_245_quantize_scale_node + Conv_249"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  3.8815ms       112  34.655us  22.273us  105.32us  Reformatting CopyNode for Input Tensor 0 to model.4.cv1.conv.module.weight + QuantizeLinear_245_quantize_scale_node + Conv_249
 GPU activities:  100.00%  3.8874ms       112  34.708us  32.450us  46.468us  cuInt8::nc32hw32ToNcqhw4_block(char4 const *, char4*, int, int, int, int, float const *, float const *)
      API calls:  100.00%  2.5216ms       112  22.514us  14.337us  90.725us  cudaLaunchKernel

==25751==       Range "Reformatting CopyNode for Output Tensor 0 to PWN(PWN(Sigmoid_185, Mul_186), Add_187)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  3.5337ms       112  31.550us  19.809us  64.516us  Reformatting CopyNode for Output Tensor 0 to PWN(PWN(Sigmoid_185, Mul_186), Add_187)
 GPU activities:  100.00%  6.6138ms       112  59.051us  54.948us  76.326us  cuInt8::nchwToNcqhw4(char const *, unsigned int*, int, int, int, int, int, int, float const *, float const *)
      API calls:  100.00%  2.3808ms       112  21.257us  13.153us  53.731us  cudaLaunchKernel

==25751==       Range "Reformatting CopyNode for Output Tensor 0 to PWN(PWN(Sigmoid_547, Mul_548), Add_549)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  3.6631ms       112  32.706us  21.154us  77.860us  Reformatting CopyNode for Output Tensor 0 to PWN(PWN(Sigmoid_547, Mul_548), Add_549)
 GPU activities:  100.00%  3.5763ms       112  31.931us  29.443us  40.100us  cuInt8::nchwToNcqhw4(char const *, unsigned int*, int, int, int, int, int, int, float const *, float const *)
      API calls:  100.00%  2.3618ms       112  21.087us  13.857us  67.268us  cudaLaunchKernel

==25751==       Range "Reformatting CopyNode for Output Tensor 0 to PWN(PWN(Sigmoid_909, Mul_910), Add_911)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  3.2838ms       112  29.319us  21.761us  114.18us  Reformatting CopyNode for Output Tensor 0 to PWN(PWN(Sigmoid_909, Mul_910), Add_911)
 GPU activities:  100.00%  2.0770ms       112  18.544us  17.089us  24.034us  cuInt8::nchwToNcqhw4(char const *, unsigned int*, int, int, int, int, int, int, float const *, float const *)
      API calls:  100.00%  2.0829ms       112  18.597us  13.824us  86.437us  cudaLaunchKernel

==25751==       Range "Reformatting CopyNode for Output Tensor 0 to PWN(Sigmoid_202, Mul_203)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  5.5529ms       112  49.579us  24.258us  1.0121ms  Reformatting CopyNode for Output Tensor 0 to PWN(Sigmoid_202, Mul_203)
 GPU activities:  100.00%  6.8390ms       112  61.062us  56.676us  148.65us  cuInt8::nchwToNcqhw4(char const *, unsigned int*, int, int, int, int, int, int, float const *, float const *)
      API calls:  100.00%  2.9476ms       112  26.318us  14.849us  368.79us  cudaLaunchKernel

==25751==       Range "Reformatting CopyNode for Output Tensor 0 to PWN(Sigmoid_564, Mul_565)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  3.5930ms       112  32.080us  21.601us  57.731us  Reformatting CopyNode for Output Tensor 0 to PWN(Sigmoid_564, Mul_565)
 GPU activities:  100.00%  2.2236ms       112  19.853us  18.209us  25.922us  cuInt8::nc32hw32ToNcqhw4_block(char4 const *, char4*, int, int, int, int, float const *, float const *)
      API calls:  100.00%  2.3049ms       112  20.579us  13.249us  46.947us  cudaLaunchKernel

==25751==       Range "Reformatting CopyNode for Output Tensor 0 to PWN(Sigmoid_71, Mul_72)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  3.9935ms       112  35.656us  23.873us  71.492us  Reformatting CopyNode for Output Tensor 0 to PWN(Sigmoid_71, Mul_72)
 GPU activities:  100.00%  8.9381ms       112  79.804us  57.412us  2.0256ms  cuInt8::nc32hw32ToNcqhw4_block(char4 const *, char4*, int, int, int, int, float const *, float const *)
      API calls:  100.00%  2.5399ms       112  22.677us  14.720us  48.323us  cudaLaunchKernel

==25751==       Range "Reformatting CopyNode for Output Tensor 0 to PWN(Sigmoid_926, Mul_927)"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  3.5423ms       112  31.627us  21.954us  65.316us  Reformatting CopyNode for Output Tensor 0 to PWN(Sigmoid_926, Mul_927)
 GPU activities:  100.00%  1.5686ms       112  14.005us  12.609us  16.833us  cuInt8::nc32hw32ToNcqhw4_block(char4 const *, char4*, int, int, int, int, float const *, float const *)
      API calls:  100.00%  2.2260ms       112  19.875us  13.537us  44.930us  cudaLaunchKernel

==25751==       Range "Reformatting CopyNode for Output Tensor 0 to QuantizeLinear_43_quantize_scale_node_clone_3"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  10.197ms       112  91.042us  62.979us  174.63us  Reformatting CopyNode for Output Tensor 0 to QuantizeLinear_43_quantize_scale_node_clone_3
 GPU activities:  100.00%  4.7297ms       112  42.229us  40.100us  121.32us  void CUTENSOR_NAMESPACE::permutationKernelPLC3<CUTENSOR_NAMESPACE::VectorWrite2DTensorView<unsigned char=0, unsigned char=1, bool=0, unsigned int=4, float, char, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::VectorRead2DTensorView<unsigned char=0, unsigned char=1, bool=0, unsigned int=4, float, char, CUTENSOR_NAMESPACE::GeneralUnarySmall<float>>, CUTENSOR_NAMESPACE::ThreadLevelElementwise<CUTENSOR_NAMESPACE::ElementwiseConfig2DSmallMode0<CUTENSOR_NAMESPACE::GeneralUnarySmall<float>, CUTENSOR_NAMESPACE::GeneralBinary<float>, int=2, int=4, int=256, int=64, char=4, bool=1, bool=0, bool=1, bool=0, bool=1>, float>, CUTENSOR_NAMESPACE::ElementwiseRuntimePLC3<char, float, char, float>::Params>(unsigned int=4)
      API calls:  100.00%  3.4825ms       112  31.093us  20.833us  101.64us  cudaLaunchKernel

==25751==       Range "Reshape_1808 + Transpose_1809"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  4.1858ms       112  37.373us  26.818us  60.387us  Reshape_1808 + Transpose_1809
 GPU activities:  100.00%  3.1822ms       112  28.412us  26.338us  37.251us  void genericReformat::copyPackedKernel<float, float, bool=1, bool=1, genericReformat::ArrayN<int=5>, int=5>(unsigned int, unsigned int, void const *, genericReformat::ArrayN<genericReformat::ArrayN<int=5>>, genericReformat::ArrayNWithReducedDivisors<genericReformat::ArrayN<int=5>>, void const *, int, int, int, float const *, void*, void const *, genericReformat::ArrayNWithReducedDivisors, genericReformat::ArrayNWithReducedDivisors, void const *, int, int, int, float const , int=5)
      API calls:  100.00%  2.2348ms       112  19.953us  15.617us  33.026us  cudaLaunchKernel

==25751==       Range "Reshape_1819 + Reshape_1827"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  207.95us       112  1.8560us  1.0880us  3.9690us  Reshape_1819 + Reshape_1827
No kernels were profiled in this range.
No API activities were profiled in this range.

==25751==       Range "Reshape_1821 + Reshape_1831"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  883.92us       112  7.8920us     736ns  766.86us  Reshape_1821 + Reshape_1831
No kernels were profiled in this range.
No API activities were profiled in this range.

==25751==       Range "Reshape_1892"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  156.84us       112  1.4000us     992ns  3.5520us  Reshape_1892
No kernels were profiled in this range.
No API activities were profiled in this range.

==25751==       Range "Reshape_1907 + Transpose_1908"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  3.1758ms       112  28.355us  21.857us  76.516us  Reshape_1907 + Transpose_1908
 GPU activities:  100.00%  1.2917ms       112  11.532us  10.337us  14.338us  void genericReformat::copyPackedKernel<float, float, bool=1, bool=1, genericReformat::ArrayN<int=5>, int=5>(unsigned int, unsigned int, void const *, genericReformat::ArrayN<genericReformat::ArrayN<int=5>>, genericReformat::ArrayNWithReducedDivisors<genericReformat::ArrayN<int=5>>, void const *, int, int, int, float const *, void*, void const *, genericReformat::ArrayNWithReducedDivisors, genericReformat::ArrayNWithReducedDivisors, void const *, int, int, int, float const , int=5)
      API calls:  100.00%  1.9517ms       112  17.425us  13.921us  66.436us  cudaLaunchKernel

==25751==       Range "Reshape_1918 + Reshape_1926"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  168.46us       112  1.5040us     992ns  17.729us  Reshape_1918 + Reshape_1926
No kernels were profiled in this range.
No API activities were profiled in this range.

==25751==       Range "Reshape_1920 + Reshape_1930"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  124.78us       112  1.1140us     864ns  3.2960us  Reshape_1920 + Reshape_1930
No kernels were profiled in this range.
No API activities were profiled in this range.

==25751==       Range "Reshape_1991"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  147.82us       112  1.3190us     928ns  3.2960us  Reshape_1991
No kernels were profiled in this range.
No API activities were profiled in this range.

==25751==       Range "Reshape_2006 + Transpose_2007"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  3.3892ms       112  30.260us  20.897us  176.72us  Reshape_2006 + Transpose_2007
 GPU activities:  100.00%  678.07us       112  6.0540us  5.6320us  8.6090us  void genericReformat::copyPackedKernel<float, float, bool=1, bool=1, genericReformat::ArrayN<int=5>, int=5>(unsigned int, unsigned int, void const *, genericReformat::ArrayN<genericReformat::ArrayN<int=5>>, genericReformat::ArrayNWithReducedDivisors<genericReformat::ArrayN<int=5>>, void const *, int, int, int, float const *, void*, void const *, genericReformat::ArrayNWithReducedDivisors, genericReformat::ArrayNWithReducedDivisors, void const *, int, int, int, float const , int=5)
      API calls:  100.00%  2.2001ms       112  19.643us  13.569us  163.53us  cudaLaunchKernel

==25751==       Range "Reshape_2017 + Reshape_2025"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  168.33us       112  1.5020us  1.0240us  18.721us  Reshape_2017 + Reshape_2025
No kernels were profiled in this range.
No API activities were profiled in this range.

==25751==       Range "Reshape_2019 + Reshape_2029"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  120.75us       112  1.0780us     640ns  1.9520us  Reshape_2019 + Reshape_2029
No kernels were profiled in this range.
No API activities were profiled in this range.

