(venv) tech@Artax:/opt/metamirror/onnx_trt_utils$ polygraphy run rvm_mobilenetv3_fp32_sim_modified_fp32.onnx --trt --onnxrt -vv [V] Model: rvm_mobilenetv3_fp32_sim_modified_fp32.onnx [I] RUNNING | Command: /opt/venv/bin/polygraphy run rvm_mobilenetv3_fp32_sim_modified_fp32.onnx --trt --onnxrt -vv [V] Loaded Module: polygraphy | Version: 0.44.2 | Path: ['/opt/venv/lib/python3.10/site-packages/polygraphy'] [V] Loaded extension modules: [] [V] Loaded Module: tensorrt | Version: 8.5.1.7 | Path: ['/opt/venv/lib/python3.10/site-packages/tensorrt'] [I] trt-runner-N0-02/13/23-19:49:35 | Activating and starting inference [V] [MemUsageChange] Init CUDA: CPU +220, GPU +0, now: CPU 247, GPU 8271 (MiB) [V] [MemUsageChange] Init builder kernel library: CPU +351, GPU +459, now: CPU 617, GPU 8747 (MiB) [V] ---------------------------------------------------------------- [V] Input filename: /opt/metamirror/onnx_trt_utils/rvm_mobilenetv3_fp32_sim_modified_fp32.onnx [V] ONNX IR version: 0.0.8 [V] Opset version: 12 [V] Producer name: onnx-typecast [V] Producer version: [V] Domain: [V] Model version: 0 [V] Doc string: [V] ---------------------------------------------------------------- [X] Registered plugin creator - ::GridAnchor_TRT version 1 [X] Registered plugin creator - ::GridAnchorRect_TRT version 1 [X] Registered plugin creator - ::NMS_TRT version 1 [X] Registered plugin creator - ::Reorg_TRT version 1 [X] Registered plugin creator - ::Region_TRT version 1 [X] Registered plugin creator - ::Clip_TRT version 1 [X] Registered plugin creator - ::LReLU_TRT version 1 [X] Registered plugin creator - ::PriorBox_TRT version 1 [X] Registered plugin creator - ::Normalize_TRT version 1 [X] Registered plugin creator - ::ScatterND version 1 [X] Registered plugin creator - ::RPROI_TRT version 1 [X] Registered plugin creator - ::BatchedNMS_TRT version 1 [X] Registered plugin creator - ::BatchedNMSDynamic_TRT version 1 [X] Registered plugin creator - ::BatchTilePlugin_TRT version 1 [X] Registered plugin creator - ::FlattenConcat_TRT version 1 [X] Registered plugin creator - ::CropAndResize version 1 [X] Registered plugin creator - ::CropAndResizeDynamic version 1 [X] Registered plugin creator - ::DetectionLayer_TRT version 1 [X] Registered plugin creator - ::EfficientNMS_TRT version 1 [X] Registered plugin creator - ::EfficientNMS_ONNX_TRT version 1 [X] Registered plugin creator - ::EfficientNMS_Explicit_TF_TRT version 1 [X] Registered plugin creator - ::EfficientNMS_Implicit_TF_TRT version 1 [X] Registered plugin creator - ::ProposalDynamic version 1 [X] Registered plugin creator - ::Proposal version 1 [X] Registered plugin creator - ::ProposalLayer_TRT version 1 [X] Registered plugin creator - ::PyramidROIAlign_TRT version 1 [X] Registered plugin creator - ::ResizeNearest_TRT version 1 [X] Registered plugin creator - ::Split version 1 [X] Registered plugin creator - ::SpecialSlice_TRT version 1 [X] Registered plugin creator - ::InstanceNormalization_TRT version 1 [X] Registered plugin creator - ::InstanceNormalization_TRT version 2 [X] Registered plugin creator - ::CoordConvAC version 1 [X] Registered plugin creator - ::DecodeBbox3DPlugin version 1 [X] Registered plugin creator - ::GenerateDetection_TRT version 1 [X] Registered plugin creator - ::MultilevelCropAndResize_TRT version 1 [X] Registered plugin creator - ::MultilevelProposeROI_TRT version 1 [X] Registered plugin creator - ::NMSDynamic_TRT version 1 [X] Registered plugin creator - ::PillarScatterPlugin version 1 [X] Registered plugin creator - ::VoxelGeneratorPlugin version 1 [X] Registered plugin creator - ::MultiscaleDeformableAttnPlugin_TRT version 1 [X] Adding network input: src with dtype: float32, dimensions: (1, 3, 1440, 2560) [X] Registering tensor: src for ONNX tensor: src [X] Adding network input: r1i with dtype: float32, dimensions: (1, 1, 1, 1) [X] Registering tensor: r1i for ONNX tensor: r1i [X] Adding network input: r2i with dtype: float32, dimensions: (1, 1, 1, 1) [X] Registering tensor: r2i for ONNX tensor: r2i [X] Adding network input: r3i with dtype: float32, dimensions: (1, 1, 1, 1) [X] Registering tensor: r3i for ONNX tensor: r3i [X] Adding network input: r4i with dtype: float32, dimensions: (1, 1, 1, 1) [X] Registering tensor: r4i for ONNX tensor: r4i [X] Importing initializer: 386 [X] Importing initializer: 388 [X] Importing initializer: 390 [X] Importing initializer: 392 [X] Importing initializer: 824 [X] Importing initializer: 825 [X] Importing initializer: 827 [X] Importing initializer: 828 [X] Importing initializer: 830 [X] Importing initializer: 831 [X] Importing initializer: 833 [X] Importing initializer: 834 [X] Importing initializer: 836 [X] Importing initializer: 837 [X] Importing initializer: 839 [X] Importing initializer: 840 [X] Importing initializer: 842 [X] Importing initializer: 843 [X] Importing initializer: 845 [X] Importing initializer: 846 [X] Importing initializer: 848 [X] Importing initializer: 849 [X] Importing initializer: 851 [X] Importing initializer: 852 [X] Importing initializer: 854 [X] Importing initializer: 855 [X] Importing initializer: backbone.features.4.block.2.fc1.weight [X] Importing initializer: backbone.features.4.block.2.fc1.bias [X] Importing initializer: backbone.features.4.block.2.fc2.weight [X] Importing initializer: backbone.features.4.block.2.fc2.bias [X] Importing initializer: 857 [X] Importing initializer: 858 [X] Importing initializer: 860 [X] Importing initializer: 861 [X] Importing initializer: 863 [X] Importing initializer: 864 [X] Importing initializer: backbone.features.5.block.2.fc1.weight [X] Importing initializer: backbone.features.5.block.2.fc1.bias [X] Importing initializer: backbone.features.5.block.2.fc2.weight [X] Importing initializer: backbone.features.5.block.2.fc2.bias [X] Importing initializer: 866 [X] Importing initializer: 867 [X] Importing initializer: 869 [X] Importing initializer: 870 [X] Importing initializer: 872 [X] Importing initializer: 873 [X] Importing initializer: backbone.features.6.block.2.fc1.weight [X] Importing initializer: backbone.features.6.block.2.fc1.bias [X] Importing initializer: backbone.features.6.block.2.fc2.weight [X] Importing initializer: backbone.features.6.block.2.fc2.bias [X] Importing initializer: 875 [X] Importing initializer: 876 [X] Importing initializer: 878 [X] Importing initializer: 879 [X] Importing initializer: 881 [X] Importing initializer: 882 [X] Importing initializer: 884 [X] Importing initializer: 885 [X] Importing initializer: 887 [X] Importing initializer: 888 [X] Importing initializer: 890 [X] Importing initializer: 891 [X] Importing initializer: 893 [X] Importing initializer: 894 [X] Importing initializer: 896 [X] Importing initializer: 897 [X] Importing initializer: 899 [X] Importing initializer: 900 [X] Importing initializer: 902 [X] Importing initializer: 903 [X] Importing initializer: 905 [X] Importing initializer: 906 [X] Importing initializer: 908 [X] Importing initializer: 909 [X] Importing initializer: 911 [X] Importing initializer: 912 [X] Importing initializer: 914 [X] Importing initializer: 915 [X] Importing initializer: 917 [X] Importing initializer: 918 [X] Importing initializer: backbone.features.11.block.2.fc1.weight [X] Importing initializer: backbone.features.11.block.2.fc1.bias [X] Importing initializer: backbone.features.11.block.2.fc2.weight [X] Importing initializer: backbone.features.11.block.2.fc2.bias [X] Importing initializer: 920 [X] Importing initializer: 921 [X] Importing initializer: 923 [X] Importing initializer: 924 [X] Importing initializer: 926 [X] Importing initializer: 927 [X] Importing initializer: backbone.features.12.block.2.fc1.weight [X] Importing initializer: backbone.features.12.block.2.fc1.bias [X] Importing initializer: backbone.features.12.block.2.fc2.weight [X] Importing initializer: backbone.features.12.block.2.fc2.bias [X] Importing initializer: 929 [X] Importing initializer: 930 [X] Importing initializer: 932 [X] Importing initializer: 933 [X] Importing initializer: 935 [X] Importing initializer: 936 [X] Importing initializer: backbone.features.13.block.2.fc1.weight [X] Importing initializer: backbone.features.13.block.2.fc1.bias [X] Importing initializer: backbone.features.13.block.2.fc2.weight [X] Importing initializer: backbone.features.13.block.2.fc2.bias [X] Importing initializer: 938 [X] Importing initializer: 939 [X] Importing initializer: 941 [X] Importing initializer: 942 [X] Importing initializer: 944 [X] Importing initializer: 945 [X] Importing initializer: backbone.features.14.block.2.fc1.weight [X] Importing initializer: backbone.features.14.block.2.fc1.bias [X] Importing initializer: backbone.features.14.block.2.fc2.weight [X] Importing initializer: backbone.features.14.block.2.fc2.bias [X] Importing initializer: 947 [X] Importing initializer: 948 [X] Importing initializer: 950 [X] Importing initializer: 951 [X] Importing initializer: 953 [X] Importing initializer: 954 [X] Importing initializer: backbone.features.15.block.2.fc1.weight [X] Importing initializer: backbone.features.15.block.2.fc1.bias [X] Importing initializer: backbone.features.15.block.2.fc2.weight [X] Importing initializer: backbone.features.15.block.2.fc2.bias [X] Importing initializer: 956 [X] Importing initializer: 957 [X] Importing initializer: 959 [X] Importing initializer: 960 [X] Importing initializer: 962 [X] Importing initializer: 963 [X] Importing initializer: aspp.aspp2.1.weight [X] Importing initializer: decoder.decode4.gru.ih.0.weight [X] Importing initializer: decoder.decode4.gru.ih.0.bias [X] Importing initializer: decoder.decode4.gru.hh.0.weight [X] Importing initializer: decoder.decode4.gru.hh.0.bias [X] Importing initializer: 990 [X] Importing initializer: 985 [X] Importing initializer: 630 [X] Importing initializer: 631 [X] Importing initializer: 632 [X] Importing initializer: 634 [X] Importing initializer: 635 [X] Importing initializer: 965 [X] Importing initializer: 966 [X] Importing initializer: decoder.decode3.gru.ih.0.weight [X] Importing initializer: decoder.decode3.gru.ih.0.bias [X] Importing initializer: decoder.decode3.gru.hh.0.weight [X] Importing initializer: decoder.decode3.gru.hh.0.bias [X] Importing initializer: 968 [X] Importing initializer: 969 [X] Importing initializer: decoder.decode2.gru.ih.0.weight [X] Importing initializer: decoder.decode2.gru.ih.0.bias [X] Importing initializer: decoder.decode2.gru.hh.0.weight [X] Importing initializer: decoder.decode2.gru.hh.0.bias [X] Importing initializer: 971 [X] Importing initializer: 972 [X] Importing initializer: decoder.decode1.gru.ih.0.weight [X] Importing initializer: decoder.decode1.gru.ih.0.bias [X] Importing initializer: decoder.decode1.gru.hh.0.weight [X] Importing initializer: decoder.decode1.gru.hh.0.bias [X] Importing initializer: 974 [X] Importing initializer: 975 [X] Importing initializer: 977 [X] Importing initializer: 978 [X] Importing initializer: project_mat.conv.weight [X] Importing initializer: project_mat.conv.bias [X] Importing initializer: refiner.box_filter.weight [X] Importing initializer: 980 [X] Importing initializer: 981 [X] Importing initializer: 983 [X] Importing initializer: 984 [X] Importing initializer: refiner.conv.6.weight [X] Importing initializer: refiner.conv.6.bias [X] Importing initializer: 815 [X] Importing initializer: 989 [X] Parsing node: Resize_3 [Resize] [X] Searching for input: src [X] Searching for input: 386 [X] Searching for input: 388 [X] Resize_3 [Resize] inputs: [src -> (1, 3, 1440, 2560)[FLOAT]], [386 -> (0)[FLOAT]], [388 -> (4)[FLOAT]], [X] Registering layer: Resize_3 for ONNX node: Resize_3 [X] Running resize layer with: Transformation mode: pytorch_half_pixel Resize mode: linear [X] Registering tensor: 389 for ONNX tensor: 389 [X] Resize_3 [Resize] outputs: [389 -> (1, 3, 288, 512)[FLOAT]], [X] Parsing node: Sub_5 [Sub] [X] Searching for input: 389 [X] Searching for input: 390 [X] Sub_5 [Sub] inputs: [389 -> (1, 3, 288, 512)[FLOAT]], [390 -> (3, 1, 1)[FLOAT]], [X] Registering layer: 390 for ONNX node: 390 [X] Registering layer: Sub_5 for ONNX node: Sub_5 [X] Registering tensor: 391 for ONNX tensor: 391 [X] Sub_5 [Sub] outputs: [391 -> (1, 3, 288, 512)[FLOAT]], [X] Parsing node: Div_7 [Div] [X] Searching for input: 391 [X] Searching for input: 392 [X] Div_7 [Div] inputs: [391 -> (1, 3, 288, 512)[FLOAT]], [392 -> (3, 1, 1)[FLOAT]], [X] Registering layer: 392 for ONNX node: 392 [X] Registering layer: Div_7 for ONNX node: Div_7 [X] Registering tensor: 393 for ONNX tensor: 393 [X] Div_7 [Div] outputs: [393 -> (1, 3, 288, 512)[FLOAT]], [X] Parsing node: Conv_8 [Conv] [X] Searching for input: 393 [X] Searching for input: 824 [X] Searching for input: 825 [X] Conv_8 [Conv] inputs: [393 -> (1, 3, 288, 512)[FLOAT]], [824 -> (16, 3, 3, 3)[FLOAT]], [825 -> (16)[FLOAT]], [X] Convolution input dimensions: (1, 3, 288, 512) [X] Registering layer: Conv_8 for ONNX node: Conv_8 [X] Using kernel: (3, 3), strides: (2, 2), prepadding: (1, 1), postpadding: (1, 1), dilations: (1, 1), numOutputs: 16 [X] Convolution output dimensions: (1, 16, 144, 256) [X] Registering tensor: 823 for ONNX tensor: 823 [X] Conv_8 [Conv] outputs: [823 -> (1, 16, 144, 256)[FLOAT]], [X] Parsing node: HardSigmoid_9 [HardSigmoid] [X] Searching for input: 823 [X] HardSigmoid_9 [HardSigmoid] inputs: [823 -> (1, 16, 144, 256)[FLOAT]], [X] Registering layer: HardSigmoid_9 for ONNX node: HardSigmoid_9 [X] Registering tensor: 396 for ONNX tensor: 396 [X] HardSigmoid_9 [HardSigmoid] outputs: [396 -> (1, 16, 144, 256)[FLOAT]], [X] Parsing node: Mul_10 [Mul] [X] Searching for input: 823 [X] Searching for input: 396 [X] Mul_10 [Mul] inputs: [823 -> (1, 16, 144, 256)[FLOAT]], [396 -> (1, 16, 144, 256)[FLOAT]], [X] Registering layer: Mul_10 for ONNX node: Mul_10 [X] Registering tensor: 397 for ONNX tensor: 397 [X] Mul_10 [Mul] outputs: [397 -> (1, 16, 144, 256)[FLOAT]], [X] Parsing node: Conv_11 [Conv] [X] Searching for input: 397 [X] Searching for input: 827 [X] Searching for input: 828 [X] Conv_11 [Conv] inputs: [397 -> (1, 16, 144, 256)[FLOAT]], [827 -> (16, 1, 3, 3)[FLOAT]], [828 -> (16)[FLOAT]], [X] Convolution input dimensions: (1, 16, 144, 256) [X] Registering layer: Conv_11 for ONNX node: Conv_11 [X] Using kernel: (3, 3), strides: (1, 1), prepadding: (1, 1), postpadding: (1, 1), dilations: (1, 1), numOutputs: 16 [X] Convolution output dimensions: (1, 16, 144, 256) [X] Registering tensor: 826 for ONNX tensor: 826 [X] Conv_11 [Conv] outputs: [826 -> (1, 16, 144, 256)[FLOAT]], [X] Parsing node: Relu_12 [Relu] [X] Searching for input: 826 [X] Relu_12 [Relu] inputs: [826 -> (1, 16, 144, 256)[FLOAT]], [X] Registering layer: Relu_12 for ONNX node: Relu_12 [X] Registering tensor: 400 for ONNX tensor: 400 [X] Relu_12 [Relu] outputs: [400 -> (1, 16, 144, 256)[FLOAT]], [X] Parsing node: Conv_13 [Conv] [X] Searching for input: 400 [X] Searching for input: 830 [X] Searching for input: 831 [X] Conv_13 [Conv] inputs: [400 -> (1, 16, 144, 256)[FLOAT]], [830 -> (16, 16, 1, 1)[FLOAT]], [831 -> (16)[FLOAT]], [X] Convolution input dimensions: (1, 16, 144, 256) [X] Registering layer: Conv_13 for ONNX node: Conv_13 [X] Using kernel: (1, 1), strides: (1, 1), prepadding: (0, 0), postpadding: (0, 0), dilations: (1, 1), numOutputs: 16 [X] Convolution output dimensions: (1, 16, 144, 256) [X] Registering tensor: 829 for ONNX tensor: 829 [X] Conv_13 [Conv] outputs: [829 -> (1, 16, 144, 256)[FLOAT]], [X] Parsing node: Add_14 [Add] [X] Searching for input: 829 [X] Searching for input: 397 [X] Add_14 [Add] inputs: [829 -> (1, 16, 144, 256)[FLOAT]], [397 -> (1, 16, 144, 256)[FLOAT]], [X] Registering layer: Add_14 for ONNX node: Add_14 [X] Registering tensor: 403 for ONNX tensor: 403 [X] Add_14 [Add] outputs: [403 -> (1, 16, 144, 256)[FLOAT]], [X] Parsing node: Conv_15 [Conv] [X] Searching for input: 403 [X] Searching for input: 833 [X] Searching for input: 834 [X] Conv_15 [Conv] inputs: [403 -> (1, 16, 144, 256)[FLOAT]], [833 -> (64, 16, 1, 1)[FLOAT]], [834 -> (64)[FLOAT]], [X] Convolution input dimensions: (1, 16, 144, 256) [X] Registering layer: Conv_15 for ONNX node: Conv_15 [X] Using kernel: (1, 1), strides: (1, 1), prepadding: (0, 0), postpadding: (0, 0), dilations: (1, 1), numOutputs: 64 [X] Convolution output dimensions: (1, 64, 144, 256) [X] Registering tensor: 832 for ONNX tensor: 832 [X] Conv_15 [Conv] outputs: [832 -> (1, 64, 144, 256)[FLOAT]], [X] Parsing node: Relu_16 [Relu] [X] Searching for input: 832 [X] Relu_16 [Relu] inputs: [832 -> (1, 64, 144, 256)[FLOAT]], [X] Registering layer: Relu_16 for ONNX node: Relu_16 [X] Registering tensor: 406 for ONNX tensor: 406 [X] Relu_16 [Relu] outputs: [406 -> (1, 64, 144, 256)[FLOAT]], [X] Parsing node: Conv_17 [Conv] [X] Searching for input: 406 [X] Searching for input: 836 [X] Searching for input: 837 [X] Conv_17 [Conv] inputs: [406 -> (1, 64, 144, 256)[FLOAT]], [836 -> (64, 1, 3, 3)[FLOAT]], [837 -> (64)[FLOAT]], [X] Convolution input dimensions: (1, 64, 144, 256) [X] Registering layer: Conv_17 for ONNX node: Conv_17 [X] Using kernel: (3, 3), strides: (2, 2), prepadding: (1, 1), postpadding: (1, 1), dilations: (1, 1), numOutputs: 64 [X] Convolution output dimensions: (1, 64, 72, 128) [X] Registering tensor: 835 for ONNX tensor: 835 [X] Conv_17 [Conv] outputs: [835 -> (1, 64, 72, 128)[FLOAT]], [X] Parsing node: Relu_18 [Relu] [X] Searching for input: 835 [X] Relu_18 [Relu] inputs: [835 -> (1, 64, 72, 128)[FLOAT]], [X] Registering layer: Relu_18 for ONNX node: Relu_18 [X] Registering tensor: 409 for ONNX tensor: 409 [X] Relu_18 [Relu] outputs: [409 -> (1, 64, 72, 128)[FLOAT]], [X] Parsing node: Conv_19 [Conv] [X] Searching for input: 409 [X] Searching for input: 839 [X] Searching for input: 840 [X] Conv_19 [Conv] inputs: [409 -> (1, 64, 72, 128)[FLOAT]], [839 -> (24, 64, 1, 1)[FLOAT]], [840 -> (24)[FLOAT]], [X] Convolution input dimensions: (1, 64, 72, 128) [X] Registering layer: Conv_19 for ONNX node: Conv_19 [X] Using kernel: (1, 1), strides: (1, 1), prepadding: (0, 0), postpadding: (0, 0), dilations: (1, 1), numOutputs: 24 [X] Convolution output dimensions: (1, 24, 72, 128) [X] Registering tensor: 838 for ONNX tensor: 838 [X] Conv_19 [Conv] outputs: [838 -> (1, 24, 72, 128)[FLOAT]], [X] Parsing node: Conv_20 [Conv] [X] Searching for input: 838 [X] Searching for input: 842 [X] Searching for input: 843 [X] Conv_20 [Conv] inputs: [838 -> (1, 24, 72, 128)[FLOAT]], [842 -> (72, 24, 1, 1)[FLOAT]], [843 -> (72)[FLOAT]], [X] Convolution input dimensions: (1, 24, 72, 128) [X] Registering layer: Conv_20 for ONNX node: Conv_20 [X] Using kernel: (1, 1), strides: (1, 1), prepadding: (0, 0), postpadding: (0, 0), dilations: (1, 1), numOutputs: 72 [X] Convolution output dimensions: (1, 72, 72, 128) [X] Registering tensor: 841 for ONNX tensor: 841 [X] Conv_20 [Conv] outputs: [841 -> (1, 72, 72, 128)[FLOAT]], [X] Parsing node: Relu_21 [Relu] [X] Searching for input: 841 [X] Relu_21 [Relu] inputs: [841 -> (1, 72, 72, 128)[FLOAT]], [X] Registering layer: Relu_21 for ONNX node: Relu_21 [X] Registering tensor: 414 for ONNX tensor: 414 [X] Relu_21 [Relu] outputs: [414 -> (1, 72, 72, 128)[FLOAT]], [X] Parsing node: Conv_22 [Conv] [X] Searching for input: 414 [X] Searching for input: 845 [X] Searching for input: 846 [X] Conv_22 [Conv] inputs: [414 -> (1, 72, 72, 128)[FLOAT]], [845 -> (72, 1, 3, 3)[FLOAT]], [846 -> (72)[FLOAT]], [X] Convolution input dimensions: (1, 72, 72, 128) [X] Registering layer: Conv_22 for ONNX node: Conv_22 [X] Using kernel: (3, 3), strides: (1, 1), prepadding: (1, 1), postpadding: (1, 1), dilations: (1, 1), numOutputs: 72 [X] Convolution output dimensions: (1, 72, 72, 128) [X] Registering tensor: 844 for ONNX tensor: 844 [X] Conv_22 [Conv] outputs: [844 -> (1, 72, 72, 128)[FLOAT]], [X] Parsing node: Relu_23 [Relu] [X] Searching for input: 844 [X] Relu_23 [Relu] inputs: [844 -> (1, 72, 72, 128)[FLOAT]], [X] Registering layer: Relu_23 for ONNX node: Relu_23 [X] Registering tensor: 417 for ONNX tensor: 417 [X] Relu_23 [Relu] outputs: [417 -> (1, 72, 72, 128)[FLOAT]], [X] Parsing node: Conv_24 [Conv] [X] Searching for input: 417 [X] Searching for input: 848 [X] Searching for input: 849 [X] Conv_24 [Conv] inputs: [417 -> (1, 72, 72, 128)[FLOAT]], [848 -> (24, 72, 1, 1)[FLOAT]], [849 -> (24)[FLOAT]], [X] Convolution input dimensions: (1, 72, 72, 128) [X] Registering layer: Conv_24 for ONNX node: Conv_24 [X] Using kernel: (1, 1), strides: (1, 1), prepadding: (0, 0), postpadding: (0, 0), dilations: (1, 1), numOutputs: 24 [X] Convolution output dimensions: (1, 24, 72, 128) [X] Registering tensor: 847 for ONNX tensor: 847 [X] Conv_24 [Conv] outputs: [847 -> (1, 24, 72, 128)[FLOAT]], [X] Parsing node: Add_25 [Add] [X] Searching for input: 847 [X] Searching for input: 838 [X] Add_25 [Add] inputs: [847 -> (1, 24, 72, 128)[FLOAT]], [838 -> (1, 24, 72, 128)[FLOAT]], [X] Registering layer: Add_25 for ONNX node: Add_25 [X] Registering tensor: 420 for ONNX tensor: 420 [X] Add_25 [Add] outputs: [420 -> (1, 24, 72, 128)[FLOAT]], [X] Parsing node: Conv_26 [Conv] [X] Searching for input: 420 [X] Searching for input: 851 [X] Searching for input: 852 [X] Conv_26 [Conv] inputs: [420 -> (1, 24, 72, 128)[FLOAT]], [851 -> (72, 24, 1, 1)[FLOAT]], [852 -> (72)[FLOAT]], [X] Convolution input dimensions: (1, 24, 72, 128) [X] Registering layer: Conv_26 for ONNX node: Conv_26 [X] Using kernel: (1, 1), strides: (1, 1), prepadding: (0, 0), postpadding: (0, 0), dilations: (1, 1), numOutputs: 72 [X] Convolution output dimensions: (1, 72, 72, 128) [X] Registering tensor: 850 for ONNX tensor: 850 [X] Conv_26 [Conv] outputs: [850 -> (1, 72, 72, 128)[FLOAT]], [X] Parsing node: Relu_27 [Relu] [X] Searching for input: 850 [X] Relu_27 [Relu] inputs: [850 -> (1, 72, 72, 128)[FLOAT]], [X] Registering layer: Relu_27 for ONNX node: Relu_27 [X] Registering tensor: 423 for ONNX tensor: 423 [X] Relu_27 [Relu] outputs: [423 -> (1, 72, 72, 128)[FLOAT]], [X] Parsing node: Conv_28 [Conv] [X] Searching for input: 423 [X] Searching for input: 854 [X] Searching for input: 855 [X] Conv_28 [Conv] inputs: [423 -> (1, 72, 72, 128)[FLOAT]], [854 -> (72, 1, 5, 5)[FLOAT]], [855 -> (72)[FLOAT]], [X] Convolution input dimensions: (1, 72, 72, 128) [X] Registering layer: Conv_28 for ONNX node: Conv_28 [X] Using kernel: (5, 5), strides: (2, 2), prepadding: (2, 2), postpadding: (2, 2), dilations: (1, 1), numOutputs: 72 [X] Convolution output dimensions: (1, 72, 36, 64) [X] Registering tensor: 853 for ONNX tensor: 853 [X] Conv_28 [Conv] outputs: [853 -> (1, 72, 36, 64)[FLOAT]], [X] Parsing node: Relu_29 [Relu] [X] Searching for input: 853 [X] Relu_29 [Relu] inputs: [853 -> (1, 72, 36, 64)[FLOAT]], [X] Registering layer: Relu_29 for ONNX node: Relu_29 [X] Registering tensor: 426 for ONNX tensor: 426 [X] Relu_29 [Relu] outputs: [426 -> (1, 72, 36, 64)[FLOAT]], [X] Parsing node: GlobalAveragePool_30 [GlobalAveragePool] [X] Searching for input: 426 [X] GlobalAveragePool_30 [GlobalAveragePool] inputs: [426 -> (1, 72, 36, 64)[FLOAT]], [X] GlobalAveragePool operators are implemented via Reduce layers rather than Pooling layers [X] Registering layer: GlobalAveragePool_30 for ONNX node: GlobalAveragePool_30 [X] Registering tensor: 427 for ONNX tensor: 427 [X] GlobalAveragePool_30 [GlobalAveragePool] outputs: [427 -> (1, 72, 1, 1)[FLOAT]], [X] Parsing node: Conv_31 [Conv] [X] Searching for input: 427 [X] Searching for input: backbone.features.4.block.2.fc1.weight [X] Searching for input: backbone.features.4.block.2.fc1.bias [X] Conv_31 [Conv] inputs: [427 -> (1, 72, 1, 1)[FLOAT]], [backbone.features.4.block.2.fc1.weight -> (24, 72, 1, 1)[FLOAT]], [backbone.features.4.block.2.fc1.bias -> (24)[FLOAT]], [X] Convolution input dimensions: (1, 72, 1, 1) [X] Registering layer: Conv_31 for ONNX node: Conv_31 [X] Using kernel: (1, 1), strides: (1, 1), prepadding: (0, 0), postpadding: (0, 0), dilations: (1, 1), numOutputs: 24 [X] Convolution output dimensions: (1, 24, 1, 1) [X] Registering tensor: 428 for ONNX tensor: 428 [X] Conv_31 [Conv] outputs: [428 -> (1, 24, 1, 1)[FLOAT]], [X] Parsing node: Relu_32 [Relu] [X] Searching for input: 428 [X] Relu_32 [Relu] inputs: [428 -> (1, 24, 1, 1)[FLOAT]], [X] Registering layer: Relu_32 for ONNX node: Relu_32 [X] Registering tensor: 429 for ONNX tensor: 429 [X] Relu_32 [Relu] outputs: [429 -> (1, 24, 1, 1)[FLOAT]], [X] Parsing node: Conv_33 [Conv] [X] Searching for input: 429 [X] Searching for input: backbone.features.4.block.2.fc2.weight [X] Searching for input: backbone.features.4.block.2.fc2.bias [X] Conv_33 [Conv] inputs: [429 -> (1, 24, 1, 1)[FLOAT]], [backbone.features.4.block.2.fc2.weight -> (72, 24, 1, 1)[FLOAT]], [backbone.features.4.block.2.fc2.bias -> (72)[FLOAT]], [X] Convolution input dimensions: (1, 24, 1, 1) [X] Registering layer: Conv_33 for ONNX node: Conv_33 [X] Using kernel: (1, 1), strides: (1, 1), prepadding: (0, 0), postpadding: (0, 0), dilations: (1, 1), numOutputs: 72 [X] Convolution output dimensions: (1, 72, 1, 1) [X] Registering tensor: 430 for ONNX tensor: 430 [X] Conv_33 [Conv] outputs: [430 -> (1, 72, 1, 1)[FLOAT]], [X] Parsing node: HardSigmoid_34 [HardSigmoid] [X] Searching for input: 430 [X] HardSigmoid_34 [HardSigmoid] inputs: [430 -> (1, 72, 1, 1)[FLOAT]], [X] Registering layer: HardSigmoid_34 for ONNX node: HardSigmoid_34 [X] Registering tensor: 431 for ONNX tensor: 431 [X] HardSigmoid_34 [HardSigmoid] outputs: [431 -> (1, 72, 1, 1)[FLOAT]], [X] Parsing node: Mul_35 [Mul] [X] Searching for input: 431 [X] Searching for input: 426 [X] Mul_35 [Mul] inputs: [431 -> (1, 72, 1, 1)[FLOAT]], [426 -> (1, 72, 36, 64)[FLOAT]], [X] Registering layer: Mul_35 for ONNX node: Mul_35 [X] Registering tensor: 432 for ONNX tensor: 432 [X] Mul_35 [Mul] outputs: [432 -> (1, 72, 36, 64)[FLOAT]], [X] Parsing node: Conv_36 [Conv] [X] Searching for input: 432 [X] Searching for input: 857 [X] Searching for input: 858 [X] Conv_36 [Conv] inputs: [432 -> (1, 72, 36, 64)[FLOAT]], [857 -> (40, 72, 1, 1)[FLOAT]], [858 -> (40)[FLOAT]], [X] Convolution input dimensions: (1, 72, 36, 64) [X] Registering layer: Conv_36 for ONNX node: Conv_36 [X] Using kernel: (1, 1), strides: (1, 1), prepadding: (0, 0), postpadding: (0, 0), dilations: (1, 1), numOutputs: 40 [X] Convolution output dimensions: (1, 40, 36, 64) [X] Registering tensor: 856 for ONNX tensor: 856 [X] Conv_36 [Conv] outputs: [856 -> (1, 40, 36, 64)[FLOAT]], [X] Parsing node: Conv_37 [Conv] [X] Searching for input: 856 [X] Searching for input: 860 [X] Searching for input: 861 [X] Conv_37 [Conv] inputs: [856 -> (1, 40, 36, 64)[FLOAT]], [860 -> (120, 40, 1, 1)[FLOAT]], [861 -> (120)[FLOAT]], [X] Convolution input dimensions: (1, 40, 36, 64) [X] Registering layer: Conv_37 for ONNX node: Conv_37 [X] Using kernel: (1, 1), strides: (1, 1), prepadding: (0, 0), postpadding: (0, 0), dilations: (1, 1), numOutputs: 120 [X] Convolution output dimensions: (1, 120, 36, 64) [X] Registering tensor: 859 for ONNX tensor: 859 [X] Conv_37 [Conv] outputs: [859 -> (1, 120, 36, 64)[FLOAT]], [X] Parsing node: Relu_38 [Relu] [X] Searching for input: 859 [X] Relu_38 [Relu] inputs: [859 -> (1, 120, 36, 64)[FLOAT]], [X] Registering layer: Relu_38 for ONNX node: Relu_38 [X] Registering tensor: 437 for ONNX tensor: 437 [X] Relu_38 [Relu] outputs: [437 -> (1, 120, 36, 64)[FLOAT]], [X] Parsing node: Conv_39 [Conv] [X] Searching for input: 437 [X] Searching for input: 863 [X] Searching for input: 864 [X] Conv_39 [Conv] inputs: [437 -> (1, 120, 36, 64)[FLOAT]], [863 -> (120, 1, 5, 5)[FLOAT]], [864 -> (120)[FLOAT]], [X] Convolution input dimensions: (1, 120, 36, 64) [X] Registering layer: Conv_39 for ONNX node: Conv_39 [X] Using kernel: (5, 5), strides: (1, 1), prepadding: (2, 2), postpadding: (2, 2), dilations: (1, 1), numOutputs: 120 [X] Convolution output dimensions: (1, 120, 36, 64) [X] Registering tensor: 862 for ONNX tensor: 862 [X] Conv_39 [Conv] outputs: [862 -> (1, 120, 36, 64)[FLOAT]], [X] Parsing node: Relu_40 [Relu] [X] Searching for input: 862 [X] Relu_40 [Relu] inputs: [862 -> (1, 120, 36, 64)[FLOAT]], [X] Registering layer: Relu_40 for ONNX node: Relu_40 [X] Registering tensor: 440 for ONNX tensor: 440 [X] Relu_40 [Relu] outputs: [440 -> (1, 120, 36, 64)[FLOAT]], [X] Parsing node: GlobalAveragePool_41 [GlobalAveragePool] [X] Searching for input: 440 [X] GlobalAveragePool_41 [GlobalAveragePool] inputs: [440 -> (1, 120, 36, 64)[FLOAT]], [X] GlobalAveragePool operators are implemented via Reduce layers rather than Pooling layers [X] Registering layer: GlobalAveragePool_41 for ONNX node: GlobalAveragePool_41 [X] Registering tensor: 441 for ONNX tensor: 441 [X] GlobalAveragePool_41 [GlobalAveragePool] outputs: [441 -> (1, 120, 1, 1)[FLOAT]], [X] Parsing node: Conv_42 [Conv] [X] Searching for input: 441 [X] Searching for input: backbone.features.5.block.2.fc1.weight [X] Searching for input: backbone.features.5.block.2.fc1.bias [X] Conv_42 [Conv] inputs: [441 -> (1, 120, 1, 1)[FLOAT]], [backbone.features.5.block.2.fc1.weight -> (32, 120, 1, 1)[FLOAT]], [backbone.features.5.block.2.fc1.bias -> (32)[FLOAT]], [X] Convolution input dimensions: (1, 120, 1, 1) [X] Registering layer: Conv_42 for ONNX node: Conv_42 [X] Using kernel: (1, 1), strides: (1, 1), prepadding: (0, 0), postpadding: (0, 0), dilations: (1, 1), numOutputs: 32 [X] Convolution output dimensions: (1, 32, 1, 1) [X] Registering tensor: 442 for ONNX tensor: 442 [X] Conv_42 [Conv] outputs: [442 -> (1, 32, 1, 1)[FLOAT]], [X] Parsing node: Relu_43 [Relu] [X] Searching for input: 442 [X] Relu_43 [Relu] inputs: [442 -> (1, 32, 1, 1)[FLOAT]], [X] Registering layer: Relu_43 for ONNX node: Relu_43 [X] Registering tensor: 443 for ONNX tensor: 443 [X] Relu_43 [Relu] outputs: [443 -> (1, 32, 1, 1)[FLOAT]], [X] Parsing node: Conv_44 [Conv] [X] Searching for input: 443 [X] Searching for input: backbone.features.5.block.2.fc2.weight [X] Searching for input: backbone.features.5.block.2.fc2.bias [X] Conv_44 [Conv] inputs: [443 -> (1, 32, 1, 1)[FLOAT]], [backbone.features.5.block.2.fc2.weight -> (120, 32, 1, 1)[FLOAT]], [backbone.features.5.block.2.fc2.bias -> (120)[FLOAT]], [X] Convolution input dimensions: (1, 32, 1, 1) [X] Registering layer: Conv_44 for ONNX node: Conv_44 [X] Using kernel: (1, 1), strides: (1, 1), prepadding: (0, 0), postpadding: (0, 0), dilations: (1, 1), numOutputs: 120 [X] Convolution output dimensions: (1, 120, 1, 1) [X] Registering tensor: 444 for ONNX tensor: 444 [X] Conv_44 [Conv] outputs: [444 -> (1, 120, 1, 1)[FLOAT]], [X] Parsing node: HardSigmoid_45 [HardSigmoid] [X] Searching for input: 444 [X] HardSigmoid_45 [HardSigmoid] inputs: [444 -> (1, 120, 1, 1)[FLOAT]], [X] Registering layer: HardSigmoid_45 for ONNX node: HardSigmoid_45 [X] Registering tensor: 445 for ONNX tensor: 445 [X] HardSigmoid_45 [HardSigmoid] outputs: [445 -> (1, 120, 1, 1)[FLOAT]], [X] Parsing node: Mul_46 [Mul] [X] Searching for input: 445 [X] Searching for input: 440 [X] Mul_46 [Mul] inputs: [445 -> (1, 120, 1, 1)[FLOAT]], [440 -> (1, 120, 36, 64)[FLOAT]], [X] Registering layer: Mul_46 for ONNX node: Mul_46 [X] Registering tensor: 446 for ONNX tensor: 446 [X] Mul_46 [Mul] outputs: [446 -> (1, 120, 36, 64)[FLOAT]], [X] Parsing node: Conv_47 [Conv] [X] Searching for input: 446 [X] Searching for input: 866 [X] Searching for input: 867 [X] Conv_47 [Conv] inputs: [446 -> (1, 120, 36, 64)[FLOAT]], [866 -> (40, 120, 1, 1)[FLOAT]], [867 -> (40)[FLOAT]], [X] Convolution input dimensions: (1, 120, 36, 64) [X] Registering layer: Conv_47 for ONNX node: Conv_47 [X] Using kernel: (1, 1), strides: (1, 1), prepadding: (0, 0), postpadding: (0, 0), dilations: (1, 1), numOutputs: 40 [X] Convolution output dimensions: (1, 40, 36, 64) [X] Registering tensor: 865 for ONNX tensor: 865 [X] Conv_47 [Conv] outputs: [865 -> (1, 40, 36, 64)[FLOAT]], [X] Parsing node: Add_48 [Add] [X] Searching for input: 865 [X] Searching for input: 856 [X] Add_48 [Add] inputs: [865 -> (1, 40, 36, 64)[FLOAT]], [856 -> (1, 40, 36, 64)[FLOAT]], [X] Registering layer: Add_48 for ONNX node: Add_48 [X] Registering tensor: 449 for ONNX tensor: 449 [X] Add_48 [Add] outputs: [449 -> (1, 40, 36, 64)[FLOAT]], [X] Parsing node: Conv_49 [Conv] [X] Searching for input: 449 [X] Searching for input: 869 [X] Searching for input: 870 [X] Conv_49 [Conv] inputs: [449 -> (1, 40, 36, 64)[FLOAT]], [869 -> (120, 40, 1, 1)[FLOAT]], [870 -> (120)[FLOAT]], [X] Convolution input dimensions: (1, 40, 36, 64) [X] Registering layer: Conv_49 for ONNX node: Conv_49 [X] Using kernel: (1, 1), strides: (1, 1), prepadding: (0, 0), postpadding: (0, 0), dilations: (1, 1), numOutputs: 120 [X] Convolution output dimensions: (1, 120, 36, 64) [X] Registering tensor: 868 for ONNX tensor: 868 [X] Conv_49 [Conv] outputs: [868 -> (1, 120, 36, 64)[FLOAT]], [X] Parsing node: Relu_50 [Relu] [X] Searching for input: 868 [X] Relu_50 [Relu] inputs: [868 -> (1, 120, 36, 64)[FLOAT]], [X] Registering layer: Relu_50 for ONNX node: Relu_50 [X] Registering tensor: 452 for ONNX tensor: 452 [X] Relu_50 [Relu] outputs: [452 -> (1, 120, 36, 64)[FLOAT]], [X] Parsing node: Conv_51 [Conv] [X] Searching for input: 452 [X] Searching for input: 872 [X] Searching for input: 873 [X] Conv_51 [Conv] inputs: [452 -> (1, 120, 36, 64)[FLOAT]], [872 -> (120, 1, 5, 5)[FLOAT]], [873 -> (120)[FLOAT]], [X] Convolution input dimensions: (1, 120, 36, 64) [X] Registering layer: Conv_51 for ONNX node: Conv_51 [X] Using kernel: (5, 5), strides: (1, 1), prepadding: (2, 2), postpadding: (2, 2), dilations: (1, 1), numOutputs: 120 [X] Convolution output dimensions: (1, 120, 36, 64) [X] Registering tensor: 871 for ONNX tensor: 871 [X] Conv_51 [Conv] outputs: [871 -> (1, 120, 36, 64)[FLOAT]], [X] Parsing node: Relu_52 [Relu] [X] Searching for input: 871 [X] Relu_52 [Relu] inputs: [871 -> (1, 120, 36, 64)[FLOAT]], [X] Registering layer: Relu_52 for ONNX node: Relu_52 [X] Registering tensor: 455 for ONNX tensor: 455 [X] Relu_52 [Relu] outputs: [455 -> (1, 120, 36, 64)[FLOAT]], [X] Parsing node: GlobalAveragePool_53 [GlobalAveragePool] [X] Searching for input: 455 [X] GlobalAveragePool_53 [GlobalAveragePool] inputs: [455 -> (1, 120, 36, 64)[FLOAT]], [X] GlobalAveragePool operators are implemented via Reduce layers rather than Pooling layers [X] Registering layer: GlobalAveragePool_53 for ONNX node: GlobalAveragePool_53 [X] Registering tensor: 456 for ONNX tensor: 456 [X] GlobalAveragePool_53 [GlobalAveragePool] outputs: [456 -> (1, 120, 1, 1)[FLOAT]], [X] Parsing node: Conv_54 [Conv] [X] Searching for input: 456 [X] Searching for input: backbone.features.6.block.2.fc1.weight [X] Searching for input: backbone.features.6.block.2.fc1.bias [X] Conv_54 [Conv] inputs: [456 -> (1, 120, 1, 1)[FLOAT]], [backbone.features.6.block.2.fc1.weight -> (32, 120, 1, 1)[FLOAT]], [backbone.features.6.block.2.fc1.bias -> (32)[FLOAT]], [X] Convolution input dimensions: (1, 120, 1, 1) [X] Registering layer: Conv_54 for ONNX node: Conv_54 [X] Using kernel: (1, 1), strides: (1, 1), prepadding: (0, 0), postpadding: (0, 0), dilations: (1, 1), numOutputs: 32 [X] Convolution output dimensions: (1, 32, 1, 1) [X] Registering tensor: 457 for ONNX tensor: 457 [X] Conv_54 [Conv] outputs: [457 -> (1, 32, 1, 1)[FLOAT]], [X] Parsing node: Relu_55 [Relu] [X] Searching for input: 457 [X] Relu_55 [Relu] inputs: [457 -> (1, 32, 1, 1)[FLOAT]], [X] Registering layer: Relu_55 for ONNX node: Relu_55 [X] Registering tensor: 458 for ONNX tensor: 458 [X] Relu_55 [Relu] outputs: [458 -> (1, 32, 1, 1)[FLOAT]], [X] Parsing node: Conv_56 [Conv] [X] Searching for input: 458 [X] Searching for input: backbone.features.6.block.2.fc2.weight [X] Searching for input: backbone.features.6.block.2.fc2.bias [X] Conv_56 [Conv] inputs: [458 -> (1, 32, 1, 1)[FLOAT]], [backbone.features.6.block.2.fc2.weight -> (120, 32, 1, 1)[FLOAT]], [backbone.features.6.block.2.fc2.bias -> (120)[FLOAT]], [X] Convolution input dimensions: (1, 32, 1, 1) [X] Registering layer: Conv_56 for ONNX node: Conv_56 [X] Using kernel: (1, 1), strides: (1, 1), prepadding: (0, 0), postpadding: (0, 0), dilations: (1, 1), numOutputs: 120 [X] Convolution output dimensions: (1, 120, 1, 1) [X] Registering tensor: 459 for ONNX tensor: 459 [X] Conv_56 [Conv] outputs: [459 -> (1, 120, 1, 1)[FLOAT]], [X] Parsing node: HardSigmoid_57 [HardSigmoid] [X] Searching for input: 459 [X] HardSigmoid_57 [HardSigmoid] inputs: [459 -> (1, 120, 1, 1)[FLOAT]], [X] Registering layer: HardSigmoid_57 for ONNX node: HardSigmoid_57 [X] Registering tensor: 460 for ONNX tensor: 460 [X] HardSigmoid_57 [HardSigmoid] outputs: [460 -> (1, 120, 1, 1)[FLOAT]], [X] Parsing node: Mul_58 [Mul] [X] Searching for input: 460 [X] Searching for input: 455 [X] Mul_58 [Mul] inputs: [460 -> (1, 120, 1, 1)[FLOAT]], [455 -> (1, 120, 36, 64)[FLOAT]], [X] Registering layer: Mul_58 for ONNX node: Mul_58 [X] Registering tensor: 461 for ONNX tensor: 461 [X] Mul_58 [Mul] outputs: [461 -> (1, 120, 36, 64)[FLOAT]], [X] Parsing node: Conv_59 [Conv] [X] Searching for input: 461 [X] Searching for input: 875 [X] Searching for input: 876 [X] Conv_59 [Conv] inputs: [461 -> (1, 120, 36, 64)[FLOAT]], [875 -> (40, 120, 1, 1)[FLOAT]], [876 -> (40)[FLOAT]], [X] Convolution input dimensions: (1, 120, 36, 64) [X] Registering layer: Conv_59 for ONNX node: Conv_59 [X] Using kernel: (1, 1), strides: (1, 1), prepadding: (0, 0), postpadding: (0, 0), dilations: (1, 1), numOutputs: 40 [X] Convolution output dimensions: (1, 40, 36, 64) [X] Registering tensor: 874 for ONNX tensor: 874 [X] Conv_59 [Conv] outputs: [874 -> (1, 40, 36, 64)[FLOAT]], [X] Parsing node: Add_60 [Add] [X] Searching for input: 874 [X] Searching for input: 449 [X] Add_60 [Add] inputs: [874 -> (1, 40, 36, 64)[FLOAT]], [449 -> (1, 40, 36, 64)[FLOAT]], [X] Registering layer: Add_60 for ONNX node: Add_60 [X] Registering tensor: 464 for ONNX tensor: 464 [X] Add_60 [Add] outputs: [464 -> (1, 40, 36, 64)[FLOAT]], [X] Parsing node: Conv_61 [Conv] [X] Searching for input: 464 [X] Searching for input: 878 [X] Searching for input: 879 [X] Conv_61 [Conv] inputs: [464 -> (1, 40, 36, 64)[FLOAT]], [878 -> (240, 40, 1, 1)[FLOAT]], [879 -> (240)[FLOAT]], [X] Convolution input dimensions: (1, 40, 36, 64) [X] Registering layer: Conv_61 for ONNX node: Conv_61 [X] Using kernel: (1, 1), strides: (1, 1), prepadding: (0, 0), postpadding: (0, 0), dilations: (1, 1), numOutputs: 240 [X] Convolution output dimensions: (1, 240, 36, 64) [X] Registering tensor: 877 for ONNX tensor: 877 [X] Conv_61 [Conv] outputs: [877 -> (1, 240, 36, 64)[FLOAT]], [X] Parsing node: HardSigmoid_62 [HardSigmoid] [X] Searching for input: 877 [X] HardSigmoid_62 [HardSigmoid] inputs: [877 -> (1, 240, 36, 64)[FLOAT]], [X] Registering layer: HardSigmoid_62 for ONNX node: HardSigmoid_62 [X] Registering tensor: 467 for ONNX tensor: 467 [X] HardSigmoid_62 [HardSigmoid] outputs: [467 -> (1, 240, 36, 64)[FLOAT]], [X] Parsing node: Mul_63 [Mul] [X] Searching for input: 877 [X] Searching for input: 467 [X] Mul_63 [Mul] inputs: [877 -> (1, 240, 36, 64)[FLOAT]], [467 -> (1, 240, 36, 64)[FLOAT]], [X] Registering layer: Mul_63 for ONNX node: Mul_63 [X] Registering tensor: 468 for ONNX tensor: 468 [X] Mul_63 [Mul] outputs: [468 -> (1, 240, 36, 64)[FLOAT]], [X] Parsing node: Conv_64 [Conv] [X] Searching for input: 468 [X] Searching for input: 881 [X] Searching for input: 882 [X] Conv_64 [Conv] inputs: [468 -> (1, 240, 36, 64)[FLOAT]], [881 -> (240, 1, 3, 3)[FLOAT]], [882 -> (240)[FLOAT]], [X] Convolution input dimensions: (1, 240, 36, 64) [X] Registering layer: Conv_64 for ONNX node: Conv_64 [X] Using kernel: (3, 3), strides: (2, 2), prepadding: (1, 1), postpadding: (1, 1), dilations: (1, 1), numOutputs: 240 [X] Convolution output dimensions: (1, 240, 18, 32) [X] Registering tensor: 880 for ONNX tensor: 880 [X] Conv_64 [Conv] outputs: [880 -> (1, 240, 18, 32)[FLOAT]], [X] Parsing node: HardSigmoid_65 [HardSigmoid] [X] Searching for input: 880 [X] HardSigmoid_65 [HardSigmoid] inputs: [880 -> (1, 240, 18, 32)[FLOAT]], [X] Registering layer: HardSigmoid_65 for ONNX node: HardSigmoid_65 [X] Registering tensor: 471 for ONNX tensor: 471 [X] HardSigmoid_65 [HardSigmoid] outputs: [471 -> (1, 240, 18, 32)[FLOAT]], [X] Parsing node: Mul_66 [Mul] [X] Searching for input: 880 [X] Searching for input: 471 [X] Mul_66 [Mul] inputs: [880 -> (1, 240, 18, 32)[FLOAT]], [471 -> (1, 240, 18, 32)[FLOAT]], [X] Registering layer: Mul_66 for ONNX node: Mul_66 [X] Registering tensor: 472 for ONNX tensor: 472 [X] Mul_66 [Mul] outputs: [472 -> (1, 240, 18, 32)[FLOAT]], [X] Parsing node: Conv_67 [Conv] [X] Searching for input: 472 [X] Searching for input: 884 [X] Searching for input: 885 [X] Conv_67 [Conv] inputs: [472 -> (1, 240, 18, 32)[FLOAT]], [884 -> (80, 240, 1, 1)[FLOAT]], [885 -> (80)[FLOAT]], [X] Convolution input dimensions: (1, 240, 18, 32) [X] Registering layer: Conv_67 for ONNX node: Conv_67 [X] Using kernel: (1, 1), strides: (1, 1), prepadding: (0, 0), postpadding: (0, 0), dilations: (1, 1), numOutputs: 80 [X] Convolution output dimensions: (1, 80, 18, 32) [X] Registering tensor: 883 for ONNX tensor: 883 [X] Conv_67 [Conv] outputs: [883 -> (1, 80, 18, 32)[FLOAT]], [X] Parsing node: Conv_68 [Conv] [X] Searching for input: 883 [X] Searching for input: 887 [X] Searching for input: 888 [X] Conv_68 [Conv] inputs: [883 -> (1, 80, 18, 32)[FLOAT]], [887 -> (200, 80, 1, 1)[FLOAT]], [888 -> (200)[FLOAT]], [X] Convolution input dimensions: (1, 80, 18, 32) [X] Registering layer: Conv_68 for ONNX node: Conv_68 [X] Using kernel: (1, 1), strides: (1, 1), prepadding: (0, 0), postpadding: (0, 0), dilations: (1, 1), numOutputs: 200 [X] Convolution output dimensions: (1, 200, 18, 32) [X] Registering tensor: 886 for ONNX tensor: 886 [X] Conv_68 [Conv] outputs: [886 -> (1, 200, 18, 32)[FLOAT]], [X] Parsing node: HardSigmoid_69 [HardSigmoid] [X] Searching for input: 886 [X] HardSigmoid_69 [HardSigmoid] inputs: [886 -> (1, 200, 18, 32)[FLOAT]], [X] Registering layer: HardSigmoid_69 for ONNX node: HardSigmoid_69 [X] Registering tensor: 477 for ONNX tensor: 477 [X] HardSigmoid_69 [HardSigmoid] outputs: [477 -> (1, 200, 18, 32)[FLOAT]], [X] Parsing node: Mul_70 [Mul] [X] Searching for input: 886 [X] Searching for input: 477 [X] Mul_70 [Mul] inputs: [886 -> (1, 200, 18, 32)[FLOAT]], [477 -> (1, 200, 18, 32)[FLOAT]], [X] Registering layer: Mul_70 for ONNX node: Mul_70 [X] Registering tensor: 478 for ONNX tensor: 478 [X] Mul_70 [Mul] outputs: [478 -> (1, 200, 18, 32)[FLOAT]], [X] Parsing node: Conv_71 [Conv] [X] Searching for input: 478 [X] Searching for input: 890 [X] Searching for input: 891 [X] Conv_71 [Conv] inputs: [478 -> (1, 200, 18, 32)[FLOAT]], [890 -> (200, 1, 3, 3)[FLOAT]], [891 -> (200)[FLOAT]], [X] Convolution input dimensions: (1, 200, 18, 32) [X] Registering layer: Conv_71 for ONNX node: Conv_71 [X] Using kernel: (3, 3), strides: (1, 1), prepadding: (1, 1), postpadding: (1, 1), dilations: (1, 1), numOutputs: 200 [X] Convolution output dimensions: (1, 200, 18, 32) [X] Registering tensor: 889 for ONNX tensor: 889 [X] Conv_71 [Conv] outputs: [889 -> (1, 200, 18, 32)[FLOAT]], [X] Parsing node: HardSigmoid_72 [HardSigmoid] [X] Searching for input: 889 [X] HardSigmoid_72 [HardSigmoid] inputs: [889 -> (1, 200, 18, 32)[FLOAT]], [X] Registering layer: HardSigmoid_72 for ONNX node: HardSigmoid_72 [X] Registering tensor: 481 for ONNX tensor: 481 [X] HardSigmoid_72 [HardSigmoid] outputs: [481 -> (1, 200, 18, 32)[FLOAT]], [X] Parsing node: Mul_73 [Mul] [X] Searching for input: 889 [X] Searching for input: 481 [X] Mul_73 [Mul] inputs: [889 -> (1, 200, 18, 32)[FLOAT]], [481 -> (1, 200, 18, 32)[FLOAT]], [X] Registering layer: Mul_73 for ONNX node: Mul_73 [X] Registering tensor: 482 for ONNX tensor: 482 [X] Mul_73 [Mul] outputs: [482 -> (1, 200, 18, 32)[FLOAT]], [X] Parsing node: Conv_74 [Conv] [X] Searching for input: 482 [X] Searching for input: 893 [X] Searching for input: 894 [X] Conv_74 [Conv] inputs: [482 -> (1, 200, 18, 32)[FLOAT]], [893 -> (80, 200, 1, 1)[FLOAT]], [894 -> (80)[FLOAT]], [X] Convolution input dimensions: (1, 200, 18, 32) [X] Registering layer: Conv_74 for ONNX node: Conv_74 [X] Using kernel: (1, 1), strides: (1, 1), prepadding: (0, 0), postpadding: (0, 0), dilations: (1, 1), numOutputs: 80 [X] Convolution output dimensions: (1, 80, 18, 32) [X] Registering tensor: 892 for ONNX tensor: 892 [X] Conv_74 [Conv] outputs: [892 -> (1, 80, 18, 32)[FLOAT]], [X] Parsing node: Add_75 [Add] [X] Searching for input: 892 [X] Searching for input: 883 [X] Add_75 [Add] inputs: [892 -> (1, 80, 18, 32)[FLOAT]], [883 -> (1, 80, 18, 32)[FLOAT]], [X] Registering layer: Add_75 for ONNX node: Add_75 [X] Registering tensor: 485 for ONNX tensor: 485 [X] Add_75 [Add] outputs: [485 -> (1, 80, 18, 32)[FLOAT]], [X] Parsing node: Conv_76 [Conv] [X] Searching for input: 485 [X] Searching for input: 896 [X] Searching for input: 897 [X] Conv_76 [Conv] inputs: [485 -> (1, 80, 18, 32)[FLOAT]], [896 -> (184, 80, 1, 1)[FLOAT]], [897 -> (184)[FLOAT]], [X] Convolution input dimensions: (1, 80, 18, 32) [X] Registering layer: Conv_76 for ONNX node: Conv_76 [X] Using kernel: (1, 1), strides: (1, 1), prepadding: (0, 0), postpadding: (0, 0), dilations: (1, 1), numOutputs: 184 [X] Convolution output dimensions: (1, 184, 18, 32) [X] Registering tensor: 895 for ONNX tensor: 895 [X] Conv_76 [Conv] outputs: [895 -> (1, 184, 18, 32)[FLOAT]], [X] Parsing node: HardSigmoid_77 [HardSigmoid] [X] Searching for input: 895 [X] HardSigmoid_77 [HardSigmoid] inputs: [895 -> (1, 184, 18, 32)[FLOAT]], [X] Registering layer: HardSigmoid_77 for ONNX node: HardSigmoid_77 [X] Registering tensor: 488 for ONNX tensor: 488 [X] HardSigmoid_77 [HardSigmoid] outputs: [488 -> (1, 184, 18, 32)[FLOAT]], [X] Parsing node: Mul_78 [Mul] [X] Searching for input: 895 [X] Searching for input: 488 [X] Mul_78 [Mul] inputs: [895 -> (1, 184, 18, 32)[FLOAT]], [488 -> (1, 184, 18, 32)[FLOAT]], [X] Registering layer: Mul_78 for ONNX node: Mul_78 [X] Registering tensor: 489 for ONNX tensor: 489 [X] Mul_78 [Mul] outputs: [489 -> (1, 184, 18, 32)[FLOAT]], [X] Parsing node: Conv_79 [Conv] [X] Searching for input: 489 [X] Searching for input: 899 [X] Searching for input: 900 [X] Conv_79 [Conv] inputs: [489 -> (1, 184, 18, 32)[FLOAT]], [899 -> (184, 1, 3, 3)[FLOAT]], [900 -> (184)[FLOAT]], [X] Convolution input dimensions: (1, 184, 18, 32) [X] Registering layer: Conv_79 for ONNX node: Conv_79 [X] Using kernel: (3, 3), strides: (1, 1), prepadding: (1, 1), postpadding: (1, 1), dilations: (1, 1), numOutputs: 184 [X] Convolution output dimensions: (1, 184, 18, 32) [X] Registering tensor: 898 for ONNX tensor: 898 [X] Conv_79 [Conv] outputs: [898 -> (1, 184, 18, 32)[FLOAT]], [X] Parsing node: HardSigmoid_80 [HardSigmoid] [X] Searching for input: 898 [X] HardSigmoid_80 [HardSigmoid] inputs: [898 -> (1, 184, 18, 32)[FLOAT]], [X] Registering layer: HardSigmoid_80 for ONNX node: HardSigmoid_80 [X] Registering tensor: 492 for ONNX tensor: 492 [X] HardSigmoid_80 [HardSigmoid] outputs: [492 -> (1, 184, 18, 32)[FLOAT]], [X] Parsing node: Mul_81 [Mul] [X] Searching for input: 898 [X] Searching for input: 492 [X] Mul_81 [Mul] inputs: [898 -> (1, 184, 18, 32)[FLOAT]], [492 -> (1, 184, 18, 32)[FLOAT]], [X] Registering layer: Mul_81 for ONNX node: Mul_81 [X] Registering tensor: 493 for ONNX tensor: 493 [X] Mul_81 [Mul] outputs: [493 -> (1, 184, 18, 32)[FLOAT]], [X] Parsing node: Conv_82 [Conv] [X] Searching for input: 493 [X] Searching for input: 902 [X] Searching for input: 903 [X] Conv_82 [Conv] inputs: [493 -> (1, 184, 18, 32)[FLOAT]], [902 -> (80, 184, 1, 1)[FLOAT]], [903 -> (80)[FLOAT]], [X] Convolution input dimensions: (1, 184, 18, 32) [X] Registering layer: Conv_82 for ONNX node: Conv_82 [X] Using kernel: (1, 1), strides: (1, 1), prepadding: (0, 0), postpadding: (0, 0), dilations: (1, 1), numOutputs: 80 [X] Convolution output dimensions: (1, 80, 18, 32) [X] Registering tensor: 901 for ONNX tensor: 901 [X] Conv_82 [Conv] outputs: [901 -> (1, 80, 18, 32)[FLOAT]], [X] Parsing node: Add_83 [Add] [X] Searching for input: 901 [X] Searching for input: 485 [X] Add_83 [Add] inputs: [901 -> (1, 80, 18, 32)[FLOAT]], [485 -> (1, 80, 18, 32)[FLOAT]], [X] Registering layer: Add_83 for ONNX node: Add_83 [X] Registering tensor: 496 for ONNX tensor: 496 [X] Add_83 [Add] outputs: [496 -> (1, 80, 18, 32)[FLOAT]], [X] Parsing node: Conv_84 [Conv] [X] Searching for input: 496 [X] Searching for input: 905 [X] Searching for input: 906 [X] Conv_84 [Conv] inputs: [496 -> (1, 80, 18, 32)[FLOAT]], [905 -> (184, 80, 1, 1)[FLOAT]], [906 -> (184)[FLOAT]], [X] Convolution input dimensions: (1, 80, 18, 32) [X] Registering layer: Conv_84 for ONNX node: Conv_84 [X] Using kernel: (1, 1), strides: (1, 1), prepadding: (0, 0), postpadding: (0, 0), dilations: (1, 1), numOutputs: 184 [X] Convolution output dimensions: (1, 184, 18, 32) [X] Registering tensor: 904 for ONNX tensor: 904 [X] Conv_84 [Conv] outputs: [904 -> (1, 184, 18, 32)[FLOAT]], [X] Parsing node: HardSigmoid_85 [HardSigmoid] [X] Searching for input: 904 [X] HardSigmoid_85 [HardSigmoid] inputs: [904 -> (1, 184, 18, 32)[FLOAT]], [X] Registering layer: HardSigmoid_85 for ONNX node: HardSigmoid_85 [X] Registering tensor: 499 for ONNX tensor: 499 [X] HardSigmoid_85 [HardSigmoid] outputs: [499 -> (1, 184, 18, 32)[FLOAT]], [X] Parsing node: Mul_86 [Mul] [X] Searching for input: 904 [X] Searching for input: 499 [X] Mul_86 [Mul] inputs: [904 -> (1, 184, 18, 32)[FLOAT]], [499 -> (1, 184, 18, 32)[FLOAT]], [X] Registering layer: Mul_86 for ONNX node: Mul_86 [X] Registering tensor: 500 for ONNX tensor: 500 [X] Mul_86 [Mul] outputs: [500 -> (1, 184, 18, 32)[FLOAT]], [X] Parsing node: Conv_87 [Conv] [X] Searching for input: 500 [X] Searching for input: 908 [X] Searching for input: 909 [X] Conv_87 [Conv] inputs: [500 -> (1, 184, 18, 32)[FLOAT]], [908 -> (184, 1, 3, 3)[FLOAT]], [909 -> (184)[FLOAT]], [X] Convolution input dimensions: (1, 184, 18, 32) [X] Registering layer: Conv_87 for ONNX node: Conv_87 [X] Using kernel: (3, 3), strides: (1, 1), prepadding: (1, 1), postpadding: (1, 1), dilations: (1, 1), numOutputs: 184 [X] Convolution output dimensions: (1, 184, 18, 32) [X] Registering tensor: 907 for ONNX tensor: 907 [X] Conv_87 [Conv] outputs: [907 -> (1, 184, 18, 32)[FLOAT]], [X] Parsing node: HardSigmoid_88 [HardSigmoid] [X] Searching for input: 907 [X] HardSigmoid_88 [HardSigmoid] inputs: [907 -> (1, 184, 18, 32)[FLOAT]], [X] Registering layer: HardSigmoid_88 for ONNX node: HardSigmoid_88 [X] Registering tensor: 503 for ONNX tensor: 503 [X] HardSigmoid_88 [HardSigmoid] outputs: [503 -> (1, 184, 18, 32)[FLOAT]], [X] Parsing node: Mul_89 [Mul] [X] Searching for input: 907 [X] Searching for input: 503 [X] Mul_89 [Mul] inputs: [907 -> (1, 184, 18, 32)[FLOAT]], [503 -> (1, 184, 18, 32)[FLOAT]], [X] Registering layer: Mul_89 for ONNX node: Mul_89 [X] Registering tensor: 504 for ONNX tensor: 504 [X] Mul_89 [Mul] outputs: [504 -> (1, 184, 18, 32)[FLOAT]], [X] Parsing node: Conv_90 [Conv] [X] Searching for input: 504 [X] Searching for input: 911 [X] Searching for input: 912 [X] Conv_90 [Conv] inputs: [504 -> (1, 184, 18, 32)[FLOAT]], [911 -> (80, 184, 1, 1)[FLOAT]], [912 -> (80)[FLOAT]], [X] Convolution input dimensions: (1, 184, 18, 32) [X] Registering layer: Conv_90 for ONNX node: Conv_90 [X] Using kernel: (1, 1), strides: (1, 1), prepadding: (0, 0), postpadding: (0, 0), dilations: (1, 1), numOutputs: 80 [X] Convolution output dimensions: (1, 80, 18, 32) [X] Registering tensor: 910 for ONNX tensor: 910 [X] Conv_90 [Conv] outputs: [910 -> (1, 80, 18, 32)[FLOAT]], [X] Parsing node: Add_91 [Add] [X] Searching for input: 910 [X] Searching for input: 496 [X] Add_91 [Add] inputs: [910 -> (1, 80, 18, 32)[FLOAT]], [496 -> (1, 80, 18, 32)[FLOAT]], [X] Registering layer: Add_91 for ONNX node: Add_91 [X] Registering tensor: 507 for ONNX tensor: 507 [X] Add_91 [Add] outputs: [507 -> (1, 80, 18, 32)[FLOAT]], [X] Parsing node: Conv_92 [Conv] [X] Searching for input: 507 [X] Searching for input: 914 [X] Searching for input: 915 [X] Conv_92 [Conv] inputs: [507 -> (1, 80, 18, 32)[FLOAT]], [914 -> (480, 80, 1, 1)[FLOAT]], [915 -> (480)[FLOAT]], [X] Convolution input dimensions: (1, 80, 18, 32) [X] Registering layer: Conv_92 for ONNX node: Conv_92 [X] Using kernel: (1, 1), strides: (1, 1), prepadding: (0, 0), postpadding: (0, 0), dilations: (1, 1), numOutputs: 480 [X] Convolution output dimensions: (1, 480, 18, 32) [X] Registering tensor: 913 for ONNX tensor: 913 [X] Conv_92 [Conv] outputs: [913 -> (1, 480, 18, 32)[FLOAT]], [X] Parsing node: HardSigmoid_93 [HardSigmoid] [X] Searching for input: 913 [X] HardSigmoid_93 [HardSigmoid] inputs: [913 -> (1, 480, 18, 32)[FLOAT]], [X] Registering layer: HardSigmoid_93 for ONNX node: HardSigmoid_93 [X] Registering tensor: 510 for ONNX tensor: 510 [X] HardSigmoid_93 [HardSigmoid] outputs: [510 -> (1, 480, 18, 32)[FLOAT]], [X] Parsing node: Mul_94 [Mul] [X] Searching for input: 913 [X] Searching for input: 510 [X] Mul_94 [Mul] inputs: [913 -> (1, 480, 18, 32)[FLOAT]], [510 -> (1, 480, 18, 32)[FLOAT]], [X] Registering layer: Mul_94 for ONNX node: Mul_94 [X] Registering tensor: 511 for ONNX tensor: 511 [X] Mul_94 [Mul] outputs: [511 -> (1, 480, 18, 32)[FLOAT]], [X] Parsing node: Conv_95 [Conv] [X] Searching for input: 511 [X] Searching for input: 917 [X] Searching for input: 918 [X] Conv_95 [Conv] inputs: [511 -> (1, 480, 18, 32)[FLOAT]], [917 -> (480, 1, 3, 3)[FLOAT]], [918 -> (480)[FLOAT]], [X] Convolution input dimensions: (1, 480, 18, 32) [X] Registering layer: Conv_95 for ONNX node: Conv_95 [X] Using kernel: (3, 3), strides: (1, 1), prepadding: (1, 1), postpadding: (1, 1), dilations: (1, 1), numOutputs: 480 [X] Convolution output dimensions: (1, 480, 18, 32) [X] Registering tensor: 916 for ONNX tensor: 916 [X] Conv_95 [Conv] outputs: [916 -> (1, 480, 18, 32)[FLOAT]], [X] Parsing node: HardSigmoid_96 [HardSigmoid] [X] Searching for input: 916 [X] HardSigmoid_96 [HardSigmoid] inputs: [916 -> (1, 480, 18, 32)[FLOAT]], [X] Registering layer: HardSigmoid_96 for ONNX node: HardSigmoid_96 [X] Registering tensor: 514 for ONNX tensor: 514 [X] HardSigmoid_96 [HardSigmoid] outputs: [514 -> (1, 480, 18, 32)[FLOAT]], [X] Parsing node: Mul_97 [Mul] [X] Searching for input: 916 [X] Searching for input: 514 [X] Mul_97 [Mul] inputs: [916 -> (1, 480, 18, 32)[FLOAT]], [514 -> (1, 480, 18, 32)[FLOAT]], [X] Registering layer: Mul_97 for ONNX node: Mul_97 [X] Registering tensor: 515 for ONNX tensor: 515 [X] Mul_97 [Mul] outputs: [515 -> (1, 480, 18, 32)[FLOAT]], [X] Parsing node: GlobalAveragePool_98 [GlobalAveragePool] [X] Searching for input: 515 [X] GlobalAveragePool_98 [GlobalAveragePool] inputs: [515 -> (1, 480, 18, 32)[FLOAT]], [X] GlobalAveragePool operators are implemented via Reduce layers rather than Pooling layers [X] Registering layer: GlobalAveragePool_98 for ONNX node: GlobalAveragePool_98 [X] Registering tensor: 516 for ONNX tensor: 516 [X] GlobalAveragePool_98 [GlobalAveragePool] outputs: [516 -> (1, 480, 1, 1)[FLOAT]], [X] Parsing node: Conv_99 [Conv] [X] Searching for input: 516 [X] Searching for input: backbone.features.11.block.2.fc1.weight [X] Searching for input: backbone.features.11.block.2.fc1.bias [X] Conv_99 [Conv] inputs: [516 -> (1, 480, 1, 1)[FLOAT]], [backbone.features.11.block.2.fc1.weight -> (120, 480, 1, 1)[FLOAT]], [backbone.features.11.block.2.fc1.bias -> (120)[FLOAT]], [X] Convolution input dimensions: (1, 480, 1, 1) [X] Registering layer: Conv_99 for ONNX node: Conv_99 [X] Using kernel: (1, 1), strides: (1, 1), prepadding: (0, 0), postpadding: (0, 0), dilations: (1, 1), numOutputs: 120 [X] Convolution output dimensions: (1, 120, 1, 1) [X] Registering tensor: 517 for ONNX tensor: 517 [X] Conv_99 [Conv] outputs: [517 -> (1, 120, 1, 1)[FLOAT]], [X] Parsing node: Relu_100 [Relu] [X] Searching for input: 517 [X] Relu_100 [Relu] inputs: [517 -> (1, 120, 1, 1)[FLOAT]], [X] Registering layer: Relu_100 for ONNX node: Relu_100 [X] Registering tensor: 518 for ONNX tensor: 518 [X] Relu_100 [Relu] outputs: [518 -> (1, 120, 1, 1)[FLOAT]], [X] Parsing node: Conv_101 [Conv] [X] Searching for input: 518 [X] Searching for input: backbone.features.11.block.2.fc2.weight [X] Searching for input: backbone.features.11.block.2.fc2.bias [X] Conv_101 [Conv] inputs: [518 -> (1, 120, 1, 1)[FLOAT]], [backbone.features.11.block.2.fc2.weight -> (480, 120, 1, 1)[FLOAT]], [backbone.features.11.block.2.fc2.bias -> (480)[FLOAT]], [X] Convolution input dimensions: (1, 120, 1, 1) [X] Registering layer: Conv_101 for ONNX node: Conv_101 [X] Using kernel: (1, 1), strides: (1, 1), prepadding: (0, 0), postpadding: (0, 0), dilations: (1, 1), numOutputs: 480 [X] Convolution output dimensions: (1, 480, 1, 1) [X] Registering tensor: 519 for ONNX tensor: 519 [X] Conv_101 [Conv] outputs: [519 -> (1, 480, 1, 1)[FLOAT]], [X] Parsing node: HardSigmoid_102 [HardSigmoid] [X] Searching for input: 519 [X] HardSigmoid_102 [HardSigmoid] inputs: [519 -> (1, 480, 1, 1)[FLOAT]], [X] Registering layer: HardSigmoid_102 for ONNX node: HardSigmoid_102 [X] Registering tensor: 520 for ONNX