==25751==       Range "Reshape_2090"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  148.65us       112  1.3270us     896ns  3.0720us  Reshape_2090
No kernels were profiled in this range.
No API activities were profiled in this range.

==25751==       Range "Resize_1159"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  3.2933ms       112  29.404us  21.665us  69.572us  Resize_1159
 GPU activities:  100.00%  9.6311ms       112  85.991us  82.375us  122.35us  void cuResizeLayer::ResizeNearestGenericKernel<float, int=2, int=128, nvinfer1::ResizeRoundMode>(float*, cuResizeLayer::ResizeNearestGenericKernel<float, int=2, int=128, nvinfer1::ResizeRoundMode> const *, cuResizeLayer::LaunchParams)
      API calls:  100.00%  2.2894ms       112  20.441us  15.713us  48.227us  cudaLaunchKernel

==25751==       Range "Resize_1323"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.5696ms       112  22.943us  17.889us  71.108us  Resize_1323
 GPU activities:  100.00%  18.503ms       112  165.21us  158.19us  238.58us  void cuResizeLayer::ResizeNearestGenericKernel<float, int=2, int=128, nvinfer1::ResizeRoundMode>(float*, cuResizeLayer::ResizeNearestGenericKernel<float, int=2, int=128, nvinfer1::ResizeRoundMode> const *, cuResizeLayer::LaunchParams)
      API calls:  100.00%  1.8093ms       112  16.154us  12.865us  55.299us  cudaLaunchKernel

==25751==       Range "Slice_14"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  3.7700ms       112  33.660us  19.170us  79.973us  Slice_14
 GPU activities:  100.00%  6.0531ms       112  54.045us  49.892us  139.66us  void cuSliceLayer::naiveSlice<float, cuSliceLayer::Mode, int=4>(cuSliceLayer::LaunchParams<float>)
      API calls:  100.00%  2.7831ms       112  24.849us  14.560us  72.868us  cudaLaunchKernel

==25751==       Range "Slice_1867"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.5185ms       112  22.486us  16.641us  111.62us  Slice_1867
 GPU activities:  100.00%  1.2204ms       112  10.896us  9.8890us  14.497us  void cuSliceLayer::naiveSlice<float, cuSliceLayer::Mode, int=5>(cuSliceLayer::LaunchParams<float>)
      API calls:  100.00%  1.8072ms       112  16.135us  12.065us  102.28us  cudaLaunchKernel

==25751==       Range "Slice_1879"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.2502ms       112  20.090us  15.328us  77.572us  Slice_1879
 GPU activities:  100.00%  1.1554ms       112  10.316us  9.2800us  14.114us  void cuSliceLayer::naiveSlice<float, cuSliceLayer::Mode, int=5>(cuSliceLayer::LaunchParams<float>)
      API calls:  100.00%  1.6777ms       112  14.979us  11.648us  68.228us  cudaLaunchKernel

==25751==       Range "Slice_1888"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.2381ms       112  19.983us  15.360us  74.916us  Slice_1888
 GPU activities:  100.00%  1.6456ms       112  14.693us  13.249us  19.329us  void cuSliceLayer::naiveSlice<float, cuSliceLayer::Mode, int=5>(cuSliceLayer::LaunchParams<float>)
      API calls:  100.00%  1.6377ms       112  14.622us  11.648us  59.171us  cudaLaunchKernel

==25751==       Range "Slice_19"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  3.4313ms       112  30.636us  15.841us  90.949us  Slice_19
 GPU activities:  100.00%  4.1031ms       112  36.634us  32.643us  102.02us  void cuSliceLayer::naiveSlice<float, cuSliceLayer::Mode, int=4>(cuSliceLayer::LaunchParams<float>)
      API calls:  100.00%  2.5615ms       112  22.870us  12.033us  54.275us  cudaLaunchKernel

==25751==       Range "Slice_1966"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.3775ms       112  21.227us  16.353us  98.181us  Slice_1966
 GPU activities:  100.00%  584.79us       112  5.2210us  4.6410us  7.8400us  void cuSliceLayer::naiveSlice<float, cuSliceLayer::Mode, int=5>(cuSliceLayer::LaunchParams<float>)
      API calls:  100.00%  1.7675ms       112  15.781us  11.841us  89.797us  cudaLaunchKernel

==25751==       Range "Slice_1978"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.4242ms       112  21.644us  15.521us  254.64us  Slice_1978
 GPU activities:  100.00%  462.91us       112  4.1330us  3.7760us  6.2400us  void cuSliceLayer::naiveSlice<float, cuSliceLayer::Mode, int=5>(cuSliceLayer::LaunchParams<float>)
      API calls:  100.00%  1.8658ms       112  16.658us  11.777us  248.37us  cudaLaunchKernel

==25751==       Range "Slice_1987"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.8842ms       112  25.752us  15.393us  794.51us  Slice_1987
 GPU activities:  100.00%  430.28us       112  3.8410us  3.4560us  5.8560us  void cuSliceLayer::naiveSlice<float, cuSliceLayer::Mode, int=5>(cuSliceLayer::LaunchParams<float>)
      API calls:  100.00%  2.3556ms       112  21.031us  11.617us  785.23us  cudaLaunchKernel

==25751==       Range "Slice_2065"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.5097ms       112  22.408us  16.193us  123.30us  Slice_2065
 GPU activities:  100.00%  222.61us       112  1.9870us  1.8560us  2.9760us  void cuSliceLayer::naiveSlice<float, cuSliceLayer::Mode, int=5>(cuSliceLayer::LaunchParams<float>)
      API calls:  100.00%  1.9101ms       112  17.054us  11.936us  113.58us  cudaLaunchKernel

==25751==       Range "Slice_2077"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.7534ms       112  24.583us  15.745us  112.52us  Slice_2077
 GPU activities:  100.00%  230.42us       112  2.0570us  1.8880us  3.6800us  void cuSliceLayer::naiveSlice<float, cuSliceLayer::Mode, int=5>(cuSliceLayer::LaunchParams<float>)
      API calls:  100.00%  2.1658ms       112  19.337us  11.745us  106.15us  cudaLaunchKernel

==25751==       Range "Slice_2086"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.9086ms       112  25.969us  15.520us  170.92us  Slice_2086
 GPU activities:  100.00%  245.18us       112  2.1890us  2.0480us  3.6800us  void cuSliceLayer::naiveSlice<float, cuSliceLayer::Mode, int=5>(cuSliceLayer::LaunchParams<float>)
      API calls:  100.00%  2.3325ms       112  20.826us  11.712us  166.92us  cudaLaunchKernel

==25751==       Range "Slice_24"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  3.3317ms       112  29.747us  17.249us  79.749us  Slice_24
 GPU activities:  100.00%  6.1253ms       112  54.690us  49.796us  138.48us  void cuSliceLayer::naiveSlice<float, cuSliceLayer::Mode, int=4>(cuSliceLayer::LaunchParams<float>)
      API calls:  100.00%  2.4710ms       112  22.062us  13.024us  71.140us  cudaLaunchKernel

==25751==       Range "Slice_29"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  3.1580ms       112  28.196us  15.713us  65.059us  Slice_29
 GPU activities:  100.00%  4.0114ms       112  35.816us  32.706us  102.28us  void cuSliceLayer::naiveSlice<float, cuSliceLayer::Mode, int=4>(cuSliceLayer::LaunchParams<float>)
      API calls:  100.00%  2.3746ms       112  21.201us  11.841us  55.043us  cudaLaunchKernel

==25751==       Range "Slice_34"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  3.1660ms       112  28.268us  16.257us  57.156us  Slice_34
 GPU activities:  100.00%  6.2865ms       112  56.129us  51.748us  138.03us  void cuSliceLayer::naiveSlice<float, cuSliceLayer::Mode, int=4>(cuSliceLayer::LaunchParams<float>)
      API calls:  100.00%  2.4044ms       112  21.467us  12.096us  43.682us  cudaLaunchKernel

==25751==       Range "Slice_39"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  3.1737ms       112  28.336us  15.681us  66.052us  Slice_39
 GPU activities:  100.00%  4.0088ms       112  35.792us  32.419us  103.05us  void cuSliceLayer::naiveSlice<float, cuSliceLayer::Mode, int=4>(cuSliceLayer::LaunchParams<float>)
      API calls:  100.00%  2.3533ms       112  21.011us  11.969us  61.667us  cudaLaunchKernel

==25751==       Range "Slice_4"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  7.6245ms       112  68.076us  38.690us  191.79us  Slice_4
 GPU activities:  100.00%  6.9809ms       112  62.329us  58.052us  144.24us  void cuSliceLayer::naiveSlice<float, cuSliceLayer::Mode, int=4>(cuSliceLayer::LaunchParams<float>)
      API calls:  100.00%  5.7260ms       112  51.124us  29.794us  144.36us  cudaLaunchKernel

==25751==       Range "Slice_9"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  3.3337ms       112  29.765us  15.841us  116.26us  Slice_9
 GPU activities:  100.00%  4.0840ms       112  36.464us  32.802us  102.38us  void cuSliceLayer::naiveSlice<float, cuSliceLayer::Mode, int=4>(cuSliceLayer::LaunchParams<float>)
      API calls:  100.00%  2.4639ms       112  21.999us  11.937us  109.51us  cudaLaunchKernel

==25751==       Range "Unsqueeze_1833"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  138.37us       112  1.2350us     704ns  28.674us  Unsqueeze_1833
No kernels were profiled in this range.
No API activities were profiled in this range.

==25751==       Range "Unsqueeze_1834"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  123.91us       112  1.1060us     704ns  2.0160us  Unsqueeze_1834
No kernels were profiled in this range.
No API activities were profiled in this range.

==25751==       Range "Unsqueeze_1932"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  108.20us       112     966ns     704ns  1.9840us  Unsqueeze_1932
No kernels were profiled in this range.
No API activities were profiled in this range.

==25751==       Range "Unsqueeze_1933"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  121.64us       112  1.0860us     736ns  2.8480us  Unsqueeze_1933
No kernels were profiled in this range.
No API activities were profiled in this range.

==25751==       Range "Unsqueeze_2031"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  108.91us       112     972ns     640ns  1.6320us  Unsqueeze_2031
No kernels were profiled in this range.
No API activities were profiled in this range.

==25751==       Range "Unsqueeze_2032"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  123.88us       112  1.1060us     736ns  3.4880us  Unsqueeze_2032
No kernels were profiled in this range.
No API activities were profiled in this range.

==25751==       Range "model.0.conv.conv.module.weight + QuantizeLinear_49_quantize_scale_node + Conv_53"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  4.8126ms       112  42.969us  29.090us  86.053us  model.0.conv.conv.module.weight + QuantizeLinear_49_quantize_scale_node + Conv_53
 GPU activities:  100.00%  66.916ms       112  597.46us  532.20us  1.7369ms  trt_volta_fp32_i8816cudnn_int8_256x64_ldg16_relu_singleBuffer_large_nt_v1
      API calls:  100.00%  3.0444ms       112  27.181us  17.409us  67.780us  cudaLaunchKernel

==25751==       Range "model.1.conv.module.weight + QuantizeLinear_65_quantize_scale_node + Conv_69"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  3.5495ms       112  31.691us  19.649us  70.436us  model.1.conv.module.weight + QuantizeLinear_65_quantize_scale_node + Conv_69
 GPU activities:  100.00%  48.232ms       112  430.64us  365.15us  2.3475ms  trt_volta_fp32_i8816cudnn_int8_128x128_ldg16_relu_small_nt_v1
      API calls:  100.00%  2.4786ms       112  22.130us  13.889us  62.884us  cudaLaunchKernel

==25751==       Range "model.10.conv.module.weight + QuantizeLinear_1150_quantize_scale_node + Conv_1154"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.3857ms       112  21.300us  16.929us  49.251us  model.10.conv.module.weight + QuantizeLinear_1150_quantize_scale_node + Conv_1154
 GPU activities:  100.00%  5.6902ms       112  50.805us  47.588us  72.229us  trt_volta_fp32_i8816cudnn_int8_128x128_ldg16_relu_small_nt_v1
      API calls:  100.00%  1.7484ms       112  15.610us  12.417us  44.483us  cudaLaunchKernel

==25751==       Range "model.13.cv1.conv.module.weight + QuantizeLinear_1169_quantize_scale_node + Conv_1173"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.4430ms       112  21.812us  17.249us  54.851us  model.13.cv1.conv.module.weight + QuantizeLinear_1169_quantize_scale_node + Conv_1173
 GPU activities:  100.00%  10.930ms       112  97.585us  87.463us  480.97us  trt_volta_fp32_i8816cudnn_int8_128x128_ldg16_relu_small_nt_v1
      API calls:  100.00%  1.7488ms       112  15.614us  12.769us  34.210us  cudaLaunchKernel

==25751==       Range "model.13.cv2.conv.module.weight + QuantizeLinear_1281_quantize_scale_node + Conv_1285"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.2661ms       112  20.232us  16.385us  44.738us  model.13.cv2.conv.module.weight + QuantizeLinear_1281_quantize_scale_node + Conv_1285
 GPU activities:  100.00%  10.198ms       112  91.049us  84.774us  130.19us  trt_volta_fp32_i8816cudnn_int8_128x128_ldg16_relu_small_nt_v1
      API calls:  100.00%  1.6944ms       112  15.128us  12.321us  38.466us  cudaLaunchKernel

==25751==       Range "model.13.cv3.conv.module.weight + QuantizeLinear_1298_quantize_scale_node + Conv_1302"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.3974ms       112  21.405us  16.929us  44.419us  model.13.cv3.conv.module.weight + QuantizeLinear_1298_quantize_scale_node + Conv_1302
 GPU activities:  100.00%  12.949ms       112  115.61us  104.46us  161.23us  trt_volta_fp32_i8816cudnn_int8_128x128_ldg16_relu_small_nt_v1
      API calls:  100.00%  1.7536ms       112  15.656us  12.257us  34.274us  cudaLaunchKernel

==25751==       Range "model.13.m.0.cv1.conv.module.weight + QuantizeLinear_1185_quantize_scale_node + Conv_1189"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.3854ms       112  21.298us  16.993us  84.933us  model.13.m.0.cv1.conv.module.weight + QuantizeLinear_1185_quantize_scale_node + Conv_1189
 GPU activities:  100.00%  5.6827ms       112  50.738us  45.507us  75.014us  trt_volta_fp32_i8816cudnn_int8_128x128_ldg16_relu_small_nt_v1
      API calls:  100.00%  1.7442ms       112  15.572us  12.513us  77.060us  cudaLaunchKernel

==25751==       Range "model.13.m.0.cv2.conv.module.weight + QuantizeLinear_1201_quantize_scale_node + Conv_1205"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.5202ms       112  22.501us  17.185us  92.165us  model.13.m.0.cv2.conv.module.weight + QuantizeLinear_1201_quantize_scale_node + Conv_1205
 GPU activities:  100.00%  17.093ms       112  152.62us  147.28us  221.75us  trt_volta_fp32_i8816cudnn_int8_256x64_ldg16_relu_singleBuffer_small_nt_v1
      API calls:  100.00%  1.8559ms       112  16.570us  12.577us  82.277us  cudaLaunchKernel

==25751==       Range "model.13.m.1.cv1.conv.module.weight + QuantizeLinear_1217_quantize_scale_node + Conv_1221"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.3870ms       112  21.312us  17.024us  66.244us  model.13.m.1.cv1.conv.module.weight + QuantizeLinear_1217_quantize_scale_node + Conv_1221
 GPU activities:  100.00%  6.0009ms       112  53.579us  48.836us  79.078us  trt_volta_fp32_i8816cudnn_int8_128x128_ldg16_relu_small_nt_v1
      API calls:  100.00%  1.7324ms       112  15.467us  12.608us  59.043us  cudaLaunchKernel

==25751==       Range "model.13.m.1.cv2.conv.module.weight + QuantizeLinear_1233_quantize_scale_node + Conv_1237"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.3321ms       112  20.822us  16.993us  41.891us  model.13.m.1.cv2.conv.module.weight + QuantizeLinear_1233_quantize_scale_node + Conv_1237
 GPU activities:  100.00%  17.196ms       112  153.53us  140.94us  731.29us  trt_volta_fp32_i8816cudnn_int8_256x64_ldg16_relu_singleBuffer_small_nt_v1
      API calls:  100.00%  1.7052ms       112  15.224us  12.480us  35.106us  cudaLaunchKernel

==25751==       Range "model.13.m.2.cv1.conv.module.weight + QuantizeLinear_1249_quantize_scale_node + Conv_1253"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.3345ms       112  20.843us  16.961us  64.292us  model.13.m.2.cv1.conv.module.weight + QuantizeLinear_1249_quantize_scale_node + Conv_1253
 GPU activities:  100.00%  5.8742ms       112  52.448us  45.603us  75.814us  trt_volta_fp32_i8816cudnn_int8_128x128_ldg16_relu_small_nt_v1
      API calls:  100.00%  1.7329ms       112  15.472us  12.352us  58.115us  cudaLaunchKernel

==25751==       Range "model.13.m.2.cv2.conv.module.weight + QuantizeLinear_1265_quantize_scale_node + Conv_1269"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.3663ms       112  21.127us  16.993us  42.594us  model.13.m.2.cv2.conv.module.weight + QuantizeLinear_1265_quantize_scale_node + Conv_1269
 GPU activities:  100.00%  16.707ms       112  149.17us  144.08us  217.49us  trt_volta_fp32_i8816cudnn_int8_256x64_ldg16_relu_singleBuffer_small_nt_v1
      API calls:  100.00%  1.7328ms       112  15.471us  12.577us  31.106us  cudaLaunchKernel

==25751==       Range "model.14.conv.module.weight + QuantizeLinear_1314_quantize_scale_node + Conv_1318"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.3084ms       112  20.610us  16.673us  50.979us  model.14.conv.module.weight + QuantizeLinear_1314_quantize_scale_node + Conv_1318
 GPU activities:  100.00%  7.3607ms       112  65.720us  60.485us  91.368us  trt_volta_fp32_i8816cudnn_int8_128x128_ldg16_relu_small_nt_v1
      API calls:  100.00%  1.6894ms       112  15.084us  12.225us  45.250us  cudaLaunchKernel

==25751==       Range "model.17.cv1.conv.module.weight + QuantizeLinear_1333_quantize_scale_node + Conv_1337"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.5178ms       112  22.480us  17.345us  83.909us  model.17.cv1.conv.module.weight + QuantizeLinear_1333_quantize_scale_node + Conv_1337
 GPU activities:  100.00%  13.845ms       112  123.61us  112.81us  168.94us  trt_volta_fp32_i8816cudnn_int8_128x128_ldg16_relu_small_nt_v1
      API calls:  100.00%  1.8704ms       112  16.700us  12.769us  75.076us  cudaLaunchKernel

==25751==       Range "model.17.cv2.conv.module.weight + QuantizeLinear_1445_quantize_scale_node + Conv_1449"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.3239ms       112  20.749us  16.609us  87.141us  model.17.cv2.conv.module.weight + QuantizeLinear_1445_quantize_scale_node + Conv_1449
 GPU activities:  100.00%  13.448ms       112  120.07us  107.79us  169.29us  trt_volta_fp32_i8816cudnn_int8_128x128_ldg16_relu_small_nt_v1
      API calls:  100.00%  1.7589ms       112  15.704us  12.384us  81.668us  cudaLaunchKernel

==25751==       Range "model.17.cv3.conv.module.weight + QuantizeLinear_1462_quantize_scale_node + Conv_1466"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.5984ms       112  23.200us  18.657us  44.931us  model.17.cv3.conv.module.weight + QuantizeLinear_1462_quantize_scale_node + Conv_1466
 GPU activities:  100.00%  21.804ms       112  194.68us  186.58us  278.77us  trt_volta_fp32_icudnn_int8x4_128x64_relu_interior_nn_v1
      API calls:  100.00%  1.8048ms       112  16.114us  12.993us  31.522us  cudaLaunchKernel

==25751==       Range "model.17.m.0.cv1.conv.module.weight + QuantizeLinear_1349_quantize_scale_node + Conv_1353"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.3595ms       112  21.066us  16.929us  46.786us  model.17.m.0.cv1.conv.module.weight + QuantizeLinear_1349_quantize_scale_node + Conv_1353
 GPU activities:  100.00%  8.8085ms       112  78.647us  71.077us  115.69us  trt_volta_fp32_i8816cudnn_int8_128x128_ldg16_relu_small_nt_v1
      API calls:  100.00%  1.7051ms       112  15.223us  12.352us  33.890us  cudaLaunchKernel

==25751==       Range "model.17.m.0.cv2.conv.module.weight + QuantizeLinear_1365_quantize_scale_node + Conv_1369"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.2468ms       112  20.060us  16.673us  39.682us  model.17.m.0.cv2.conv.module.weight + QuantizeLinear_1365_quantize_scale_node + Conv_1369
 GPU activities:  100.00%  19.185ms       112  171.30us  154.96us  247.22us  trt_volta_fp32_i8816cudnn_int8_128x128_ldg16_relu_small_nt_v1
      API calls:  100.00%  1.6472ms       112  14.707us  12.160us  32.706us  cudaLaunchKernel