tensor: 520 [X] HardSigmoid_102 [HardSigmoid] outputs: [520 -> (1, 480, 1, 1)[FLOAT]], [X] Parsing node: Mul_103 [Mul] [X] Searching for input: 520 [X] Searching for input: 515 [X] Mul_103 [Mul] inputs: [520 -> (1, 480, 1, 1)[FLOAT]], [515 -> (1, 480, 18, 32)[FLOAT]], [X] Registering layer: Mul_103 for ONNX node: Mul_103 [X] Registering tensor: 521 for ONNX tensor: 521 [X] Mul_103 [Mul] outputs: [521 -> (1, 480, 18, 32)[FLOAT]], [X] Parsing node: Conv_104 [Conv] [X] Searching for input: 521 [X] Searching for input: 920 [X] Searching for input: 921 [X] Conv_104 [Conv] inputs: [521 -> (1, 480, 18, 32)[FLOAT]], [920 -> (112, 480, 1, 1)[FLOAT]], [921 -> (112)[FLOAT]], [X] Convolution input dimensions: (1, 480, 18, 32) [X] Registering layer: Conv_104 for ONNX node: Conv_104 [X] Using kernel: (1, 1), strides: (1, 1), prepadding: (0, 0), postpadding: (0, 0), dilations: (1, 1), numOutputs: 112 [X] Convolution output dimensions: (1, 112, 18, 32) [X] Registering tensor: 919 for ONNX tensor: 919 [X] Conv_104 [Conv] outputs: [919 -> (1, 112, 18, 32)[FLOAT]], [X] Parsing node: Conv_105 [Conv] [X] Searching for input: 919 [X] Searching for input: 923 [X] Searching for input: 924 [X] Conv_105 [Conv] inputs: [919 -> (1, 112, 18, 32)[FLOAT]], [923 -> (672, 112, 1, 1)[FLOAT]], [924 -> (672)[FLOAT]], [X] Convolution input dimensions: (1, 112, 18, 32) [X] Registering layer: Conv_105 for ONNX node: Conv_105 [X] Using kernel: (1, 1), strides: (1, 1), prepadding: (0, 0), postpadding: (0, 0), dilations: (1, 1), numOutputs: 672 [X] Convolution output dimensions: (1, 672, 18, 32) [X] Registering tensor: 922 for ONNX tensor: 922 [X] Conv_105 [Conv] outputs: [922 -> (1, 672, 18, 32)[FLOAT]], [X] Parsing node: HardSigmoid_106 [HardSigmoid] [X] Searching for input: 922 [X] HardSigmoid_106 [HardSigmoid] inputs: [922 -> (1, 672, 18, 32)[FLOAT]], [X] Registering layer: HardSigmoid_106 for ONNX node: HardSigmoid_106 [X] Registering tensor: 526 for ONNX tensor: 526 [X] HardSigmoid_106 [HardSigmoid] outputs: [526 -> (1, 672, 18, 32)[FLOAT]], [X] Parsing node: Mul_107 [Mul] [X] Searching for input: 922 [X] Searching for input: 526 [X] Mul_107 [Mul] inputs: [922 -> (1, 672, 18, 32)[FLOAT]], [526 -> (1, 672, 18, 32)[FLOAT]], [X] Registering layer: Mul_107 for ONNX node: Mul_107 [X] Registering tensor: 527 for ONNX tensor: 527 [X] Mul_107 [Mul] outputs: [527 -> (1, 672, 18, 32)[FLOAT]], [X] Parsing node: Conv_108 [Conv] [X] Searching for input: 527 [X] Searching for input: 926 [X] Searching for input: 927 [X] Conv_108 [Conv] inputs: [527 -> (1, 672, 18, 32)[FLOAT]], [926 -> (672, 1, 3, 3)[FLOAT]], [927 -> (672)[FLOAT]], [X] Convolution input dimensions: (1, 672, 18, 32) [X] Registering layer: Conv_108 for ONNX node: Conv_108 [X] Using kernel: (3, 3), strides: (1, 1), prepadding: (1, 1), postpadding: (1, 1), dilations: (1, 1), numOutputs: 672 [X] Convolution output dimensions: (1, 672, 18, 32) [X] Registering tensor: 925 for ONNX tensor: 925 [X] Conv_108 [Conv] outputs: [925 -> (1, 672, 18, 32)[FLOAT]], [X] Parsing node: HardSigmoid_109 [HardSigmoid] [X] Searching for input: 925 [X] HardSigmoid_109 [HardSigmoid] inputs: [925 -> (1, 672, 18, 32)[FLOAT]], [X] Registering layer: HardSigmoid_109 for ONNX node: HardSigmoid_109 [X] Registering tensor: 530 for ONNX tensor: 530 [X] HardSigmoid_109 [HardSigmoid] outputs: [530 -> (1, 672, 18, 32)[FLOAT]], [X] Parsing node: Mul_110 [Mul] [X] Searching for input: 925 [X] Searching for input: 530 [X] Mul_110 [Mul] inputs: [925 -> (1, 672, 18, 32)[FLOAT]], [530 -> (1, 672, 18, 32)[FLOAT]], [X] Registering layer: Mul_110 for ONNX node: Mul_110 [X] Registering tensor: 531 for ONNX tensor: 531 [X] Mul_110 [Mul] outputs: [531 -> (1, 672, 18, 32)[FLOAT]], [X] Parsing node: GlobalAveragePool_111 [GlobalAveragePool] [X] Searching for input: 531 [X] GlobalAveragePool_111 [GlobalAveragePool] inputs: [531 -> (1, 672, 18, 32)[FLOAT]], [X] GlobalAveragePool operators are implemented via Reduce layers rather than Pooling layers [X] Registering layer: GlobalAveragePool_111 for ONNX node: GlobalAveragePool_111 [X] Registering tensor: 532 for ONNX tensor: 532 [X] GlobalAveragePool_111 [GlobalAveragePool] outputs: [532 -> (1, 672, 1, 1)[FLOAT]], [X] Parsing node: Conv_112 [Conv] [X] Searching for input: 532 [X] Searching for input: backbone.features.12.block.2.fc1.weight [X] Searching for input: backbone.features.12.block.2.fc1.bias [X] Conv_112 [Conv] inputs: [532 -> (1, 672, 1, 1)[FLOAT]], [backbone.features.12.block.2.fc1.weight -> (168, 672, 1, 1)[FLOAT]], [backbone.features.12.block.2.fc1.bias -> (168)[FLOAT]], [X] Convolution input dimensions: (1, 672, 1, 1) [X] Registering layer: Conv_112 for ONNX node: Conv_112 [X] Using kernel: (1, 1), strides: (1, 1), prepadding: (0, 0), postpadding: (0, 0), dilations: (1, 1), numOutputs: 168 [X] Convolution output dimensions: (1, 168, 1, 1) [X] Registering tensor: 533 for ONNX tensor: 533 [X] Conv_112 [Conv] outputs: [533 -> (1, 168, 1, 1)[FLOAT]], [X] Parsing node: Relu_113 [Relu] [X] Searching for input: 533 [X] Relu_113 [Relu] inputs: [533 -> (1, 168, 1, 1)[FLOAT]], [X] Registering layer: Relu_113 for ONNX node: Relu_113 [X] Registering tensor: 534 for ONNX tensor: 534 [X] Relu_113 [Relu] outputs: [534 -> (1, 168, 1, 1)[FLOAT]], [X] Parsing node: Conv_114 [Conv] [X] Searching for input: 534 [X] Searching for input: backbone.features.12.block.2.fc2.weight [X] Searching for input: backbone.features.12.block.2.fc2.bias [X] Conv_114 [Conv] inputs: [534 -> (1, 168, 1, 1)[FLOAT]], [backbone.features.12.block.2.fc2.weight -> (672, 168, 1, 1)[FLOAT]], [backbone.features.12.block.2.fc2.bias -> (672)[FLOAT]], [X] Convolution input dimensions: (1, 168, 1, 1) [X] Registering layer: Conv_114 for ONNX node: Conv_114 [X] Using kernel: (1, 1), strides: (1, 1), prepadding: (0, 0), postpadding: (0, 0), dilations: (1, 1), numOutputs: 672 [X] Convolution output dimensions: (1, 672, 1, 1) [X] Registering tensor: 535 for ONNX tensor: 535 [X] Conv_114 [Conv] outputs: [535 -> (1, 672, 1, 1)[FLOAT]], [X] Parsing node: HardSigmoid_115 [HardSigmoid] [X] Searching for input: 535 [X] HardSigmoid_115 [HardSigmoid] inputs: [535 -> (1, 672, 1, 1)[FLOAT]], [X] Registering layer: HardSigmoid_115 for ONNX node: HardSigmoid_115 [X] Registering tensor: 536 for ONNX tensor: 536 [X] HardSigmoid_115 [HardSigmoid] outputs: [536 -> (1, 672, 1, 1)[FLOAT]], [X] Parsing node: Mul_116 [Mul] [X] Searching for input: 536 [X] Searching for input: 531 [X] Mul_116 [Mul] inputs: [536 -> (1, 672, 1, 1)[FLOAT]], [531 -> (1, 672, 18, 32)[FLOAT]], [X] Registering layer: Mul_116 for ONNX node: Mul_116 [X] Registering tensor: 537 for ONNX tensor: 537 [X] Mul_116 [Mul] outputs: [537 -> (1, 672, 18, 32)[FLOAT]], [X] Parsing node: Conv_117 [Conv] [X] Searching for input: 537 [X] Searching for input: 929 [X] Searching for input: 930 [X] Conv_117 [Conv] inputs: [537 -> (1, 672, 18, 32)[FLOAT]], [929 -> (112, 672, 1, 1)[FLOAT]], [930 -> (112)[FLOAT]], [X] Convolution input dimensions: (1, 672, 18, 32) [X] Registering layer: Conv_117 for ONNX node: Conv_117 [X] Using kernel: (1, 1), strides: (1, 1), prepadding: (0, 0), postpadding: (0, 0), dilations: (1, 1), numOutputs: 112 [X] Convolution output dimensions: (1, 112, 18, 32) [X] Registering tensor: 928 for ONNX tensor: 928 [X] Conv_117 [Conv] outputs: [928 -> (1, 112, 18, 32)[FLOAT]], [X] Parsing node: Add_118 [Add] [X] Searching for input: 928 [X] Searching for input: 919 [X] Add_118 [Add] inputs: [928 -> (1, 112, 18, 32)[FLOAT]], [919 -> (1, 112, 18, 32)[FLOAT]], [X] Registering layer: Add_118 for ONNX node: Add_118 [X] Registering tensor: 540 for ONNX tensor: 540 [X] Add_118 [Add] outputs: [540 -> (1, 112, 18, 32)[FLOAT]], [X] Parsing node: Conv_119 [Conv] [X] Searching for input: 540 [X] Searching for input: 932 [X] Searching for input: 933 [X] Conv_119 [Conv] inputs: [540 -> (1, 112, 18, 32)[FLOAT]], [932 -> (672, 112, 1, 1)[FLOAT]], [933 -> (672)[FLOAT]], [X] Convolution input dimensions: (1, 112, 18, 32) [X] Registering layer: Conv_119 for ONNX node: Conv_119 [X] Using kernel: (1, 1), strides: (1, 1), prepadding: (0, 0), postpadding: (0, 0), dilations: (1, 1), numOutputs: 672 [X] Convolution output dimensions: (1, 672, 18, 32) [X] Registering tensor: 931 for ONNX tensor: 931 [X] Conv_119 [Conv] outputs: [931 -> (1, 672, 18, 32)[FLOAT]], [X] Parsing node: HardSigmoid_120 [HardSigmoid] [X] Searching for input: 931 [X] HardSigmoid_120 [HardSigmoid] inputs: [931 -> (1, 672, 18, 32)[FLOAT]], [X] Registering layer: HardSigmoid_120 for ONNX node: HardSigmoid_120 [X] Registering tensor: 543 for ONNX tensor: 543 [X] HardSigmoid_120 [HardSigmoid] outputs: [543 -> (1, 672, 18, 32)[FLOAT]], [X] Parsing node: Mul_121 [Mul] [X] Searching for input: 931 [X] Searching for input: 543 [X] Mul_121 [Mul] inputs: [931 -> (1, 672, 18, 32)[FLOAT]], [543 -> (1, 672, 18, 32)[FLOAT]], [X] Registering layer: Mul_121 for ONNX node: Mul_121 [X] Registering tensor: 544 for ONNX tensor: 544 [X] Mul_121 [Mul] outputs: [544 -> (1, 672, 18, 32)[FLOAT]], [X] Parsing node: Conv_122 [Conv] [X] Searching for input: 544 [X] Searching for input: 935 [X] Searching for input: 936 [X] Conv_122 [Conv] inputs: [544 -> (1, 672, 18, 32)[FLOAT]], [935 -> (672, 1, 5, 5)[FLOAT]], [936 -> (672)[FLOAT]], [X] Convolution input dimensions: (1, 672, 18, 32) [X] Registering layer: Conv_122 for ONNX node: Conv_122 [X] Using kernel: (5, 5), strides: (1, 1), prepadding: (4, 4), postpadding: (4, 4), dilations: (2, 2), numOutputs: 672 [X] Convolution output dimensions: (1, 672, 18, 32) [X] Registering tensor: 934 for ONNX tensor: 934 [X] Conv_122 [Conv] outputs: [934 -> (1, 672, 18, 32)[FLOAT]], [X] Parsing node: HardSigmoid_123 [HardSigmoid] [X] Searching for input: 934 [X] HardSigmoid_123 [HardSigmoid] inputs: [934 -> (1, 672, 18, 32)[FLOAT]], [X] Registering layer: HardSigmoid_123 for ONNX node: HardSigmoid_123 [X] Registering tensor: 547 for ONNX tensor: 547 [X] HardSigmoid_123 [HardSigmoid] outputs: [547 -> (1, 672, 18, 32)[FLOAT]], [X] Parsing node: Mul_124 [Mul] [X] Searching for input: 934 [X] Searching for input: 547 [X] Mul_124 [Mul] inputs: [934 -> (1, 672, 18, 32)[FLOAT]], [547 -> (1, 672, 18, 32)[FLOAT]], [X] Registering layer: Mul_124 for ONNX node: Mul_124 [X] Registering tensor: 548 for ONNX tensor: 548 [X] Mul_124 [Mul] outputs: [548 -> (1, 672, 18, 32)[FLOAT]], [X] Parsing node: GlobalAveragePool_125 [GlobalAveragePool] [X] Searching for input: 548 [X] GlobalAveragePool_125 [GlobalAveragePool] inputs: [548 -> (1, 672, 18, 32)[FLOAT]], [X] GlobalAveragePool operators are implemented via Reduce layers rather than Pooling layers [X] Registering layer: GlobalAveragePool_125 for ONNX node: GlobalAveragePool_125 [X] Registering tensor: 549 for ONNX tensor: 549 [X] GlobalAveragePool_125 [GlobalAveragePool] outputs: [549 -> (1, 672, 1, 1)[FLOAT]], [X] Parsing node: Conv_126 [Conv] [X] Searching for input: 549 [X] Searching for input: backbone.features.13.block.2.fc1.weight [X] Searching for input: backbone.features.13.block.2.fc1.bias [X] Conv_126 [Conv] inputs: [549 -> (1, 672, 1, 1)[FLOAT]], [backbone.features.13.block.2.fc1.weight -> (168, 672, 1, 1)[FLOAT]], [backbone.features.13.block.2.fc1.bias -> (168)[FLOAT]], [X] Convolution input dimensions: (1, 672, 1, 1) [X] Registering layer: Conv_126 for ONNX node: Conv_126 [X] Using kernel: (1, 1), strides: (1, 1), prepadding: (0, 0), postpadding: (0, 0), dilations: (1, 1), numOutputs: 168 [X] Convolution output dimensions: (1, 168, 1, 1) [X] Registering tensor: 550 for ONNX tensor: 550 [X] Conv_126 [Conv] outputs: [550 -> (1, 168, 1, 1)[FLOAT]], [X] Parsing node: Relu_127 [Relu] [X] Searching for input: 550 [X] Relu_127 [Relu] inputs: [550 -> (1, 168, 1, 1)[FLOAT]], [X] Registering layer: Relu_127 for ONNX node: Relu_127 [X] Registering tensor: 551 for ONNX tensor: 551 [X] Relu_127 [Relu] outputs: [551 -> (1, 168, 1, 1)[FLOAT]], [X] Parsing node: Conv_128 [Conv] [X] Searching for input: 551 [X] Searching for input: backbone.features.13.block.2.fc2.weight [X] Searching for input: backbone.features.13.block.2.fc2.bias [X] Conv_128 [Conv] inputs: [551 -> (1, 168, 1, 1)[FLOAT]], [backbone.features.13.block.2.fc2.weight -> (672, 168, 1, 1)[FLOAT]], [backbone.features.13.block.2.fc2.bias -> (672)[FLOAT]], [X] Convolution input dimensions: (1, 168, 1, 1) [X] Registering layer: Conv_128 for ONNX node: Conv_128 [X] Using kernel: (1, 1), strides: (1, 1), prepadding: (0, 0), postpadding: (0, 0), dilations: (1, 1), numOutputs: 672 [X] Convolution output dimensions: (1, 672, 1, 1) [X] Registering tensor: 552 for ONNX tensor: 552 [X] Conv_128 [Conv] outputs: [552 -> (1, 672, 1, 1)[FLOAT]], [X] Parsing node: HardSigmoid_129 [HardSigmoid] [X] Searching for input: 552 [X] HardSigmoid_129 [HardSigmoid] inputs: [552 -> (1, 672, 1, 1)[FLOAT]], [X] Registering layer: HardSigmoid_129 for ONNX node: HardSigmoid_129 [X] Registering tensor: 553 for ONNX tensor: 553 [X] HardSigmoid_129 [HardSigmoid] outputs: [553 -> (1, 672, 1, 1)[FLOAT]], [X] Parsing node: Mul_130 [Mul] [X] Searching for input: 553 [X] Searching for input: 548 [X] Mul_130 [Mul] inputs: [553 -> (1, 672, 1, 1)[FLOAT]], [548 -> (1, 672, 18, 32)[FLOAT]], [X] Registering layer: Mul_130 for ONNX node: Mul_130 [X] Registering tensor: 554 for ONNX tensor: 554 [X] Mul_130 [Mul] outputs: [554 -> (1, 672, 18, 32)[FLOAT]], [X] Parsing node: Conv_131 [Conv] [X] Searching for input: 554 [X] Searching for input: 938 [X] Searching for input: 939 [X] Conv_131 [Conv] inputs: [554 -> (1, 672, 18, 32)[FLOAT]], [938 -> (160, 672, 1, 1)[FLOAT]], [939 -> (160)[FLOAT]], [X] Convolution input dimensions: (1, 672, 18, 32) [X] Registering layer: Conv_131 for ONNX node: Conv_131 [X] Using kernel: (1, 1), strides: (1, 1), prepadding: (0, 0), postpadding: (0, 0), dilations: (1, 1), numOutputs: 160 [X] Convolution output dimensions: (1, 160, 18, 32) [X] Registering tensor: 937 for ONNX tensor: 937 [X] Conv_131 [Conv] outputs: [937 -> (1, 160, 18, 32)[FLOAT]], [X] Parsing node: Conv_132 [Conv] [X] Searching for input: 937 [X] Searching for input: 941 [X] Searching for input: 942 [X] Conv_132 [Conv] inputs: [937 -> (1, 160, 18, 32)[FLOAT]], [941 -> (960, 160, 1, 1)[FLOAT]], [942 -> (960)[FLOAT]], [X] Convolution input dimensions: (1, 160, 18, 