==25751==       Range "model.17.m.1.cv1.conv.module.weight + QuantizeLinear_1381_quantize_scale_node + Conv_1385"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.3952ms       112  21.385us  17.057us  51.523us  model.17.m.1.cv1.conv.module.weight + QuantizeLinear_1381_quantize_scale_node + Conv_1385
 GPU activities:  100.00%  8.8061ms       112  78.626us  71.558us  113.19us  trt_volta_fp32_i8816cudnn_int8_128x128_ldg16_relu_small_nt_v1
      API calls:  100.00%  1.7401ms       112  15.536us  12.512us  45.859us  cudaLaunchKernel

==25751==       Range "model.17.m.1.cv2.conv.module.weight + QuantizeLinear_1397_quantize_scale_node + Conv_1401"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.3519ms       112  20.998us  16.993us  90.533us  model.17.m.1.cv2.conv.module.weight + QuantizeLinear_1397_quantize_scale_node + Conv_1401
 GPU activities:  100.00%  18.909ms       112  168.83us  154.70us  257.14us  trt_volta_fp32_i8816cudnn_int8_128x128_ldg16_relu_small_nt_v1
      API calls:  100.00%  1.6993ms       112  15.172us  12.352us  79.492us  cudaLaunchKernel

==25751==       Range "model.17.m.2.cv1.conv.module.weight + QuantizeLinear_1413_quantize_scale_node + Conv_1417"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.2843ms       112  20.395us  16.833us  43.170us  model.17.m.2.cv1.conv.module.weight + QuantizeLinear_1413_quantize_scale_node + Conv_1417
 GPU activities:  100.00%  9.1403ms       112  81.609us  66.054us  391.65us  trt_volta_fp32_i8816cudnn_int8_128x128_ldg16_relu_small_nt_v1
      API calls:  100.00%  1.6527ms       112  14.755us  12.417us  37.794us  cudaLaunchKernel

==25751==       Range "model.17.m.2.cv2.conv.module.weight + QuantizeLinear_1429_quantize_scale_node + Conv_1433"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.3953ms       112  21.386us  16.673us  70.436us  model.17.m.2.cv2.conv.module.weight + QuantizeLinear_1429_quantize_scale_node + Conv_1433
 GPU activities:  100.00%  18.911ms       112  168.85us  152.72us  245.01us  trt_volta_fp32_i8816cudnn_int8_128x128_ldg16_relu_small_nt_v1
      API calls:  100.00%  1.7781ms       112  15.875us  12.256us  64.484us  cudaLaunchKernel

==25751==       Range "model.18.conv.module.weight + QuantizeLinear_1478_quantize_scale_node + Conv_1482"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.2933ms       112  20.476us  16.961us  53.635us  model.18.conv.module.weight + QuantizeLinear_1478_quantize_scale_node + Conv_1482
 GPU activities:  100.00%  17.812ms       112  159.03us  151.56us  227.09us  trt_volta_fp32_i8816cudnn_int8_128x128_ldg16_relu_small_nt_v1
      API calls:  100.00%  1.7124ms       112  15.289us  12.609us  48.227us  cudaLaunchKernel

==25751==       Range "model.2.cv1.conv.module.weight + QuantizeLinear_81_quantize_scale_node + Conv_85"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  3.7325ms       112  33.325us  21.857us  130.57us  model.2.cv1.conv.module.weight + QuantizeLinear_81_quantize_scale_node + Conv_85
 GPU activities:  100.00%  15.383ms       112  137.35us  128.75us  383.29us  trt_volta_fp32_icudnn_int8x4_128x64_relu_small_nn_v1
      API calls:  100.00%  2.7576ms       112  24.621us  15.361us  119.94us  cudaLaunchKernel

==25751==       Range "model.2.cv2.conv.module.weight + QuantizeLinear_196_quantize_scale_node + Conv_200"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  3.1612ms       112  28.225us  17.121us  71.908us  model.2.cv2.conv.module.weight + QuantizeLinear_196_quantize_scale_node + Conv_200
 GPU activities:  100.00%  15.077ms       112  134.61us  127.11us  375.58us  trt_volta_fp32_icudnn_int8x4_128x64_relu_small_nn_v1
      API calls:  100.00%  2.3277ms       112  20.783us  12.705us  64.964us  cudaLaunchKernel

==25751==       Range "model.2.cv3.conv.module.weight + QuantizeLinear_213_quantize_scale_node + Conv_217"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  3.3261ms       112  29.697us  18.689us  109.13us  model.2.cv3.conv.module.weight + QuantizeLinear_213_quantize_scale_node + Conv_217
 GPU activities:  100.00%  29.177ms       112  260.51us  243.92us  1.0296ms  trt_volta_fp32_icudnn_int8x4_128x64_relu_interior_nn_v1
      API calls:  100.00%  2.4124ms       112  21.539us  13.248us  101.38us  cudaLaunchKernel

==25751==       Range "model.2.m.0.cv1.conv.module.weight + QuantizeLinear_97_quantize_scale_node + Conv_101"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  3.8365ms       112  34.254us  22.017us  75.621us  model.2.m.0.cv1.conv.module.weight + QuantizeLinear_97_quantize_scale_node + Conv_101
 GPU activities:  100.00%  13.638ms       112  121.77us  115.50us  306.71us  trt_volta_fp32_icudnn_int8x4_128x64_relu_interior_nn_v1
      API calls:  100.00%  2.6709ms       112  23.847us  14.817us  46.787us  cudaLaunchKernel

==25751==       Range "model.2.m.0.cv2.conv.module.weight + QuantizeLinear_113_quantize_scale_node + Conv_117"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  3.2961ms       112  29.429us  17.857us  87.493us  model.2.m.0.cv2.conv.module.weight + QuantizeLinear_113_quantize_scale_node + Conv_117
 GPU activities:  100.00%  48.106ms       112  429.51us  378.72us  5.3590ms  trt_volta_fp32_icudnn_int8x4_128x64_relu_small_nn_v1
      API calls:  100.00%  2.4191ms       112  21.598us  12.865us  78.949us  cudaLaunchKernel

==25751==       Range "model.2.m.1.cv1.conv.module.weight + QuantizeLinear_130_quantize_scale_node + Conv_134"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  3.3820ms       112  30.196us  19.297us  111.30us  model.2.m.1.cv1.conv.module.weight + QuantizeLinear_130_quantize_scale_node + Conv_134
 GPU activities:  100.00%  13.347ms       112  119.17us  114.06us  157.39us  trt_volta_fp32_icudnn_int8x4_128x64_relu_interior_nn_v1
      API calls:  100.00%  2.4406ms       112  21.790us  13.856us  103.01us  cudaLaunchKernel

==25751==       Range "model.2.m.1.cv2.conv.module.weight + QuantizeLinear_146_quantize_scale_node + Conv_150"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  3.0994ms       112  27.673us  17.569us  54.819us  model.2.m.1.cv2.conv.module.weight + QuantizeLinear_146_quantize_scale_node + Conv_150
 GPU activities:  100.00%  44.044ms       112  393.25us  376.99us  1.2457ms  trt_volta_fp32_icudnn_int8x4_128x64_relu_small_nn_v1
      API calls:  100.00%  2.2987ms       112  20.524us  12.704us  47.907us  cudaLaunchKernel

==25751==       Range "model.2.m.2.cv1.conv.module.weight + QuantizeLinear_163_quantize_scale_node + Conv_167"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  3.3205ms       112  29.647us  18.689us  88.550us  model.2.m.2.cv1.conv.module.weight + QuantizeLinear_163_quantize_scale_node + Conv_167
 GPU activities:  100.00%  13.443ms       112  120.03us  115.02us  159.31us  trt_volta_fp32_icudnn_int8x4_128x64_relu_interior_nn_v1
      API calls:  100.00%  2.4125ms       112  21.540us  13.537us  79.620us  cudaLaunchKernel

==25751==       Range "model.2.m.2.cv2.conv.module.weight + QuantizeLinear_179_quantize_scale_node + Conv_183"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  3.2288ms       112  28.828us  17.921us  172.14us  model.2.m.2.cv2.conv.module.weight + QuantizeLinear_179_quantize_scale_node + Conv_183
 GPU activities:  100.00%  43.989ms       112  392.76us  378.40us  1.2566ms  trt_volta_fp32_icudnn_int8x4_128x64_relu_small_nn_v1
      API calls:  100.00%  2.4450ms       112  21.830us  13.249us  163.95us  cudaLaunchKernel

==25751==       Range "model.20.cv1.conv.module.weight + QuantizeLinear_1495_quantize_scale_node + Conv_1499"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.4533ms       112  21.904us  17.121us  82.661us  model.20.cv1.conv.module.weight + QuantizeLinear_1495_quantize_scale_node + Conv_1499
 GPU activities:  100.00%  7.7530ms       112  69.223us  65.222us  97.704us  trt_volta_fp32_i8816cudnn_int8_128x128_ldg16_relu_small_nt_v1
      API calls:  100.00%  1.6854ms       112  15.048us  12.448us  41.346us  cudaLaunchKernel

==25751==       Range "model.20.cv2.conv.module.weight + QuantizeLinear_1607_quantize_scale_node + Conv_1611"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.2410ms       112  20.008us  16.737us  42.499us  model.20.cv2.conv.module.weight + QuantizeLinear_1607_quantize_scale_node + Conv_1611
 GPU activities:  100.00%  7.3278ms       112  65.427us  59.909us  91.143us  trt_volta_fp32_i8816cudnn_int8_128x128_ldg16_relu_small_nt_v1
      API calls:  100.00%  1.6554ms       112  14.780us  12.449us  35.522us  cudaLaunchKernel

==25751==       Range "model.20.cv3.conv.module.weight + QuantizeLinear_1624_quantize_scale_node + Conv_1628"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.6397ms       112  23.569us  18.017us  65.668us  model.20.cv3.conv.module.weight + QuantizeLinear_1624_quantize_scale_node + Conv_1628
 GPU activities:  100.00%  20.765ms       112  185.40us  179.37us  284.18us  trt_volta_fp32_icudnn_int8x4_128x64_relu_interior_nn_v1
      API calls:  100.00%  1.8272ms       112  16.314us  12.801us  35.682us  cudaLaunchKernel

==25751==       Range "model.20.m.0.cv1.conv.module.weight + QuantizeLinear_1511_quantize_scale_node + Conv_1515"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.4115ms       112  21.531us  16.737us  116.55us  model.20.m.0.cv1.conv.module.weight + QuantizeLinear_1511_quantize_scale_node + Conv_1515
 GPU activities:  100.00%  5.5249ms       112  49.329us  44.260us  69.957us  trt_volta_fp32_i8816cudnn_int8_128x128_ldg16_relu_small_nt_v1
      API calls:  100.00%  1.6523ms       112  14.752us  12.225us  34.562us  cudaLaunchKernel

==25751==       Range "model.20.m.0.cv2.conv.module.weight + QuantizeLinear_1527_quantize_scale_node + Conv_1531"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.4492ms       112  21.867us  17.185us  51.555us  model.20.m.0.cv2.conv.module.weight + QuantizeLinear_1527_quantize_scale_node + Conv_1531
 GPU activities:  100.00%  17.004ms       112  151.83us  146.67us  220.85us  trt_volta_fp32_i8816cudnn_int8_256x64_ldg16_relu_singleBuffer_small_nt_v1
      API calls:  100.00%  1.7919ms       112  15.999us  12.385us  42.883us  cudaLaunchKernel

==25751==       Range "model.20.m.1.cv1.conv.module.weight + QuantizeLinear_1543_quantize_scale_node + Conv_1547"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.3608ms       112  21.078us  17.153us  40.834us  model.20.m.1.cv1.conv.module.weight + QuantizeLinear_1543_quantize_scale_node + Conv_1547
 GPU activities:  100.00%  6.8739ms       112  61.373us  51.044us  606.86us  trt_volta_fp32_i8816cudnn_int8_128x128_ldg16_relu_small_nt_v1
      API calls:  100.00%  1.6810ms       112  15.008us  12.193us  35.842us  cudaLaunchKernel

==25751==       Range "model.20.m.1.cv2.conv.module.weight + QuantizeLinear_1559_quantize_scale_node + Conv_1563"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.3650ms       112  21.116us  16.865us  81.701us  model.20.m.1.cv2.conv.module.weight + QuantizeLinear_1559_quantize_scale_node + Conv_1563
 GPU activities:  100.00%  17.053ms       112  152.26us  147.02us  219.70us  trt_volta_fp32_i8816cudnn_int8_256x64_ldg16_relu_singleBuffer_small_nt_v1
      API calls:  100.00%  1.6936ms       112  15.121us  12.320us  43.427us  cudaLaunchKernel

==25751==       Range "model.20.m.2.cv1.conv.module.weight + QuantizeLinear_1575_quantize_scale_node + Conv_1579"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.3137ms       112  20.657us  16.865us  78.948us  model.20.m.2.cv1.conv.module.weight + QuantizeLinear_1575_quantize_scale_node + Conv_1579
 GPU activities:  100.00%  5.9627ms       112  53.238us  48.004us  75.142us  trt_volta_fp32_i8816cudnn_int8_128x128_ldg16_relu_small_nt_v1
      API calls:  100.00%  1.6603ms       112  14.823us  12.352us  73.316us  cudaLaunchKernel

==25751==       Range "model.20.m.2.cv2.conv.module.weight + QuantizeLinear_1591_quantize_scale_node + Conv_1595"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.3251ms       112  20.760us  16.833us  43.715us  model.20.m.2.cv2.conv.module.weight + QuantizeLinear_1591_quantize_scale_node + Conv_1595
 GPU activities:  100.00%  16.849ms       112  150.44us  144.81us  218.39us  trt_volta_fp32_i8816cudnn_int8_256x64_ldg16_relu_singleBuffer_small_nt_v1
      API calls:  100.00%  1.7059ms       112  15.231us  12.257us  36.546us  cudaLaunchKernel

==25751==       Range "model.21.conv.module.weight + QuantizeLinear_1640_quantize_scale_node + Conv_1644"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.5079ms       112  22.392us  17.089us  114.47us  model.21.conv.module.weight + QuantizeLinear_1640_quantize_scale_node + Conv_1644
 GPU activities:  100.00%  16.461ms       112  146.98us  141.32us  205.39us  trt_volta_fp32_i8816cudnn_int8_256x64_ldg16_relu_singleBuffer_small_nt_v1
      API calls:  100.00%  1.8733ms       112  16.725us  12.641us  106.95us  cudaLaunchKernel

==25751==       Range "model.23.cv1.conv.module.weight + QuantizeLinear_1657_quantize_scale_node + Conv_1661"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  4.2506ms       112  37.952us  16.993us  1.9135ms  model.23.cv1.conv.module.weight + QuantizeLinear_1657_quantize_scale_node + Conv_1661
 GPU activities:  100.00%  6.1043ms       112  54.502us  49.156us  361.08us  trt_volta_fp32_i8816cudnn_int8_128x128_ldg16_relu_small_nt_v1
      API calls:  100.00%  3.5549ms       112  31.740us  12.289us  1.9007ms  cudaLaunchKernel

==25751==       Range "model.23.cv2.conv.module.weight + QuantizeLinear_1769_quantize_scale_node + Conv_1773"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.4553ms       112  21.922us  16.769us  209.39us  model.23.cv2.conv.module.weight + QuantizeLinear_1769_quantize_scale_node + Conv_1773
 GPU activities:  100.00%  5.5198ms       112  49.283us  46.403us  70.597us  trt_volta_fp32_i8816cudnn_int8_128x128_ldg16_relu_small_nt_v1
      API calls:  100.00%  1.8560ms       112  16.571us  12.385us  197.26us  cudaLaunchKernel

==25751==       Range "model.23.cv3.conv.module.weight + QuantizeLinear_1786_quantize_scale_node + Conv_1790"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.5469ms       112  22.739us  17.121us  96.453us  model.23.cv3.conv.module.weight + QuantizeLinear_1786_quantize_scale_node + Conv_1790
 GPU activities:  100.00%  11.053ms       112  98.684us  88.935us  663.96us  trt_volta_fp32_i8816cudnn_int8_256x64_ldg16_relu_singleBuffer_small_nt_v1
      API calls:  100.00%  1.9133ms       112  17.082us  12.417us  87.493us  cudaLaunchKernel

==25751==       Range "model.23.m.0.cv1.conv.module.weight + QuantizeLinear_1673_quantize_scale_node + Conv_1677"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.3851ms       112  21.295us  16.705us  102.63us  model.23.m.0.cv1.conv.module.weight + QuantizeLinear_1673_quantize_scale_node + Conv_1677
 GPU activities:  100.00%  3.9755ms       112  35.495us  32.579us  51.140us  trt_volta_fp32_i8816cudnn_int8_128x128_ldg16_relu_small_nt_v1
      API calls:  100.00%  1.7835ms       112  15.924us  12.193us  94.405us  cudaLaunchKernel

==25751==       Range "model.23.m.0.cv2.conv.module.weight + QuantizeLinear_1689_quantize_scale_node + Conv_1693"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.2909ms       112  20.454us  16.769us  65.092us  model.23.m.0.cv2.conv.module.weight + QuantizeLinear_1689_quantize_scale_node + Conv_1693
 GPU activities:  100.00%  15.711ms       112  140.28us  136.14us  202.00us  trt_volta_fp32_i8816cudnn_int8_256x64_ldg16_relu_singleBuffer_small_nt_v1
      API calls:  100.00%  1.6048ms       112  14.328us  12.161us  24.033us  cudaLaunchKernel

==25751==       Range "model.23.m.1.cv1.conv.module.weight + QuantizeLinear_1705_quantize_scale_node + Conv_1709"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.4760ms       112  22.106us  17.089us  111.65us  model.23.m.1.cv1.conv.module.weight + QuantizeLinear_1705_quantize_scale_node + Conv_1709
 GPU activities:  100.00%  4.4539ms       112  39.766us  38.147us  55.812us  trt_volta_fp32_i8816cudnn_int8_128x128_ldg16_relu_small_nt_v1
      API calls:  100.00%  1.7528ms       112  15.650us  12.096us  103.08us  cudaLaunchKernel

==25751==       Range "model.23.m.1.cv2.conv.module.weight + QuantizeLinear_1721_quantize_scale_node + Conv_1725"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.4535ms       112  21.906us  16.961us  102.66us  model.23.m.1.cv2.conv.module.weight + QuantizeLinear_1721_quantize_scale_node + Conv_1725
 GPU activities:  100.00%  16.052ms       112  143.32us  136.75us  442.98us  trt_volta_fp32_i8816cudnn_int8_256x64_ldg16_relu_singleBuffer_small_nt_v1
      API calls:  100.00%  1.8246ms       112  16.290us  12.353us  95.238us  cudaLaunchKernel

==25751==       Range "model.23.m.2.cv1.conv.module.weight + QuantizeLinear_1737_quantize_scale_node + Conv_1741"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.5289ms       112  22.579us  16.673us  138.82us  model.23.m.2.cv1.conv.module.weight + QuantizeLinear_1737_quantize_scale_node + Conv_1741
 GPU activities:  100.00%  4.0982ms       112  36.590us  34.723us  51.908us  trt_volta_fp32_i8816cudnn_int8_128x128_ldg16_relu_small_nt_v1
      API calls:  100.00%  1.8775ms       112  16.763us  12.000us  131.30us  cudaLaunchKernel

==25751==       Range "model.23.m.2.cv2.conv.module.weight + QuantizeLinear_1753_quantize_scale_node + Conv_1757"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.4611ms       112  21.974us  16.737us  130.09us  model.23.m.2.cv2.conv.module.weight + QuantizeLinear_1753_quantize_scale_node + Conv_1757
 GPU activities:  100.00%  15.354ms       112  137.09us  130.54us  199.60us  trt_volta_fp32_i8816cudnn_int8_256x64_ldg16_relu_singleBuffer_small_nt_v1
      API calls:  100.00%  1.8444ms       112  16.467us  12.160us  122.41us  cudaLaunchKernel

==25751==       Range "model.3.conv.module.weight + QuantizeLinear_229_quantize_scale_node + Conv_233"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  3.2757ms       112  29.247us  18.401us  104.58us  model.3.conv.module.weight + QuantizeLinear_229_quantize_scale_node + Conv_233
 GPU activities:  100.00%  38.313ms       112  342.08us  301.69us  1.9295ms  trt_volta_fp32_i8816cudnn_int8_128x128_ldg16_relu_small_nt_v1
      API calls:  100.00%  2.3065ms       112  20.594us  13.281us  44.738us  cudaLaunchKernel

==25751==       Range "model.4.cv1.conv.module.weight + QuantizeLinear_245_quantize_scale_node + Conv_249"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  3.3654ms       112  30.048us  19.649us  125.64us  model.4.cv1.conv.module.weight + QuantizeLinear_245_quantize_scale_node + Conv_249
 GPU activities:  100.00%  12.136ms       112  108.36us  103.40us  151.95us  trt_volta_fp32_icudnn_int8x4_128x64_relu_interior_nn_v1
      API calls:  100.00%  2.3756ms       112  21.210us  13.441us  112.84us  cudaLaunchKernel

==25751==       Range "model.4.cv2.conv.module.weight + QuantizeLinear_558_quantize_scale_node + Conv_562"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  3.0600ms       112  27.321us  17.153us  98.150us  model.4.cv2.conv.module.weight + QuantizeLinear_558_quantize_scale_node + Conv_562
 GPU activities:  100.00%  10.532ms       112  94.033us  84.711us  143.72us  trt_volta_fp32_i8816cudnn_int8_128x128_ldg16_relu_small_nt_v1
      API calls:  100.00%  2.2793ms       112  20.350us  12.641us  90.085us  cudaLaunchKernel

==25751==       Range "model.4.cv3.conv.module.weight + QuantizeLinear_575_quantize_scale_node + Conv_579"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  3.1612ms       112  28.225us  18.689us  43.043us  model.4.cv3.conv.module.weight + QuantizeLinear_575_quantize_scale_node + Conv_579
 GPU activities:  100.00%  21.729ms       112  194.01us  185.46us  275.80us  trt_volta_fp32_icudnn_int8x4_128x64_relu_interior_nn_v1
      API calls:  100.00%  2.2389ms       112  19.990us  12.929us  35.106us  cudaLaunchKernel

==25751==       Range "model.4.m.0.cv1.conv.module.weight + QuantizeLinear_261_quantize_scale_node + Conv_265"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  3.2348ms       112  28.881us  17.569us  103.91us  model.4.m.0.cv1.conv.module.weight + QuantizeLinear_261_quantize_scale_node + Conv_265
 GPU activities:  100.00%  9.1225ms       112  81.450us  73.445us  114.38us  trt_volta_fp32_i8816cudnn_int8_128x128_ldg16_relu_small_nt_v1
      API calls:  100.00%  2.3983ms       112  21.413us  12.832us  96.709us  cudaLaunchKernel

==25751==       Range "model.4.m.0.cv2.conv.module.weight + QuantizeLinear_277_quantize_scale_node + Conv_281"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  3.3424ms       112  29.842us  17.697us  128.97us  model.4.m.0.cv2.conv.module.weight + QuantizeLinear_277_quantize_scale_node + Conv_281
 GPU activities:  100.00%  18.602ms       112  166.09us  149.87us  265.46us  trt_volta_fp32_i8816cudnn_int8_128x128_ldg16_relu_small_nt_v1
      API calls:  100.00%  2.5033ms       112  22.351us  12.897us  119.27us  cudaLaunchKernel

==25751==       Range "model.4.m.1.cv1.conv.module.weight + QuantizeLinear_294_quantize_scale_node + Conv_298"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  3.1601ms       112  28.215us  17.601us  114.60us  model.4.m.1.cv1.conv.module.weight + QuantizeLinear_294_quantize_scale_node + Conv_298
 GPU activities:  100.00%  9.0622ms       112  80.912us  75.110us  120.91us  trt_volta_fp32_i8816cudnn_int8_128x128_ldg16_relu_small_nt_v1
      API calls:  100.00%  2.3389ms       112  20.883us  13.088us  105.74us  cudaLaunchKernel

==25751==       Range "model.4.m.1.cv2.conv.module.weight + QuantizeLinear_310_quantize_scale_node + Conv_314"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  3.0676ms       112  27.389us  17.313us  77.060us  model.4.m.1.cv2.conv.module.weight + QuantizeLinear_310_quantize_scale_node + Conv_314
 GPU activities:  100.00%  18.737ms       112  167.29us  148.88us  248.76us  trt_volta_fp32_i8816cudnn_int8_128x128_ldg16_relu_small_nt_v1
      API calls:  100.00%  2.1941ms       112  19.590us  12.769us  65.155us  cudaLaunchKernel

==25751==       Range "model.4.m.2.cv1.conv.module.weight + QuantizeLinear_327_quantize_scale_node + Conv_331"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  3.1551ms       112  28.170us  18.369us  99.590us  model.4.m.2.cv1.conv.module.weight + QuantizeLinear_327_quantize_scale_node + Conv_331
 GPU activities:  100.00%  9.0487ms       112  80.792us  74.854us  118.92us  trt_volta_fp32_i8816cudnn_int8_128x128_ldg16_relu_small_nt_v1
      API calls:  100.00%  2.2842ms       112  20.394us  13.409us  91.845us  cudaLaunchKernel

==25751==       Range "model.4.m.2.cv2.conv.module.weight + QuantizeLinear_343_quantize_scale_node + Conv_347"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  3.0656ms       112  27.371us  17.217us  109.29us  model.4.m.2.cv2.conv.module.weight + QuantizeLinear_343_quantize_scale_node + Conv_347
 GPU activities:  100.00%  19.436ms       112  173.54us  153.29us  924.33us  trt_volta_fp32_i8816cudnn_int8_128x128_ldg16_relu_small_nt_v1
      API calls:  100.00%  2.2302ms       112  19.912us  12.481us  90.533us  cudaLaunchKernel

==25751==       Range "model.4.m.3.cv1.conv.module.weight + QuantizeLinear_360_quantize_scale_node + Conv_364"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  3.0474ms       112  27.208us  18.178us  57.635us  model.4.m.3.cv1.conv.module.weight + QuantizeLinear_360_quantize_scale_node + Conv_364
 GPU activities:  100.00%  9.0709ms       112  80.989us  73.542us  120.20us  trt_volta_fp32_i8816cudnn_int8_128x128_ldg16_relu_small_nt_v1
      API calls:  100.00%  2.2416ms       112  20.013us  13.248us  50.563us  cudaLaunchKernel

==25751==       Range "model.4.m.3.cv2.conv.module.weight + QuantizeLinear_376_quantize_scale_node + Conv_380"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.9213ms       112  26.083us  17.089us  43.043us  model.4.m.3.cv2.conv.module.weight + QuantizeLinear_376_quantize_scale_node + Conv_380
 GPU activities:  100.00%  18.953ms       112  169.22us  154.89us  249.01us  trt_volta_fp32_i8816cudnn_int8_128x128_ldg16_relu_small_nt_v1
      API calls:  100.00%  2.1011ms       112  18.759us  12.641us  32.674us  cudaLaunchKernel

==25751==       Range "model.4.m.4.cv1.conv.module.weight + QuantizeLinear_393_quantize_scale_node + Conv_397"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  3.0732ms       112  27.439us  17.537us  52.611us  model.4.m.4.cv1.conv.module.weight + QuantizeLinear_393_quantize_scale_node + Conv_397
 GPU activities:  100.00%  9.0913ms       112  81.172us  74.918us  127.40us  trt_volta_fp32_i8816cudnn_int8_128x128_ldg16_relu_small_nt_v1
      API calls:  100.00%  2.2241ms       112  19.857us  12.832us  35.874us  cudaLaunchKernel

==25751==       Range "model.4.m.4.cv2.conv.module.weight + QuantizeLinear_409_quantize_scale_node + Conv_413"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  3.0670ms       112  27.383us  17.281us  119.14us  model.4.m.4.cv2.conv.module.weight + QuantizeLinear_409_quantize_scale_node + Conv_413
 GPU activities:  100.00%  19.040ms       112  170.00us  153.84us  244.76us  trt_volta_fp32_i8816cudnn_int8_128x128_ldg16_relu_small_nt_v1
      API calls:  100.00%  2.2328ms       112  19.935us  12.321us  109.32us  cudaLaunchKernel

==25751==       Range "model.4.m.5.cv1.conv.module.weight + QuantizeLinear_426_quantize_scale_node + Conv_430"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  3.0312ms       112  27.064us  17.857us  48.067us  model.4.m.5.cv1.conv.module.weight + QuantizeLinear_426_quantize_scale_node + Conv_430
 GPU activities:  100.00%  8.9812ms       112  80.189us  73.222us  117.51us  trt_volta_fp32_i8816cudnn_int8_128x128_ldg16_relu_small_nt_v1
      API calls:  100.00%  2.2297ms       112  19.907us  13.120us  36.962us  cudaLaunchKernel

==25751==       Range "model.4.m.5.cv2.conv.module.weight + QuantizeLinear_442_quantize_scale_node + Conv_446"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  3.0013ms       112  26.797us  17.761us  84.965us  model.4.m.5.cv2.conv.module.weight + QuantizeLinear_442_quantize_scale_node + Conv_446
 GPU activities:  100.00%  18.838ms       112  168.19us  153.32us  246.04us  trt_volta_fp32_i8816cudnn_int8_128x128_ldg16_relu_small_nt_v1
      API calls:  100.00%  2.2090ms       112  19.723us  12.609us  77.828us  cudaLaunchKernel

==25751==       Range "model.4.m.6.cv1.conv.module.weight + QuantizeLinear_459_quantize_scale_node + Conv_463"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  3.1036ms       112  27.711us  17.409us  74.244us  model.4.m.6.cv1.conv.module.weight + QuantizeLinear_459_quantize_scale_node + Conv_463
 GPU activities:  100.00%  9.1012ms       112  81.260us  74.054us  114.92us  trt_volta_fp32_i8816cudnn_int8_128x128_ldg16_relu_small_nt_v1
      API calls:  100.00%  2.2769ms       112  20.329us  12.769us  50.146us  cudaLaunchKernel

==25751==       Range "model.4.m.6.cv2.conv.module.weight + QuantizeLinear_475_quantize_scale_node + Conv_479"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  3.1593ms       112  28.208us  17.569us  104.26us  model.4.m.6.cv2.conv.module.weight + QuantizeLinear_475_quantize_scale_node + Conv_479
 GPU activities:  100.00%  18.847ms       112  168.27us  150.73us  238.35us  trt_volta_fp32_i8816cudnn_int8_128x128_ldg16_relu_small_nt_v1
      API calls:  100.00%  2.2097ms       112  19.729us  12.481us  80.517us  cudaLaunchKernel

==25751==       Range "model.4.m.7.cv1.conv.module.weight + QuantizeLinear_492_quantize_scale_node + Conv_496"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  3.1471ms       112  28.099us  18.017us  91.750us  model.4.m.7.cv1.conv.module.weight + QuantizeLinear_492_quantize_scale_node + Conv_496
 GPU activities:  100.00%  9.0329ms       112  80.650us  74.758us  124.30us  trt_volta_fp32_i8816cudnn_int8_128x128_ldg16_relu_small_nt_v1
      API calls:  100.00%  2.2292ms       112  19.903us  13.185us  56.995us  cudaLaunchKernel

==25751==       Range "model.4.m.7.cv2.conv.module.weight + QuantizeLinear_508_quantize_scale_node + Conv_512"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.9106ms       112  25.987us  17.121us  97.638us  model.4.m.7.cv2.conv.module.weight + QuantizeLinear_508_quantize_scale_node + Conv_512
 GPU activities:  100.00%  18.841ms       112  168.22us  154.25us  238.29us  trt_volta_fp32_i8816cudnn_int8_128x128_ldg16_relu_small_nt_v1
      API calls:  100.00%  2.1410ms       112  19.115us  12.577us  89.734us  cudaLaunchKernel

==25751==       Range "model.4.m.8.cv1.conv.module.weight + QuantizeLinear_525_quantize_scale_node + Conv_529"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  3.0199ms       112  26.962us  17.857us  69.028us  model.4.m.8.cv1.conv.module.weight + QuantizeLinear_525_quantize_scale_node + Conv_529
 GPU activities:  100.00%  8.9688ms       112  80.078us  70.373us  116.59us  trt_volta_fp32_i8816cudnn_int8_128x128_ldg16_relu_small_nt_v1
      API calls:  100.00%  2.2497ms       112  20.086us  13.120us  60.388us  cudaLaunchKernel

==25751==       Range "model.4.m.8.cv2.conv.module.weight + QuantizeLinear_541_quantize_scale_node + Conv_545"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.8673ms       112  25.600us  17.473us  43.842us  model.4.m.8.cv2.conv.module.weight + QuantizeLinear_541_quantize_scale_node + Conv_545
 GPU activities:  100.00%  18.763ms       112  167.53us  152.59us  244.95us  trt_volta_fp32_i8816cudnn_int8_128x128_ldg16_relu_small_nt_v1
      API calls:  100.00%  2.0873ms       112  18.636us  12.705us  35.330us  cudaLaunchKernel

==25751==       Range "model.5.conv.module.weight + QuantizeLinear_591_quantize_scale_node + Conv_595"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  3.0286ms       112  27.041us  17.953us  54.819us  model.5.conv.module.weight + QuantizeLinear_591_quantize_scale_node + Conv_595
 GPU activities:  100.00%  33.454ms       112  298.70us  275.86us  417.31us  trt_volta_fp32_i8816cudnn_int8_128x128_ldg16_relu_small_nt_v1
      API calls:  100.00%  2.2125ms       112  19.754us  13.344us  41.826us  cudaLaunchKernel

==25751==       Range "model.6.cv1.conv.module.weight + QuantizeLinear_607_quantize_scale_node + Conv_611"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.9742ms       112  26.554us  17.185us  53.731us  model.6.cv1.conv.module.weight + QuantizeLinear_607_quantize_scale_node + Conv_611
 GPU activities:  100.00%  7.3939ms       112  66.016us  60.293us  94.151us  trt_volta_fp32_i8816cudnn_int8_128x128_ldg16_relu_small_nt_v1
      API calls:  100.00%  2.1673ms       112  19.351us  12.705us  47.651us  cudaLaunchKernel

==25751==       Range "model.6.cv2.conv.module.weight + QuantizeLinear_920_quantize_scale_node + Conv_924"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  3.0094ms       112  26.870us  17.025us  120.39us  model.6.cv2.conv.module.weight + QuantizeLinear_920_quantize_scale_node + Conv_924
 GPU activities:  100.00%  7.2938ms       112  65.122us  58.980us  95.272us  trt_volta_fp32_i8816cudnn_int8_128x128_ldg16_relu_small_nt_v1
      API calls:  100.00%  2.2882ms       112  20.430us  12.704us  113.32us  cudaLaunchKernel

==25751==       Range "model.6.cv3.conv.module.weight + QuantizeLinear_937_quantize_scale_node + Conv_941"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.9440ms       112  26.286us  18.849us  70.052us  model.6.cv3.conv.module.weight + QuantizeLinear_937_quantize_scale_node + Conv_941
 GPU activities:  100.00%  21.110ms       112  188.48us  178.70us  493.32us  trt_volta_fp32_icudnn_int8x4_128x64_relu_interior_nn_v1
      API calls:  100.00%  2.0414ms       112  18.226us  13.313us  58.947us  cudaLaunchKernel

==25751==       Range "model.6.m.0.cv1.conv.module.weight + QuantizeLinear_623_quantize_scale_node + Conv_627"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.8184ms       112  25.163us  17.121us  41.026us  model.6.m.0.cv1.conv.module.weight + QuantizeLinear_623_quantize_scale_node + Conv_627
 GPU activities:  100.00%  6.2700ms       112  55.982us  51.748us  78.982us  trt_volta_fp32_i8816cudnn_int8_128x128_ldg16_relu_small_nt_v1
      API calls:  100.00%  2.0624ms       112  18.414us  12.384us  33.698us  cudaLaunchKernel

==25751==       Range "model.6.m.0.cv2.conv.module.weight + QuantizeLinear_639_quantize_scale_node + Conv_643"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  3.1452ms       112  28.082us  20.577us  77.701us  model.6.m.0.cv2.conv.module.weight + QuantizeLinear_639_quantize_scale_node + Conv_643
 GPU activities:  100.00%  17.170ms       112  153.31us  148.56us  221.43us  trt_volta_fp32_i8816cudnn_int8_256x64_ldg16_relu_singleBuffer_small_nt_v1
      API calls:  100.00%  2.3022ms       112  20.555us  14.785us  63.427us  cudaLaunchKernel

==25751==       Range "model.6.m.1.cv1.conv.module.weight + QuantizeLinear_656_quantize_scale_node + Conv_660"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.8956ms       112  25.853us  18.113us  72.101us  model.6.m.1.cv1.conv.module.weight + QuantizeLinear_656_quantize_scale_node + Conv_660
 GPU activities:  100.00%  6.2326ms       112  55.647us  51.652us  86.727us  trt_volta_fp32_i8816cudnn_int8_128x128_ldg16_relu_small_nt_v1
      API calls:  100.00%  2.1348ms       112  19.061us  12.993us  64.771us  cudaLaunchKernel

==25751==       Range "model.6.m.1.cv2.conv.module.weight + QuantizeLinear_672_quantize_scale_node + Conv_676"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.9846ms       112  26.648us  18.017us  87.397us  model.6.m.1.cv2.conv.module.weight + QuantizeLinear_672_quantize_scale_node + Conv_676
 GPU activities:  100.00%  17.144ms       112  153.07us  147.88us  219.95us  trt_volta_fp32_i8816cudnn_int8_256x64_ldg16_relu_singleBuffer_small_nt_v1
      API calls:  100.00%  2.1794ms       112  19.458us  12.705us  70.340us  cudaLaunchKernel

==25751==       Range "model.6.m.2.cv1.conv.module.weight + QuantizeLinear_689_quantize_scale_node + Conv_693"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  3.0427ms       112  27.167us  19.073us  156.43us  model.6.m.2.cv1.conv.module.weight + QuantizeLinear_689_quantize_scale_node + Conv_693
 GPU activities:  100.00%  6.2047ms       112  55.399us  51.044us  79.014us  trt_volta_fp32_i8816cudnn_int8_128x128_ldg16_relu_small_nt_v1
      API calls:  100.00%  2.2671ms       112  20.242us  13.697us  148.10us  cudaLaunchKernel

==25751==       Range "model.6.m.2.cv2.conv.module.weight + QuantizeLinear_705_quantize_scale_node + Conv_709"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.7566ms       112  24.612us  16.993us  48.643us  model.6.m.2.cv2.conv.module.weight + QuantizeLinear_705_quantize_scale_node + Conv_709
 GPU activities:  100.00%  17.134ms       112  152.98us  147.88us  221.97us  trt_volta_fp32_i8816cudnn_int8_256x64_ldg16_relu_singleBuffer_small_nt_v1
      API calls:  100.00%  2.0186ms       112  18.022us  12.577us  38.210us  cudaLaunchKernel

==25751==       Range "model.6.m.3.cv1.conv.module.weight + QuantizeLinear_722_quantize_scale_node + Conv_726"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.9566ms       112  26.398us  18.049us  93.189us  model.6.m.3.cv1.conv.module.weight + QuantizeLinear_722_quantize_scale_node + Conv_726
 GPU activities:  100.00%  6.2555ms       112  55.852us  52.132us  78.502us  trt_volta_fp32_i8816cudnn_int8_128x128_ldg16_relu_small_nt_v1
      API calls:  100.00%  2.1397ms       112  19.104us  13.088us  80.900us  cudaLaunchKernel

==25751==       Range "model.6.m.3.cv2.conv.module.weight + QuantizeLinear_738_quantize_scale_node + Conv_742"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.8454ms       112  25.405us  17.729us  77.156us  model.6.m.3.cv2.conv.module.weight + QuantizeLinear_738_quantize_scale_node + Conv_742
 GPU activities:  100.00%  17.122ms       112  152.88us  146.86us  221.71us  trt_volta_fp32_i8816cudnn_int8_256x64_ldg16_relu_singleBuffer_small_nt_v1
      API calls:  100.00%  2.0963ms       112  18.717us  12.673us  70.372us  cudaLaunchKernel

==25751==       Range "model.6.m.4.cv1.conv.module.weight + QuantizeLinear_755_quantize_scale_node + Conv_759"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.8134ms       112  25.119us  17.889us  80.740us  model.6.m.4.cv1.conv.module.weight + QuantizeLinear_755_quantize_scale_node + Conv_759
 GPU activities:  100.00%  6.2767ms       112  56.042us  52.516us  79.334us  trt_volta_fp32_i8816cudnn_int8_128x128_ldg16_relu_small_nt_v1
      API calls:  100.00%  2.1015ms       112  18.763us  12.897us  73.284us  cudaLaunchKernel

==25751==       Range "model.6.m.4.cv2.conv.module.weight + QuantizeLinear_771_quantize_scale_node + Conv_775"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  3.6657ms       112  32.729us  17.697us  965.50us  model.6.m.4.cv2.conv.module.weight + QuantizeLinear_771_quantize_scale_node + Conv_775
 GPU activities:  100.00%  17.098ms       112  152.66us  147.05us  220.59us  trt_volta_fp32_i8816cudnn_int8_256x64_ldg16_relu_singleBuffer_small_nt_v1
      API calls:  100.00%  2.9488ms       112  26.328us  12.993us  955.29us  cudaLaunchKernel

==25751==       Range "model.6.m.5.cv1.conv.module.weight + QuantizeLinear_788_quantize_scale_node + Conv_792"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.6988ms       112  24.096us  17.889us  60.836us  model.6.m.5.cv1.conv.module.weight + QuantizeLinear_788_quantize_scale_node + Conv_792
 GPU activities:  100.00%  6.1944ms       112  55.307us  50.308us  79.750us  trt_volta_fp32_i8816cudnn_int8_128x128_ldg16_relu_small_nt_v1
      API calls:  100.00%  1.9759ms       112  17.642us  13.089us  54.883us  cudaLaunchKernel

==25751==       Range "model.6.m.5.cv2.conv.module.weight + QuantizeLinear_804_quantize_scale_node + Conv_808"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.6492ms       112  23.653us  16.929us  45.091us  model.6.m.5.cv2.conv.module.weight + QuantizeLinear_804_quantize_scale_node + Conv_808
 GPU activities:  100.00%  17.096ms       112  152.64us  148.08us  220.27us  trt_volta_fp32_i8816cudnn_int8_256x64_ldg16_relu_singleBuffer_small_nt_v1
      API calls:  100.00%  1.9511ms       112  17.420us  12.416us  38.082us  cudaLaunchKernel

==25751==       Range "model.6.m.6.cv1.conv.module.weight + QuantizeLinear_821_quantize_scale_node + Conv_825"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.5850ms       112  23.080us  17.377us  34.882us  model.6.m.6.cv1.conv.module.weight + QuantizeLinear_821_quantize_scale_node + Conv_825
 GPU activities:  100.00%  6.2223ms       112  55.556us  51.780us  80.679us  trt_volta_fp32_i8816cudnn_int8_128x128_ldg16_relu_small_nt_v1
      API calls:  100.00%  1.8876ms       112  16.853us  12.864us  28.705us  cudaLaunchKernel

==25751==       Range "model.6.m.6.cv2.conv.module.weight + QuantizeLinear_837_quantize_scale_node + Conv_841"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.5982ms       112  23.198us  17.537us  58.500us  model.6.m.6.cv2.conv.module.weight + QuantizeLinear_837_quantize_scale_node + Conv_841
 GPU activities:  100.00%  17.171ms       112  153.31us  148.11us  219.76us  trt_volta_fp32_i8816cudnn_int8_256x64_ldg16_relu_singleBuffer_small_nt_v1
      API calls:  100.00%  1.9083ms       112  17.038us  12.961us  51.426us  cudaLaunchKernel

==25751==       Range "model.6.m.7.cv1.conv.module.weight + QuantizeLinear_854_quantize_scale_node + Conv_858"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.7882ms       112  24.894us  18.049us  90.533us  model.6.m.7.cv1.conv.module.weight + QuantizeLinear_854_quantize_scale_node + Conv_858
 GPU activities:  100.00%  6.2484ms       112  55.789us  51.140us  79.622us  trt_volta_fp32_i8816cudnn_int8_128x128_ldg16_relu_small_nt_v1
      API calls:  100.00%  2.0934ms       112  18.691us  12.929us  82.052us  cudaLaunchKernel

==25751==       Range "model.6.m.7.cv2.conv.module.weight + QuantizeLinear_870_quantize_scale_node + Conv_874"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.6406ms       112  23.576us  17.697us  97.414us  model.6.m.7.cv2.conv.module.weight + QuantizeLinear_870_quantize_scale_node + Conv_874
 GPU activities:  100.00%  17.196ms       112  153.54us  148.24us  220.91us  trt_volta_fp32_i8816cudnn_int8_256x64_ldg16_relu_singleBuffer_small_nt_v1
      API calls:  100.00%  1.8825ms       112  16.808us  12.800us  45.955us  cudaLaunchKernel

==25751==       Range "model.6.m.8.cv1.conv.module.weight + QuantizeLinear_887_quantize_scale_node + Conv_891"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.4856ms       112  22.192us  17.569us  44.067us  model.6.m.8.cv1.conv.module.weight + QuantizeLinear_887_quantize_scale_node + Conv_891
 GPU activities:  100.00%  6.1955ms       112  55.316us  50.916us  80.550us  trt_volta_fp32_i8816cudnn_int8_128x128_ldg16_relu_small_nt_v1
      API calls:  100.00%  1.8452ms       112  16.474us  12.865us  37.282us  cudaLaunchKernel

==25751==       Range "model.6.m.8.cv2.conv.module.weight + QuantizeLinear_903_quantize_scale_node + Conv_907"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.7394ms       112  24.458us  17.313us  90.789us  model.6.m.8.cv2.conv.module.weight + QuantizeLinear_903_quantize_scale_node + Conv_907
 GPU activities:  100.00%  17.199ms       112  153.56us  147.98us  220.75us  trt_volta_fp32_i8816cudnn_int8_256x64_ldg16_relu_singleBuffer_small_nt_v1
      API calls:  100.00%  1.9480ms       112  17.392us  12.737us  58.243us  cudaLaunchKernel

==25751==       Range "model.7.conv.module.weight + QuantizeLinear_953_quantize_scale_node + Conv_957"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.7404ms       112  24.467us  17.633us  97.702us  model.7.conv.module.weight + QuantizeLinear_953_quantize_scale_node + Conv_957
 GPU activities:  100.00%  30.457ms       112  271.94us  263.35us  393.50us  trt_volta_fp32_i8816cudnn_int8_256x64_ldg16_relu_singleBuffer_small_nt_v1
      API calls:  100.00%  2.0069ms       112  17.918us  12.737us  74.500us  cudaLaunchKernel

==25751==       Range "model.8.cv1.conv.module.weight + QuantizeLinear_969_quantize_scale_node + Conv_973"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.4865ms       112  22.200us  17.121us  73.220us  model.8.cv1.conv.module.weight + QuantizeLinear_969_quantize_scale_node + Conv_973
 GPU activities:  100.00%  5.9063ms       112  52.735us  49.956us  75.014us  trt_volta_fp32_i8816cudnn_int8_128x128_ldg16_relu_small_nt_v1
      API calls:  100.00%  1.8262ms       112  16.304us  12.609us  66.660us  cudaLaunchKernel

==25751==       Range "model.8.cv2.conv.module.weight + QuantizeLinear_989_quantize_scale_node + Conv_993"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.5989ms       112  23.204us  17.761us  103.27us  model.8.cv2.conv.module.weight + QuantizeLinear_989_quantize_scale_node + Conv_993
 GPU activities:  100.00%  16.721ms       112  149.29us  141.42us  212.56us  trt_volta_fp32_i8816cudnn_int8_128x128_ldg16_relu_small_nt_v1
      API calls:  100.00%  1.8719ms       112  16.713us  12.800us  94.693us  cudaLaunchKernel

==25751==       Range "model.9.cv1.conv.module.weight + QuantizeLinear_1005_quantize_scale_node + Conv_1009"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.4799ms       112  22.141us  16.993us  62.787us  model.9.cv1.conv.module.weight + QuantizeLinear_1005_quantize_scale_node + Conv_1009
 GPU activities:  100.00%  5.6622ms       112  50.555us  48.163us  71.014us  trt_volta_fp32_i8816cudnn_int8_128x128_ldg16_relu_small_nt_v1
      API calls:  100.00%  1.7358ms       112  15.498us  12.481us  35.042us  cudaLaunchKernel

==25751==       Range "model.9.cv2.conv.module.weight + QuantizeLinear_1117_quantize_scale_node + Conv_1121"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.6557ms       112  23.711us  16.993us  208.88us  model.9.cv2.conv.module.weight + QuantizeLinear_1117_quantize_scale_node + Conv_1121
 GPU activities:  100.00%  5.5571ms       112  49.617us  47.620us  70.502us  trt_volta_fp32_i8816cudnn_int8_128x128_ldg16_relu_small_nt_v1
      API calls:  100.00%  1.9797ms       112  17.676us  12.545us  152.20us  cudaLaunchKernel

==25751==       Range "model.9.cv3.conv.module.weight + QuantizeLinear_1134_quantize_scale_node + Conv_1138"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.4279ms       112  21.677us  16.961us  63.940us  model.9.cv3.conv.module.weight + QuantizeLinear_1134_quantize_scale_node + Conv_1138
 GPU activities:  100.00%  10.275ms       112  91.739us  86.535us  134.06us  trt_volta_fp32_i8816cudnn_int8_256x64_ldg16_relu_singleBuffer_small_nt_v1
      API calls:  100.00%  1.7930ms       112  16.008us  12.545us  58.404us  cudaLaunchKernel

==25751==       Range "model.9.m.0.cv1.conv.module.weight + QuantizeLinear_1021_quantize_scale_node + Conv_1025"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.5074ms       112  22.387us  16.833us  66.595us  model.9.m.0.cv1.conv.module.weight + QuantizeLinear_1021_quantize_scale_node + Conv_1025
 GPU activities:  100.00%  4.0167ms       112  35.863us  32.834us  52.068us  trt_volta_fp32_i8816cudnn_int8_128x128_ldg16_relu_small_nt_v1
      API calls:  100.00%  1.8246ms       112  16.291us  12.416us  59.587us  cudaLaunchKernel

==25751==       Range "model.9.m.0.cv2.conv.module.weight + QuantizeLinear_1037_quantize_scale_node + Conv_1041"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.4945ms       112  22.272us  17.825us  51.843us  model.9.m.0.cv2.conv.module.weight + QuantizeLinear_1037_quantize_scale_node + Conv_1041
 GPU activities:  100.00%  16.050ms       112  143.30us  138.28us  201.36us  trt_volta_fp32_i8816cudnn_int8_256x64_ldg16_relu_singleBuffer_small_nt_v1
      API calls:  100.00%  1.8226ms       112  16.273us  13.185us  41.058us  cudaLaunchKernel

==25751==       Range "model.9.m.1.cv1.conv.module.weight + QuantizeLinear_1053_quantize_scale_node + Conv_1057"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.3506ms       112  20.987us  17.217us  40.706us  model.9.m.1.cv1.conv.module.weight + QuantizeLinear_1053_quantize_scale_node + Conv_1057
 GPU activities:  100.00%  4.0677ms       112  36.318us  34.883us  52.292us  trt_volta_fp32_i8816cudnn_int8_128x128_ldg16_relu_small_nt_v1
      API calls:  100.00%  1.7185ms       112  15.343us  12.673us  33.442us  cudaLaunchKernel

==25751==       Range "model.9.m.1.cv2.conv.module.weight + QuantizeLinear_1069_quantize_scale_node + Conv_1073"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  3.1909ms       112  28.490us  17.025us  734.41us  model.9.m.1.cv2.conv.module.weight + QuantizeLinear_1069_quantize_scale_node + Conv_1073
 GPU activities:  100.00%  15.426ms       112  137.73us  133.26us  199.18us  trt_volta_fp32_i8816cudnn_int8_256x64_ldg16_relu_singleBuffer_small_nt_v1
      API calls:  100.00%  2.5205ms       112  22.504us  12.577us  726.73us  cudaLaunchKernel

==25751==       Range "model.9.m.2.cv1.conv.module.weight + QuantizeLinear_1085_quantize_scale_node + Conv_1089"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.4335ms       112  21.727us  17.185us  59.395us  model.9.m.2.cv1.conv.module.weight + QuantizeLinear_1085_quantize_scale_node + Conv_1089
 GPU activities:  100.00%  4.1454ms       112  37.012us  35.298us  52.900us  trt_volta_fp32_i8816cudnn_int8_128x128_ldg16_relu_small_nt_v1
      API calls:  100.00%  1.8008ms       112  16.078us  12.673us  54.147us  cudaLaunchKernel

==25751==       Range "model.9.m.2.cv2.conv.module.weight + QuantizeLinear_1101_quantize_scale_node + Conv_1105"
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
          Range:  100.00%  2.6298ms       112  23.480us  17.377us  84.357us  model.9.m.2.cv2.conv.module.weight + QuantizeLinear_1101_quantize_scale_node + Conv_1105
 GPU activities:  100.00%  15.513ms       112  138.51us  133.61us  198.90us  trt_volta_fp32_i8816cudnn_int8_256x64_ldg16_relu_singleBuffer_small_nt_v1
      API calls:  100.00%  1.9652ms       112  17.546us  12.513us  78.853us  cudaLaunchKernel