32) [X] Registering layer: Conv_132 for ONNX node: Conv_132 [X] Using kernel: (1, 1), strides: (1, 1), prepadding: (0, 0), postpadding: (0, 0), dilations: (1, 1), numOutputs: 960 [X] Convolution output dimensions: (1, 960, 18, 32) [X] Registering tensor: 940 for ONNX tensor: 940 [X] Conv_132 [Conv] outputs: [940 -> (1, 960, 18, 32)[FLOAT]], [X] Parsing node: HardSigmoid_133 [HardSigmoid] [X] Searching for input: 940 [X] HardSigmoid_133 [HardSigmoid] inputs: [940 -> (1, 960, 18, 32)[FLOAT]], [X] Registering layer: HardSigmoid_133 for ONNX node: HardSigmoid_133 [X] Registering tensor: 559 for ONNX tensor: 559 [X] HardSigmoid_133 [HardSigmoid] outputs: [559 -> (1, 960, 18, 32)[FLOAT]], [X] Parsing node: Mul_134 [Mul] [X] Searching for input: 940 [X] Searching for input: 559 [X] Mul_134 [Mul] inputs: [940 -> (1, 960, 18, 32)[FLOAT]], [559 -> (1, 960, 18, 32)[FLOAT]], [X] Registering layer: Mul_134 for ONNX node: Mul_134 [X] Registering tensor: 560 for ONNX tensor: 560 [X] Mul_134 [Mul] outputs: [560 -> (1, 960, 18, 32)[FLOAT]], [X] Parsing node: Conv_135 [Conv] [X] Searching for input: 560 [X] Searching for input: 944 [X] Searching for input: 945 [X] Conv_135 [Conv] inputs: [560 -> (1, 960, 18, 32)[FLOAT]], [944 -> (960, 1, 5, 5)[FLOAT]], [945 -> (960)[FLOAT]], [X] Convolution input dimensions: (1, 960, 18, 32) [X] Registering layer: Conv_135 for ONNX node: Conv_135 [X] Using kernel: (5, 5), strides: (1, 1), prepadding: (4, 4), postpadding: (4, 4), dilations: (2, 2), numOutputs: 960 [X] Convolution output dimensions: (1, 960, 18, 32) [X] Registering tensor: 943 for ONNX tensor: 943 [X] Conv_135 [Conv] outputs: [943 -> (1, 960, 18, 32)[FLOAT]], [X] Parsing node: HardSigmoid_136 [HardSigmoid] [X] Searching for input: 943 [X] HardSigmoid_136 [HardSigmoid] inputs: [943 -> (1, 960, 18, 32)[FLOAT]], [X] Registering layer: HardSigmoid_136 for ONNX node: HardSigmoid_136 [X] Registering tensor: 563 for ONNX tensor: 563 [X] HardSigmoid_136 [HardSigmoid] outputs: [563 -> (1, 960, 18, 32)[FLOAT]], [X] Parsing node: Mul_137 [Mul] [X] Searching for input: 943 [X] Searching for input: 563 [X] Mul_137 [Mul] inputs: [943 -> (1, 960, 18, 32)[FLOAT]], [563 -> (1, 960, 18, 32)[FLOAT]], [X] Registering layer: Mul_137 for ONNX node: Mul_137 [X] Registering tensor: 564 for ONNX tensor: 564 [X] Mul_137 [Mul] outputs: [564 -> (1, 960, 18, 32)[FLOAT]], [X] Parsing node: GlobalAveragePool_138 [GlobalAveragePool] [X] Searching for input: 564 [X] GlobalAveragePool_138 [GlobalAveragePool] inputs: [564 -> (1, 960, 18, 32)[FLOAT]], [X] GlobalAveragePool operators are implemented via Reduce layers rather than Pooling layers [X] Registering layer: GlobalAveragePool_138 for ONNX node: GlobalAveragePool_138 [X] Registering tensor: 565 for ONNX tensor: 565 [X] GlobalAveragePool_138 [GlobalAveragePool] outputs: [565 -> (1, 960, 1, 1)[FLOAT]], [X] Parsing node: Conv_139 [Conv] [X] Searching for input: 565 [X] Searching for input: backbone.features.14.block.2.fc1.weight [X] Searching for input: backbone.features.14.block.2.fc1.bias [X] Conv_139 [Conv] inputs: [565 -> (1, 960, 1, 1)[FLOAT]], [backbone.features.14.block.2.fc1.weight -> (240, 960, 1, 1)[FLOAT]], [backbone.features.14.block.2.fc1.bias -> (240)[FLOAT]], [X] Convolution input dimensions: (1, 960, 1, 1) [X] Registering layer: Conv_139 for ONNX node: Conv_139 [X] Using kernel: (1, 1), strides: (1, 1), prepadding: (0, 0), postpadding: (0, 0), dilations: (1, 1), numOutputs: 240 [X] Convolution output dimensions: (1, 240, 1, 1) [X] Registering tensor: 566 for ONNX tensor: 566 [X] Conv_139 [Conv] outputs: [566 -> (1, 240, 1, 1)[FLOAT]], [X] Parsing node: Relu_140 [Relu] [X] Searching for input: 566 [X] Relu_140 [Relu] inputs: [566 -> (1, 240, 1, 1)[FLOAT]], [X] Registering layer: Relu_140 for ONNX node: Relu_140 [X] Registering tensor: 567 for ONNX tensor: 567 [X] Relu_140 [Relu] outputs: [567 -> (1, 240, 1, 1)[FLOAT]], [X] Parsing node: Conv_141 [Conv] [X] Searching for input: 567 [X] Searching for input: backbone.features.14.block.2.fc2.weight [X] Searching for input: backbone.features.14.block.2.fc2.bias [X] Conv_141 [Conv] inputs: [567 -> (1, 240, 1, 1)[FLOAT]], [backbone.features.14.block.2.fc2.weight -> (960, 240, 1, 1)[FLOAT]], [backbone.features.14.block.2.fc2.bias -> (960)[FLOAT]], [X] Convolution input dimensions: (1, 240, 1, 1) [X] Registering layer: Conv_141 for ONNX node: Conv_141 [X] Using kernel: (1, 1), strides: (1, 1), prepadding: (0, 0), postpadding: (0, 0), dilations: (1, 1), numOutputs: 960 [X] Convolution output dimensions: (1, 960, 1, 1) [X] Registering tensor: 568 for ONNX tensor: 568 [X] Conv_141 [Conv] outputs: [568 -> (1, 960, 1, 1)[FLOAT]], [X] Parsing node: HardSigmoid_142 [HardSigmoid] [X] Searching for input: 568 [X] HardSigmoid_142 [HardSigmoid] inputs: [568 -> (1, 960, 1, 1)[FLOAT]], [X] Registering layer: HardSigmoid_142 for ONNX node: HardSigmoid_142 [X] Registering tensor: 569 for ONNX tensor: 569 [X] HardSigmoid_142 [HardSigmoid] outputs: [569 -> (1, 960, 1, 1)[FLOAT]], [X] Parsing node: Mul_143 [Mul] [X] Searching for input: 569 [X] Searching for input: 564 [X] Mul_143 [Mul] inputs: [569 -> (1, 960, 1, 1)[FLOAT]], [564 -> (1, 960, 18, 32)[FLOAT]], [X] Registering layer: Mul_143 for ONNX node: Mul_143 [X] Registering tensor: 570 for ONNX tensor: 570 [X] Mul_143 [Mul] outputs: [570 -> (1, 960, 18, 32)[FLOAT]], [X] Parsing node: Conv_144 [Conv] [X] Searching for input: 570 [X] Searching for input: 947 [X] Searching for input: 948 [X] Conv_144 [Conv] inputs: [570 -> (1, 960, 18, 32)[FLOAT]], [947 -> (160, 960, 1, 1)[FLOAT]], [948 -> (160)[FLOAT]], [X] Convolution input dimensions: (1, 960, 18, 32) [X] Registering layer: Conv_144 for ONNX node: Conv_144 [X] Using kernel: (1, 1), strides: (1, 1), prepadding: (0, 0), postpadding: (0, 0), dilations: (1, 1), numOutputs: 160 [X] Convolution output dimensions: (1, 160, 18, 32) [X] Registering tensor: 946 for ONNX tensor: 946 [X] Conv_144 [Conv] outputs: [946 -> (1, 160, 18, 32)[FLOAT]], [X] Parsing node: Add_145 [Add] [X] Searching for input: 946 [X] Searching for input: 937 [X] Add_145 [Add] inputs: [946 -> (1, 160, 18, 32)[FLOAT]], [937 -> (1, 160, 18, 32)[FLOAT]], [X] Registering layer: Add_145 for ONNX node: Add_145 [X] Registering tensor: 573 for ONNX tensor: 573 [X] Add_145 [Add] outputs: [573 -> (1, 160, 18, 32)[FLOAT]], [X] Parsing node: Conv_146 [Conv] [X] Searching for input: 573 [X] Searching for input: 950 [X] Searching for input: 951 [X] Conv_146 [Conv] inputs: [573 -> (1, 160, 18, 32)[FLOAT]], [950 -> (960, 160, 1, 1)[FLOAT]], [951 -> (960)[FLOAT]], [X] Convolution input dimensions: (1, 160, 18, 32) [X] Registering layer: Conv_146 for ONNX node: Conv_146 [X] Using kernel: (1, 1), strides: (1, 1), prepadding: (0, 0), postpadding: (0, 0), dilations: (1, 1), numOutputs: 960 [X] Convolution output dimensions: (1, 960, 18, 32) [X] Registering tensor: 949 for ONNX tensor: 949 [X] Conv_146 [Conv] outputs: [949 -> (1, 960, 18, 32)[FLOAT]], [X] Parsing node: HardSigmoid_147 [HardSigmoid] [X] Searching for input: 949 [X] HardSigmoid_147 [HardSigmoid] inputs: [949 -> (1, 960, 18, 32)[FLOAT]], [X] Registering layer: HardSigmoid_147 for ONNX node: HardSigmoid_147 [X] Registering tensor: 576 for ONNX tensor: 576 [X] HardSigmoid_147 [HardSigmoid] outputs: [576 -> (1, 960, 18, 32)[FLOAT]], [X] Parsing node: Mul_148 [Mul] [X] Searching for input: 949 [X] Searching for input: 576 [X] Mul_148 [Mul] inputs: [949 -> (1, 960, 18, 32)[FLOAT]], [576 -> (1, 960, 18, 32)[FLOAT]], [X] Registering layer: Mul_148 for ONNX node: Mul_148 [X] Registering tensor: 577 for ONNX tensor: 577 [X] Mul_148 [Mul] outputs: [577 -> (1, 960, 18, 32)[FLOAT]], [X] Parsing node: Conv_149 [Conv] [X] Searching for input: 577 [X] Searching for input: 953 [X] Searching for input: 954 [X] Conv_149 [Conv] inputs: [577 -> (1, 960, 18, 32)[FLOAT]], [953 -> (960, 1, 5, 5)[FLOAT]], [954 -> (960)[FLOAT]], [X] Convolution input dimensions: (1, 960, 18, 32) [X] Registering layer: Conv_149 for ONNX node: Conv_149 [X] Using kernel: (5, 5), strides: (1, 1), prepadding: (4, 4), postpadding: (4, 4), dilations: (2, 2), numOutputs: 960 [X] Convolution output dimensions: (1, 960, 18, 32) [X] Registering tensor: 952 for ONNX tensor: 952 [X] Conv_149 [Conv] outputs: [952 -> (1, 960, 18, 32)[FLOAT]], [X] Parsing node: HardSigmoid_150 [HardSigmoid] [X] Searching for input: 952 [X] HardSigmoid_150 [HardSigmoid] inputs: [952 -> (1, 960, 18, 32)[FLOAT]], [X] Registering layer: HardSigmoid_150 for ONNX node: HardSigmoid_150 [X] Registering tensor: 580 for ONNX tensor: 580 [X] HardSigmoid_150 [HardSigmoid] outputs: [580 -> (1, 960, 18, 32)[FLOAT]], [X] Parsing node: Mul_151 [Mul] [X] Searching for input: 952 [X] Searching for input: 580 [X] Mul_151 [Mul] inputs: [952 -> (1, 960, 18, 32)[FLOAT]], [580 -> (1, 960, 18, 32)[FLOAT]], [X] Registering layer: Mul_151 for ONNX node: Mul_151 [X] Registering tensor: 581 for ONNX tensor: 581 [X] Mul_151 [Mul] outputs: [581 -> (1, 960, 18, 32)[FLOAT]], [X] Parsing node: GlobalAveragePool_152 [GlobalAveragePool] [X] Searching for input: 581 [X] GlobalAveragePool_152 [GlobalAveragePool] inputs: [581 -> (1, 960, 18, 32)[FLOAT]], [X] GlobalAveragePool operators are implemented via Reduce layers rather than Pooling layers [X] Registering layer: GlobalAveragePool_152 for ONNX node: GlobalAveragePool_152 [X] Registering tensor: 582 for ONNX tensor: 582 [X] GlobalAveragePool_152 [GlobalAveragePool] outputs: [582 -> (1, 960, 1, 1)[FLOAT]], [X] Parsing node: Conv_153 [Conv] [X] Searching for input: 582 [X] Searching for input: backbone.features.15.block.2.fc1.weight [X] Searching for input: backbone.features.15.block.2.fc1.bias [X] Conv_153 [Conv] inputs: [582 -> (1, 960, 1, 1)[FLOAT]], [backbone.features.15.block.2.fc1.weight -> (240, 960, 1, 1)[FLOAT]], [backbone.features.15.block.2.fc1.bias -> (240)[FLOAT]], [X] Convolution input dimensions: (1, 960, 1, 1) [X] Registering layer: Conv_153 for ONNX node: Conv_153 [X] Using kernel: (1, 1), strides: (1, 1), prepadding: (0, 0), postpadding: (0, 0), dilations: (1, 1), numOutputs: 240 [X] Convolution output dimensions: (1, 240, 1, 1) [X] Registering tensor: 583 for ONNX tensor: 583 [X] Conv_153 [Conv] outputs: [583 -> (1, 240, 1, 1)[FLOAT]], [X] Parsing node: Relu_154 [Relu] [X] Searching for input: 583 [X] Relu_154 [Relu] inputs: [583 -> (1, 240, 1, 1)[FLOAT]], [X] Registering layer: Relu_154 for ONNX node: Relu_154 [X] Registering tensor: 584 for ONNX tensor: 584 [X] Relu_154 [Relu] outputs: [584 -> (1, 240, 1, 1)[FLOAT]], [X] Parsing node: Conv_155 [Conv] [X] Searching for input: 584 [X] Searching for input: backbone.features.15.block.2.fc2.weight [X] Searching for input: backbone.features.15.block.2.fc2.bias [X] Conv_155 [Conv] inputs: [584 -> (1, 240, 1, 1)[FLOAT]], [backbone.features.15.block.2.fc2.weight -> (960, 240, 1, 1)[FLOAT]], [backbone.features.15.block.2.fc2.bias -> (960)[FLOAT]], [X] Convolution input dimensions: (1, 240, 1, 1) [X] Registering layer: Conv_155 for ONNX node: Conv_155 [X] Using kernel: (1, 1), strides: (1, 1), prepadding: (0, 0), postpadding: (0, 0), dilations: (1, 1), numOutputs: 960 [X] Convolution output dimensions: (1, 960, 1, 1) [X] Registering tensor: 585 for ONNX tensor: 585 [X] Conv_155 [Conv] outputs: [585 -> (1, 960, 1, 1)[FLOAT]], [X] Parsing node: HardSigmoid_156 [HardSigmoid] [X] Searching for input: 585 [X] HardSigmoid_156 [HardSigmoid] inputs: [585 -> (1, 960, 1, 1)[FLOAT]], [X] Registering layer: HardSigmoid_156 for ONNX node: HardSigmoid_156 [X] Registering tensor: 586 for ONNX tensor: 586 [X] HardSigmoid_156 [HardSigmoid] outputs: [586 -> (1, 960, 1, 1)[FLOAT]], [X] Parsing node: Mul_157 [Mul] [X] Searching for input: 586 [X] Searching for input: 581 [X] Mul_157 [Mul] inputs: [586 -> (1, 960, 1, 1)[FLOAT]], [581 -> (1, 960, 18, 32)[FLOAT]], [X] Registering layer: Mul_157 for ONNX node: Mul_157 [X] Registering tensor: 587 for ONNX tensor: 587 [X] Mul_157 [Mul] outputs: [587 -> (1, 960, 18, 32)[FLOAT]], [X] Parsing node: Conv_158 [Conv] [X] Searching for input: 587 [X] Searching for input: 956 [X] Searching for input: 957 [X] Conv_158 [Conv] inputs: [587 -> (1, 960, 18, 32)[FLOAT]], [956 -> (160, 960, 1, 1)[FLOAT]], [957 -> (160)[FLOAT]], [X] Convolution input dimensions: (1, 960, 18, 32) [X] Registering layer: Conv_158 for ONNX node: Conv_158 [X] Using kernel: (1, 1), strides: (1, 1), prepadding: (0, 0), postpadding: (0, 0), dilations: (1, 1), numOutputs: 160 [X] Convolution output dimensions: (1, 160, 18, 32) [X] Registering tensor: 955 for ONNX tensor: 955 [X] Conv_158 [Conv] outputs: [955 -> (1, 160, 18, 32)[FLOAT]], [X] Parsing node: Add_159 [Add] [X] Searching for input: 955 [X] Searching for input: 573 [X] Add_159 [Add] inputs: [955 -> (1, 160, 18, 32)[FLOAT]], [573 -> (1, 160, 18, 32)[FLOAT]], [X] Registering layer: Add_159 for ONNX node: Add_159 [X] Registering tensor: 590 for ONNX tensor: 590 [X] Add_159 [Add] outputs: [590 -> (1, 160, 18, 32)[FLOAT]], [X] Parsing node: Conv_160 [Conv] [X] Searching for input: 590 [X] Searching for input: 959 [X] Searching for input: 960 [X] Conv_160 [Conv] inputs: [590 -> (1, 160, 18, 32)[FLOAT]], [959 -> (960, 160, 1, 1)[FLOAT]], [960 -> (960)[FLOAT]], [X] Convolution input dimensions: (1, 160, 18, 32) [X] Registering layer: Conv_160 for ONNX node: Conv_160 [X] Using kernel: (1, 1), strides: (1, 1), prepadding: (0, 0), postpadding: (0, 0), dilations: (1, 1), numOutputs: 960 [X] Convolution output dimensions: (1, 960, 18, 32) [X] Registering tensor: 958 for ONNX tensor: 958 [X] Conv_160 [Conv] outputs: [958 -> (1, 960, 18, 32)[FLOAT]], [X] Parsing node: HardSigmoid_161 [HardSigmoid] [X] Searching for input: 958 [X] HardSigmoid_161 [HardSigmoid] inputs: [958 -> (1, 960, 18, 32)[FLOAT]], [X] Registering layer: HardSigmoid_161 for ONNX node: HardSigmoid_161 [X] Registering tensor: 593 for ONNX tensor: 593 [X] HardSigmoid_161 [HardSigmoid] outputs: [593 -> (1, 960, 18, 32)[FLOAT]], [X] Parsing node: Mul_162 [Mul] [X] Searching for input: 958 [X] Searching for input: 593 [X] Mul_162 [Mul] inputs: [958 -> (1, 960, 18, 32)[FLOAT]], [593 -> (1, 960, 18, 32)[FLOAT]], [X] Registering layer: Mul_162 for ONNX node: Mul_162 [X] Registering tensor: 594 for ONNX tensor: 594 [X] Mul_162 [Mul] outputs: [594 -> (1, 960, 18, 32)[FLOAT]], [X] Parsing node: Conv_163 [Conv] [X] Searching for input: 594 [X] Searching for input: 962 [X] Searching for input: 963 [X] Conv_163 [Conv] inputs: [594 -> (1, 960, 18, 32)[FLOAT]], [962 -> (128, 960, 1, 1)[FLOAT]], [963 -> (128)[FLOAT]], [X] Convolution input dimensions: (1, 960, 18, 32) [X] Registering layer: Conv_163 for ONNX node: Conv_163 [X] Using kernel: (1, 1), strides: (1, 1), prepadding: (0, 0), postpadding: (0, 0), dilations: (1, 1), numOutputs: 128 [X] Convolution output dimensions: (1, 128, 18, 32) [X] Registering tensor: 961 for ONNX tensor: 961 [X] Conv_163 [Conv] outputs: [961 -> (1, 128, 18, 32)[FLOAT]], [X] Parsing node: Relu_164 [Relu] [X] Searching for input: 961 [X] Relu_164 [Relu] inputs: [961 -> (1, 128, 18, 32)[FLOAT]], [X] Registering layer: Relu_164 for ONNX node: Relu_164 [X] Registering tensor: 597 for ONNX tensor: 597 [X] Relu_164 [Relu] outputs: [597 -> (1, 128, 18, 32)[FLOAT]], [X] Parsing node: GlobalAveragePool_165 [GlobalAveragePool] [X] Searching for input: 594 [X] GlobalAveragePool_165 [GlobalAveragePool] inputs: [594 -> (1, 960, 18, 32)[FLOAT]], [X] GlobalAveragePool operators are implemented via Reduce layers rather than Pooling layers [X] Registering layer: GlobalAveragePool_165 for ONNX node: GlobalAveragePool_165 [X] Registering tensor: 598 for ONNX tensor: 598 [X] GlobalAveragePool_165 [GlobalAveragePool] outputs: [598 -> (1, 960, 1, 1)[FLOAT]], [X] Parsing node: Conv_166 [Conv] [X] Searching for input: 598 [X] Searching for input: aspp.aspp2.1.weight [X] Conv_166 [Conv] inputs: [598 -> (1, 960, 1, 1)[FLOAT]], [aspp.aspp2.1.weight -> (128, 960, 1, 1)[FLOAT]], [X] Convolution input dimensions: (1, 960, 1, 1) [X] Registering layer: Conv_166 for ONNX node: Conv_166 [X] Using kernel: (1, 1), strides: (1, 1), prepadding: (0, 0), postpadding: (0, 0), dilations: (1, 1), numOutputs: 128 [X] Convolution output dimensions: (1, 128, 1, 1) [X] Registering tensor: 599 for ONNX tensor: 599 [X] Conv_166 [Conv] outputs: [599 -> (1, 128, 1, 1)[FLOAT]], [X] Parsing node: Sigmoid_167 [Sigmoid] [X] Searching for input: 599 [X] Sigmoid_167 [Sigmoid] inputs: [599 -> (1, 128, 1, 1)[FLOAT]], [X] Registering layer: Sigmoid_167 for ONNX node: Sigmoid_167 [X] Registering tensor: 600 for ONNX tensor: 600 [X] Sigmoid_167 [Sigmoid] outputs: [600 -> (1, 128, 1, 1)[FLOAT]], [X] Parsing node: Mul_168 [Mul] [X] Searching for input: 597 [X] Searching for input: 600 [X] Mul_168 [Mul] inputs: [597 -> (1, 128, 18, 32)[FLOAT]], [600 -> (1, 128, 1, 1)[FLOAT]], [X] Registering layer: Mul_168 for ONNX node: Mul_168 [X] Registering tensor: 601 for ONNX tensor: 601 [X] Mul_168 [Mul] outputs: [601 -> (1, 128, 18, 32)[FLOAT]], [X] Parsing node: AveragePool_169 [AveragePool] [X] Searching for input: 389 [X] AveragePool_169 [AveragePool] inputs: [389 -> (1, 3, 288, 512)[FLOAT]], [X] Registering layer: AveragePool_169 for ONNX node: AveragePool_169 [X] Registering tensor: 602 for ONNX tensor: 602 [X] AveragePool_169 [AveragePool] outputs: [602 -> (1, 3, 144, 256)[FLOAT]], [X] Parsing node: AveragePool_170 [AveragePool] [X] Searching for input: 602 [X] AveragePool_170 [AveragePool] inputs: [602 -> (1, 3, 144, 256)[FLOAT]], [X] Registering layer: AveragePool_170 for ONNX node: AveragePool_170 [X] Registering tensor: 603 for ONNX tensor: 603 [X] AveragePool_170 [AveragePool] outputs: [603 -> (1, 3, 72, 128)[FLOAT]], [X] Parsing node: AveragePool_171 [AveragePool] [X] Searching for input: 603 [X] AveragePool_171 [AveragePool] inputs: [603 -> (1, 3, 72, 128)[FLOAT]], [X] Registering layer: AveragePool_171 for ONNX node: AveragePool_171 [X] Registering tensor: 604 for ONNX tensor: 604 [X] AveragePool_171 [AveragePool] outputs: [604 -> (1, 3, 36, 64)[FLOAT]], [X] Parsing node: Split_172 [Split] [X] Searching for input: 601 [X] Split_172 [Split] inputs: [601 -> (1, 128, 18, 32)[FLOAT]], [X] Registering layer: Split_172 for ONNX node: Split_172 [X] Registering layer: Split_172_0 for ONNX node: Split_172 [X] Registering tensor: 605 for ONNX tensor: 605 [X] Registering tensor: 606 for ONNX tensor: 606 [X] Split_172 [Split] outputs: [605 -> (1, 64, 18, 32)[FLOAT]], [606 -> (1, 64, 18, 32)[FLOAT]], [X] Parsing node: Shape_173 [Shape] [X] Searching for input: 606 [X] Shape_173 [Shape] inputs: [606 -> (1, 64, 18, 32)[FLOAT]], [X] Registering layer: Shape_173 for ONNX node: Shape_173 [X] Registering tensor: 607 for ONNX tensor: 607 [X] Shape_173 [Shape] outputs: [607 -> (4)[INT32]], [X] Parsing node: Expand_174 [Expand] [X] Searching for input: r4i [X] Searching for input: 607 [X] Expand_174 [Expand] inputs: [r4i -> (1, 1, 1, 1)[FLOAT]], [607 -> (4)[INT32]], [W] onnx2trt_utils.cpp:367: Your ONNX model has been generated with INT64 weights, while TensorRT does not natively support INT64. Attempting to cast down to INT32. [X] Registering layer: Expand_174 for ONNX node: Expand_174 [X] Registering tensor: 608 for ONNX tensor: 608 [X] Expand_174 [Expand] outputs: [608 -> (1, 64, 18, 32)[FLOAT]], [X] Parsing node: Concat_175 [Concat] [X] Searching for input: 606 [X] Searching for input: 608 [X] Concat_175 [Concat] inputs: [606 -> (1, 64, 18, 32)[FLOAT]], [608 -> (1, 64, 18, 32)[FLOAT]], [X] Registering layer: Concat_175 for ONNX node: Concat_175 [X] Registering tensor: 609 for ONNX tensor: 609 [X] Concat_175 [Concat] outputs: [609 -> (1, 128, 18, 32)[FLOAT]], [X] Parsing node: Conv_176 [Conv] [X] Searching for input: 609 [X] Searching for input: decoder.decode4.gru.ih.0.weight [X] Searching for input: decoder.decode4.gru.ih.0.bias [X] Conv_176 [Conv] inputs: [609 -> (1, 128, 18, 32)[FLOAT]], [decoder.decode4.gru.ih.0.weight -> (128, 128, 3, 3)[FLOAT]], [decoder.decode4.gru.ih.0.bias -> (128)[FLOAT]], [X] Convolution input dimensions: (1, 128, 18, 32) [X] Registering layer: Conv_176 for ONNX node: Conv_176 [X] Using kernel: (3, 3), strides: (1, 1), prepadding: (1, 1), postpadding: (1, 1), dilations: (1, 1), numOutputs: 128 [X] Convolution output dimensions: (1, 128, 18, 32) [X] Registering tensor: 610 for ONNX tensor: 610 [X] Conv_176 [Conv] outputs: [610 -> (1, 128, 18, 32)[FLOAT]], [X] Parsing node: Sigmoid_177 [Sigmoid] [X] Searching for input: 610 [X] Sigmoid_177 [Sigmoid] inputs: [610 -> (1, 128, 18, 32)[FLOAT]], [X] Registering layer: Sigmoid_177 for ONNX node: Sigmoid_177 [X] Registering tensor: 611 for ONNX tensor: 611 [X] Sigmoid_177 [Sigmoid] outputs: [611 -> (1, 128, 18, 32)[FLOAT]], [X] Parsing node: Split_178 [Split] [X] Searching for input: 611 [X] Split_178 [Split] inputs: [611 -> (1, 128, 18, 32)[FLOAT]], [X] Registering layer: Split_178 for ONNX node: Split_178 [X] Registering layer: Split_178_2 for ONNX node: Split_178 [X] Registering tensor: 612 for ONNX tensor: 612 [X] Registering tensor: 613 for ONNX tensor: 613 [X] Split_178 [Split] outputs: [612 -> (1, 64, 18, 32)[FLOAT]], [613 -> (1, 64, 18, 32)[FLOAT]], [X] Parsing node: Mul_179 [Mul] [X] Searching for input: 612 [X] Searching for input: 608 [X] Mul_179 [Mul] inputs: [612 -> (1, 64, 18, 32)[FLOAT]], [608 -> (1, 64, 18, 32)[FLOAT]], [X] Registering layer: Mul_179 for ONNX node: Mul_179 [X] Registering tensor: 614 for ONNX tensor: 614 [X] Mul_179 [Mul] outputs: [614 -> (1, 64, 18, 32)[FLOAT]], [X] Parsing node: Concat_180 [Concat] [X] Searching for input: 606 [X] Searching for input: 614 [X] Concat_180 [Concat] inputs: [606 -> (1, 64, 18, 32)[FLOAT]], [614 -> (1, 64, 18, 32)[FLOAT]], [X] Registering layer: Concat_180 for ONNX node: Concat_180 [X] Registering tensor: 615 for ONNX tensor: 615 [X] Concat_180 [Concat] outputs: [615 -> (1, 128, 18, 32)[FLOAT]], [X] Parsing node: Conv_181 [Conv] [X] Searching for input: 615 [X] Searching for input: decoder.decode4.gru.hh.0.weight [X] Searching for input: decoder.decode4.gru.hh.0.bias [X] Conv_181 [Conv] inputs: [615 -> (1, 128, 18, 32)[FLOAT]], [decoder.decode4.gru.hh.0.weight -> (64, 128, 3, 3)[FLOAT]], [decoder.decode4.gru.hh.0.bias -> (64)[FLOAT]], [X] Convolution input dimensions: (1, 128, 18, 32) [X] Registering layer: Conv_181 for ONNX node: Conv_181 [X] Using kernel: (3, 3), strides: (1, 1), prepadding: (1, 1), postpadding: (1, 1), dilations: (1, 1), numOutputs: 64 [X] Convolution output dimensions: (1, 64, 18, 32) [X] Registering tensor: 616 for ONNX tensor: 616 [X] Conv_181 [Conv] outputs: [616 -> (1, 64, 18, 32)[FLOAT]], [X] Parsing node: Tanh_182 [Tanh] [X] Searching for input: 616 [X] Tanh_182 [Tanh] inputs: [616 -> (1, 64, 18, 32)[FLOAT]], [X] Registering layer: Tanh_182 for ONNX node: Tanh_182 [X] Registering tensor: 617 for ONNX tensor: 617 [X] Tanh_182 [Tanh] outputs: [617 -> (1, 64, 18, 32)[FLOAT]], [X] Parsing node: Sub_184 [Sub] [X] Searching for input: 990 [X] Searching for input: 613 [X] Sub_184 [Sub] inputs: [990 -> ()[FLOAT]], [613 -> (1, 64, 18, 32)[FLOAT]], [X] Registering layer: 990 for ONNX node: 990 [X] Registering layer: Sub_184 for ONNX node: Sub_184 [X] Registering tensor: 619 for ONNX tensor: 619 [X] Sub_184 [Sub] outputs: [619 -> (1, 64, 18, 32)[FLOAT]], [X] Parsing node: Mul_185 [Mul] [X] Searching for input: 619 [X] Searching for input: 608 [X] Mul_185 [Mul] inputs: [619 -> (1, 64, 18, 32)[FLOAT]], [608 -> (1, 64, 18, 32)[FLOAT]], [X] Registering layer: Mul_185 for ONNX node: Mul_185 [X] Registering tensor: 620 for ONNX tensor: 620 [X] Mul_185 [Mul] outputs: [620 -> (1, 64, 18, 32)[FLOAT]], [X] Parsing node: Mul_186 [Mul] [X] Searching for input: 613 [X] Searching for input: 617 [X] Mul_186 [Mul] inputs: [613 -> (1, 64, 18, 32)[FLOAT]], [617 -> (1, 64, 18, 32)[FLOAT]], [X] Registering layer: Mul_186 for ONNX node: Mul_186 [X] Registering tensor: 621 for ONNX tensor: 621 [X] Mul_186 [Mul] outputs: [621 -> (1, 64, 18, 32)[FLOAT]], [X] Parsing node: Add_187 [Add] [X] Searching for input: 620 [X] Searching for input: 621 [X] Add_187 [Add] inputs: [620 -> (1, 64, 18, 32)[FLOAT]], [621 -> (1, 64, 18, 32)[FLOAT]], [X] Registering layer: Add_187 for ONNX node: Add_187 [X] Registering tensor: r4o_3 for ONNX tensor: r4o [X] Add_187 [Add] outputs: [r4o -> (1, 64, 18, 32)[FLOAT]], [X] Parsing node: Concat_188 [Concat] [X] Searching for input: 605 [X] Searching for input: r4o [X] Concat_188 [Concat] inputs: [605 -> (1, 64, 18, 32)[FLOAT]], [r4o -> (1, 64, 18, 32)[FLOAT]], [X] Registering layer: Concat_188 for ONNX node: Concat_188 [X] Registering tensor: 623 for ONNX tensor: 623 [X] Concat_188 [Concat] outputs: [623 -> (1, 128, 18, 32)[FLOAT]], [X] Parsing node: Resize_190 [Resize] [X] Searching for input: 623 [X] Searching for input: 386 [X] Searching for input: 985 [X] Resize_190 [Resize] inputs: [623 -> (1, 128, 18, 32)[FLOAT]], [386 -> (0)[FLOAT]], [985 -> (4)[FLOAT]], [X] Registering layer: Resize_190 for ONNX node: Resize_190 [X] Running resize layer with: Transformation mode: pytorch_half_pixel Resize mode: linear [X] Registering tensor: 628 for ONNX tensor: 628 [X] Resize_190 [Resize] outputs: [628 -> (1, 128, 36, 64)[FLOAT]], [X] Parsing node: Shape_191 [Shape] [X] Searching for input: 604 [X] Shape_191 [Shape] inputs: [604 -> (1, 3, 36, 64)[FLOAT]], [X] Registering layer: Shape_191 for ONNX node: Shape_191 [X] Registering tensor: 629 for ONNX tensor: 629 [X] Shape_191 [Shape] outputs: [629 -> (4)[INT32]], [X] Parsing node: Slice_195 [Slice] [X] Searching for input: 629 [X] Searching for input: 630 [X] Searching for input: 631 [X] Searching for input: 632 [X] Slice_195 [Slice] inputs: [629 -> (4)[INT32]], [630 -> (1)[INT32]], [631 -> (1)[INT32]], [632 -> (1)[INT32]], [X] Registering layer: Slice_195 for ONNX node: Slice_195 [X] Registering tensor: 633 for ONNX tensor: 633 [X] Slice_195 [Slice] outputs: [633 -> (2)[INT32]], [X] Parsing node: Slice_198 [Slice] [X] Searching for input: 628 [X] Searching for input: 634 [X] Searching for input: 633 [X] Searching for input: 635 [X] Slice_198 [Slice] inputs: [628 -> (1, 128, 36, 64)[FLOAT]], [634 -> (2)[INT32]], [633 -> (2)[INT32]], [635 -> (2)[INT32]], [X] Registering layer: Slice_198 for ONNX node: Slice_198 [X] Registering tensor: 636 for ONNX tensor: 636 [X] Slice_198 [Slice] outputs: [636 -> (1, 128, 36, 64)[FLOAT]], [X] Parsing node: Concat_199 [Concat] [X] Searching for input: 636 [X] Searching for input: 464 [X] Searching for input: 604 [X] Concat_199 [Concat] inputs: [636 -> (1, 128, 36, 64)[FLOAT]], [464 -> (1, 40, 36, 64)[FLOAT]], [604 -> (1, 3, 36, 64)[FLOAT]], [X] Registering layer: Concat_199 for ONNX node: Concat_199 [X] Registering tensor: 637 for ONNX tensor: 637 [X] Concat_199 [Concat] outputs: [637 -> (1, 171, 36, 64)[FLOAT]], [X] Parsing node: Conv_200 [Conv] [X] Searching for input: 637 [X] Searching for input: 965 [X] Searching for input: 966 [X] Conv_200 [Conv] inputs: [637 -> (1, 171, 36, 64)[FLOAT]], [965 -> (80, 171, 3, 3)[FLOAT]], [966 -> (80)[FLOAT]], [X] Convolution input dimensions: (1, 171, 36, 64) [X] Registering layer: Conv_200 for ONNX node: Conv_200 [X] Using kernel: (3, 3), strides: (1, 1), prepadding: (1, 1), postpadding: (1, 1), dilations: (1, 1), numOutputs: 80 [X] Convolution output dimensions: (1, 80, 36, 64) [X] Registering tensor: 964 for ONNX tensor: 964 [X] Conv_200 [Conv] outputs: [964 -> (1, 80, 36, 64)[FLOAT]], [X] Parsing node: Relu_201 [Relu] [X] Searching for input: 964 [X] Relu_201 [Relu] inputs: [964 -> (1, 80, 36, 64)[FLOAT]], [X] Registering layer: Relu_201 for ONNX node: Relu_201 [X] Registering tensor: 640 for ONNX tensor: 640 [X] Relu_201 [Relu] outputs: [640 -> (1, 80, 36, 64)[FLOAT]], [X] Parsing node: Split_202 [Split] [X] Searching for input: 640 [X] Split_202 [Split] inputs: [640 -> (1, 80, 36, 64)[FLOAT]], [X] Registering layer: Split_202 for ONNX node: Split_202 [X] Registering layer: Split_202_10 for ONNX node: Split_202 [X] Registering tensor: 641 for ONNX tensor: 641 [X] Registering tensor: 642 for ONNX tensor: 642 [X] Split_202 [Split] outputs: [641 -> (1, 40, 36, 64)[FLOAT]], [642 -> (1, 40, 36, 64)[FLOAT]], [X] Parsing node: Shape_203 [Shape] [X] Searching for input: 642 [X] Shape_203 [Shape] inputs: [642 -> (1, 40, 36, 64)[FLOAT]], [X] Registering layer: Shape_203 for ONNX node: Shape_203 [X] Registering tensor: 643 for ONNX tensor: 643 [X] Shape_203 [Shape] outputs: [643 -> (4)[INT32]], [X] Parsing node: Expand_204 [Expand] [X] Searching for input: r3i [X] Searching for input: 643 [X] Expand_204 [Expand] inputs: [r3i -> (1, 1, 1, 1)[FLOAT]], [643 -> (4)[INT32]], [X] Registering layer: Expand_204 for ONNX node: Expand_204 [X] Registering tensor: 644 for ONNX tensor: 644 [X] Expand_204 [Expand] outputs: [644 -> (1, 40, 36, 64)[FLOAT]], [X] Parsing node: Concat_205 [Concat] [X] Searching for input: 642 [X] Searching for input: 644 [X] Concat_205 [Concat] inputs: [642 -> (1, 40, 36, 64)[FLOAT]], [644 -> (1, 40, 36, 64)[FLOAT]], [X] Registering layer: Concat_205 for ONNX node: Concat_205 [X] Registering tensor: 645 for ONNX tensor: 645 [X] Concat_205 [Concat] outputs: [645 -> (1, 80, 36, 64)[FLOAT]], [X] Parsing node: Conv_206 [Conv] [X] Searching for input: 645 [X] Searching for input: decoder.decode3.gru.ih.0.weight [X] Searching for input: decoder.decode3.gru.ih.0.bias [X] Conv_206 [Conv] inputs: [645 -> (1, 80, 36, 64)[FLOAT]], [decoder.decode3.gru.ih.0.weight -> (80, 80, 3, 3)[FLOAT]], [decoder.decode3.gru.ih.0.bias -> (80)[FLOAT]], [X] Convolution input dimensions: (1, 80, 36, 64) [X] Registering layer: Conv_206 for ONNX node: Conv_206 [X] Using kernel: (3, 3), strides: (1, 1), prepadding: (1, 1), postpadding: (1, 1), dilations: (1, 1), numOutputs: 80 [X] Convolution output dimensions: (1, 80, 36, 64) [X] Registering tensor: 646 for ONNX tensor: 646 [X] Conv_206 [Conv] outputs: [646 -> (1, 80, 36, 64)[FLOAT]], [X] Parsing node: Sigmoid_207 [Sigmoid] [X] Searching for input: 646 [X] Sigmoid_207 [Sigmoid] inputs: [646 -> (1, 80, 36, 64)[FLOAT]], [X] Registering layer: Sigmoid_207 for ONNX node: Sigmoid_207 [X] Registering tensor: 647 for ONNX tensor: 647 [X] Sigmoid_207 [Sigmoid] outputs: [647 -> (1, 80, 36, 64)[FLOAT]], [X] Parsing node: Split_208 [Split] [X] Searching for input: 647 [X] Split_208 [Split] inputs: [647 -> (1, 80, 36, 64)[FLOAT]], [X] Registering layer: Split_208 for ONNX node: Split_208 [X] Registering layer: Split_208_13 for ONNX node: Split_208 [X] Registering tensor: 648 for ONNX tensor: 648 [X] Registering tensor: 649 for ONNX tensor: 649 [X] Split_208 [Split] outputs: [648 -> (1, 40, 36, 64)[FLOAT]], [649 -> (1, 40, 36, 64)[FLOAT]], [X] Parsing node: Mul_209 [Mul] [X] Searching for input: 648 [X] Searching for input: 644 [X] Mul_209 [Mul] inputs: [648 -> (1, 40, 36, 64)[FLOAT]], [644 -> (1, 40, 36, 64)[FLOAT]], [X] Registering layer: Mul_209 for ONNX node: Mul_209 [X] Registering tensor: 650 for ONNX tensor: 650 [X] Mul_209 [Mul] outputs: [650 -> (1, 40, 36, 64)[FLOAT]], [X] Parsing node: Concat_210 [Concat] [X] Searching for input: 642 [X] Searching for input: 650 [X] Concat_210 [Concat] inputs: [642 -> (1, 40, 36, 64)[FLOAT]], [650 -> (1, 40, 36, 64)[FLOAT]], [X] Registering layer: Concat_210 for ONNX node: Concat_210 [X] Registering tensor: 651 for ONNX tensor: 651 [X] Concat_210 [Concat] outputs: [651 -> (1, 80, 36, 64)[FLOAT]], [X] Parsing node: Conv_211 [Conv] [X] Searching for input: 651 [X] Searching for input: decoder.decode3.gru.hh.0.weight [X] Searching for input: decoder.decode3.gru.hh.0.bias [X] Conv_211 [Conv] inputs: [651 -> (1, 80, 36, 64)[FLOAT]], [decoder.decode3.gru.hh.0.weight -> (40, 80, 3, 3)[FLOAT]], [decoder.decode3.gru.hh.0.bias -> (40)[FLOAT]], [X] Convolution input dimensions: (1, 80, 36, 64) [X] Registering layer: Conv_211 for ONNX node: Conv_211 [X] Using kernel: (3, 3), strides: (1, 1), prepadding: (1, 1), postpadding: (1, 1), dilations: (1, 1), numOutputs: 40 [X] Convolution output dimensions: (1, 40, 36, 64) [X] Registering tensor: 652 for ONNX tensor: 652 [X] Conv_211 [Conv] outputs: [652 -> (1, 40, 36, 64)[FLOAT]], [X] Parsing node: Tanh_212 [Tanh] [X] Searching for input: 652 [X] Tanh_212 [Tanh] inputs: [652 -> (1, 40, 36, 64)[FLOAT]], [X] Registering layer: Tanh_212 for ONNX node: Tanh_212 [X] Registering tensor: 653 for ONNX tensor: 653 [X] Tanh_212 [Tanh] outputs: [653 -> (1, 40, 36, 64)[FLOAT]], [X] Parsing node: Sub_214 [Sub] [X] Searching for input: 990 [X] Searching for input: 649 [X] Sub_214 [Sub] inputs: [990 -> ()[FLOAT]], [649 -> (1, 40, 36, 64)[FLOAT]], [X] Registering layer: Sub_214 for ONNX node: Sub_214 [X] Registering tensor: 655 for ONNX tensor: 655 [X] Sub_214 [Sub] outputs: [655 -> (1, 40, 36, 64)[FLOAT]], [X] Parsing node: Mul_215 [Mul] [X] Searching for input: 655 [X] Searching for input: 644 [X] Mul_215 [Mul] inputs: [655 -> (1, 40, 36, 64)[FLOAT]], [644 -> (1, 40, 36, 64)[FLOAT]], [X] Registering layer: Mul_215 for ONNX node: Mul_215 [X] Registering tensor: 656 for ONNX tensor: 656 [X] Mul_215 [Mul] outputs: [656 -> (1, 40, 36, 64)[FLOAT]], [X] Parsing node: Mul_216 [Mul] [X] Searching for input: 649 [X] Searching for input: 653 [X] Mul_216 [Mul] inputs: [649 -> (1, 40, 36, 64)[FLOAT]], [653 -> (1, 40, 36, 64)[FLOAT]], [X] Registering layer: Mul_216 for ONNX node: Mul_216 [X] Registering tensor: 657 for ONNX tensor: 657 [X] Mul_216 [Mul] outputs: [657 -> (1, 40, 36, 64)[FLOAT]], [X] Parsing node: Add_217 [Add] [X] Searching for input: 656 [X] Searching for input: 657 [X] Add_217 [Add] inputs: [656 -> (1, 40, 36, 64)[FLOAT]], [657 -> (1, 40, 36, 64)[FLOAT]], [X] Registering layer: Add_217 for ONNX node: Add_217 [X] Registering tensor: r3o_14 for ONNX tensor: r3o [X] Add_217 [Add] outputs: [r3o -> (1, 40, 36, 64)[FLOAT]], [X] Parsing node: Concat_218 [Concat] [X] Searching for input: 641 [X] Searching for input: r3o [X] Concat_218 [Concat] inputs: [641 -> (1, 40, 36, 64)[FLOAT]], [r3o -> (1, 40, 36, 64)[FLOAT]], [X] Registering layer: Concat_218 for ONNX node: Concat_218 [X] Registering tensor: 659 for ONNX tensor: 659 [X] Concat_218 [Concat] outputs: [659 -> (1, 80, 36, 64)[FLOAT]], [X] Parsing node: Resize_220 [Resize] [X] Searching for input: 659 [X] Searching for input: 386 [X] Searching for input: 985 [X] Resize_220 [Resize] inputs: [659 -> (1, 80, 36, 64)[FLOAT]], [386 -> (0)[FLOAT]], [985 -> (4)[FLOAT]], [X] Registering layer: Resize_220 for ONNX node: Resize_220 [X] Running resize layer with: Transformation mode: pytorch_half_pixel Resize mode: linear [X] Registering tensor: 664 for ONNX tensor: 664 [X] Resize_220 [Resize] outputs: [664 -> (1, 80, 72, 128)[FLOAT]], [X] Parsing node: Shape_221 [Shape] [X] Searching for input: 603 [X] Shape_221 [Shape] inputs: [603 -> (1, 3, 72, 128)[FLOAT]], [X] Registering layer: Shape_221 for ONNX node: Shape_221 [X] Registering tensor: 665 for ONNX tensor: 665 [X] Shape_221 [Shape] outputs: [665 -> (4)[INT32]], [X] Parsing node: Slice_225 [Slice] [X] Searching for input: 665 [X] Searching for input: 630 [X] Searching for input: 631 [X] Searching for input: 632 [X] Slice_225 [Slice] inputs: [665 -> (4)[INT32]], [630 -> (1)[INT32]], [631 -> (1)[INT32]], [632 -> (1)[INT32]], [X] Registering layer: Slice_225 for ONNX node: Slice_225 [X] Registering tensor: 669 for ONNX tensor: 669 [X] Slice_225 [Slice] outputs: [669 -> (2)[INT32]], [X] Parsing node: Slice_228 [Slice] [X] Searching for input: 664 [X] Searching for input: 634 [X] Searching for input: 669 [X] Searching for input: 635 [X] Slice_228 [Slice] inputs: [664 -> (1, 80, 72, 128)[FLOAT]], [634 -> (2)[INT32]], [669 -> (2)[INT32]], [635 -> (2)[INT32]], [X] Registering layer: Slice_228 for ONNX node: Slice_228 [X] Registering tensor: 672 for ONNX tensor: 672 [X] Slice_228 [Slice] outputs: [672 -> (1, 80, 72, 128)[FLOAT]], [X] Parsing node: Concat_229 [Concat] [X] Searching for input: 672 [X] Searching for input: 420 [X] Searching for input: 603 [X] Concat_229 [Concat] inputs: [672 -> (1, 80, 72, 128)[FLOAT]], [420 -> (1, 24, 72, 128)[FLOAT]], [603 -> (1, 3, 72, 128)[FLOAT]], [X] Registering layer: Concat_229 for ONNX node: Concat_229 [X] Registering tensor: 673 for ONNX tensor: 673 [X] Concat_229 [Concat] outputs: [673 -> (1, 107, 72, 128)[FLOAT]], [X] Parsing node: Conv_230 [Conv] [X] Searching for input: 673 [X] Searching for input: 968 [X] Searching for input: 969 [X] Conv_230 [Conv] inputs: [673 -> (1, 107, 72, 128)[FLOAT]], [968 -> (40, 107, 3, 3)[FLOAT]], [969 -> (40)[FLOAT]], [X] Convolution input dimensions: (1, 107, 72, 128) [X] Registering layer: Conv_230 for ONNX node: Conv_230 [X] Using kernel: (3, 3), strides: (1, 1), prepadding: (1, 1), postpadding: (1, 1), dilations: (1, 1), numOutputs: 40 [X] Convolution output dimensions: (1, 40, 72, 128) [X] Registering tensor: 967 for ONNX tensor: 967 [X] Conv_230 [Conv] outputs: [967 -> (1, 40, 72, 128)[FLOAT]], [X] Parsing node: Relu_231 [Relu] [X] Searching for input: 967 [X] Relu_231 [Relu] inputs: [967 -> (1, 40, 72, 128)[FLOAT]], [X] Registering layer: Relu_231 for ONNX node: Relu_231 [X] Registering tensor: 676 for ONNX tensor: 676 [X] Relu_231 [Relu] outputs: [676 -> (1, 40, 72, 128)[FLOAT]], [X] Parsing node: Split_232 [Split] [X] Searching for input: 676 [X] Split_232 [Split] inputs: [676 -> (1, 40, 72, 128)[FLOAT]], [X] Registering layer: Split_232 for ONNX node: Split_232 [X] Registering layer: Split_232_21 for ONNX node: Split_232 [X] Registering tensor: 677 for ONNX tensor: 677 [X] Registering tensor: 678 for ONNX tensor: 678 [X] Split_232 [Split] outputs: [677 -> (1, 20, 72, 128)[FLOAT]], [678 -> (1, 20, 72, 128)[FLOAT]], [X] Parsing node: Shape_233 [Shape] [X] Searching for input: 678 [X] Shape_233 [Shape] inputs: [678 -> (1, 20, 72, 128)[FLOAT]], [X] Registering layer: Shape_233 for ONNX node: Shape_233 [X] Registering tensor: 679 for ONNX tensor: 679 [X] Shape_233 [Shape] outputs: [679 -> (4)[INT32]], [X] Parsing node: Expand_234 [Expand] [X] Searching for input: r2i [X] Searching for input: 679 [X] Expand_234 [Expand] inputs: [r2i -> (1, 1, 1, 1)[FLOAT]], [679 -> (4)[INT32]], [X] Registering layer: Expand_234 for ONNX node: Expand_234 [X] Registering tensor: 680 for ONNX tensor: 680 [X] Expand_234 [Expand] outputs: [680 -> (1, 20, 72, 128)[FLOAT]], [X] Parsing node: Concat_235 [Concat] [X] Searching for input: 678 [X] Searching for input: 680 [X] Concat_235 [Concat] inputs: [678 -> (1, 20, 72, 128)[FLOAT]], [680 -> (1, 20, 72, 128)[FLOAT]], [X] Registering layer: Concat_235 for ONNX node: Concat_235 [X] Registering tensor: 681 for ONNX tensor: 681 [X] Concat_235 [Concat] outputs: [681 -> (1, 40, 72, 128)[FLOAT]], [X] Parsing node: Conv_236 [Conv] [X] Searching for input: 681 [X] Searching for input: decoder.decode2.gru.ih.0.weight [X] Searching for input: decoder.decode2.gru.ih.0.bias [X] Conv_236 [Conv] inputs: [681 -> (1, 40, 72, 128)[FLOAT]], [decoder.decode2.gru.ih.0.weight -> (40, 40, 3, 3)[FLOAT]], [decoder.decode2.gru.ih.0.bias -> (40)[FLOAT]], [X] Convolution input dimensions: (1, 40, 72, 128) [X] Registering layer: Conv_236 for ONNX node: Conv_236 [X] Using kernel: (3, 3), strides: (1, 1), prepadding: (1, 1), postpadding: (1, 1), dilations: (1, 1), numOutputs: 40 [X] Convolution output dimensions: (1, 40, 72, 128) [X] Registering tensor: 682 for ONNX tensor: 682 [X] Conv_236 [Conv] outputs: [682 -> (1, 40, 72, 128)[FLOAT]], [X] Parsing node: Sigmoid_237 [Sigmoid] [X] Searching for input: 682 [X] Sigmoid_237 [Sigmoid] inputs: [682 -> (1, 40, 72, 128)[FLOAT]], [X] Registering layer: Sigmoid_237 for ONNX node: Sigmoid_237 [X] Registering tensor: 683 for ONNX tensor: 683 [X] Sigmoid_237 [Sigmoid] outputs: [683 -> (1, 40, 72, 128)[FLOAT]], [X] Parsing node: Split_238 [Split] [X] Searching for input: 683 [X] Split_238 [Split] inputs: [683 -> (1, 40, 72, 128)[FLOAT]], [X] Registering layer: Split_238 for ONNX node: Split_238 [X] Registering layer: Split_238_24 for ONNX node: Split_238 [X] Registering tensor: 684 for ONNX tensor: 684 [X] Registering tensor: 685 for ONNX tensor: 685 [X] Split_238 [Split] outputs: [684 -> (1, 20, 72, 128)[FLOAT]], [685 -> (1, 20, 72, 128)[FLOAT]], [X] Parsing node: Mul_239 [Mul] [X] Searching for input: 684 [X] Searching for input: 680 [X] Mul_239 [Mul] inputs: [684 -> (1, 20, 72, 128)[FLOAT]], [680 -> (1, 20, 72, 128)[FLOAT]], [X] Registering layer: Mul_239 for ONNX node: Mul_239 [X] Registering tensor: 686 for ONNX tensor: 686 [X] Mul_239 [Mul] outputs: [686 -> (1, 20, 72, 128)[FLOAT]], [X] Parsing node: Concat_240 [Concat] [X] Searching for input: 678 [X] Searching for input: 686 [X] Concat_240 [Concat] inputs: [678 -> (1, 20, 72, 128)[FLOAT]], [686 -> (1, 20, 72, 128)[FLOAT]], [X] Registering layer: Concat_240 for ONNX node: Concat_240 [X] Registering tensor: 687 for ONNX tensor: 687 [X] Concat_240 [Concat] outputs: [687 -> (1, 40, 72, 128)[FLOAT]], [X] Parsing node: Conv_241 [Conv] [X] Searching for input: 687 [X] Searching for input: decoder.decode2.gru.hh.0.weight [X] Searching for input: decoder.decode2.gru.hh.0.bias [X] Conv_241 [Conv] inputs: [687 -> (1, 40, 72, 128)[FLOAT]], [decoder.decode2.gru.hh.0.weight -> (20, 40, 3, 3)[FLOAT]], [decoder.decode2.gru.hh.0.bias -> (20)[FLOAT]], [X] Convolution input dimensions: (1, 40, 72, 128) [X] Registering layer: Conv_241 for ONNX node: Conv_241 [X] Using kernel: (3, 3), strides: (1, 1), prepadding: (1, 1), postpadding: (1, 1), dilations: (1, 1), numOutputs: 20 [X] Convolution output dimensions: (1, 20, 72, 128) [X] Registering tensor: 688 for ONNX tensor: 688 [X] Conv_241 [Conv] outputs: [688 -> (1, 20, 72, 128)[FLOAT]], [X] Parsing node: Tanh_242 [Tanh] [X] Searching for input: 688 [X] Tanh_242 [Tanh] inputs: [688 -> (1, 20, 72, 128)[FLOAT]], [X] Registering layer: Tanh_242 for ONNX node: Tanh_242 [X] Registering tensor: 689 for ONNX tensor: 689 [X] Tanh_242 [Tanh] outputs: [689 -> (1, 20, 72, 128)[FLOAT]], [X] Parsing node: Sub_244 [Sub] [X] Searching for input: 990 [X] Searching for input: 685 [X] Sub_244 [Sub] inputs: [990 -> ()[FLOAT]], [685 -> (1, 20, 72, 128)[FLOAT]], [X] Registering layer: Sub_244 for ONNX node: Sub_244 [X] Registering tensor: 691 for ONNX tensor: 691 [X] Sub_244 [Sub] outputs: [691 -> (1, 20, 72, 128)[FLOAT]], [X] Parsing node: Mul_245 [Mul] [X] Searching for input: 691 [X] Searching for input: 680 [X] Mul_245 [Mul] inputs: [691 -> (1, 20, 72, 128)[FLOAT]], [680 -> (1, 20, 72, 128)[FLOAT]], [X] Registering layer: Mul_245 for ONNX node: Mul_245 [X] Registering tensor: 692 for ONNX tensor: 692 [X] Mul_245 [Mul] outputs: [692 -> (1, 20, 72, 128)[FLOAT]], [X] Parsing node: Mul_246 [Mul] [X] Searching for input: 685 [X] Searching for input: 689 [X] Mul_246 [Mul] inputs: [685 -> (1, 20, 72, 128)[FLOAT]], [689 -> (1, 20, 72, 128)[FLOAT]], [X] Registering layer: Mul_246 for ONNX node: Mul_246 [X] Registering tensor: 693 for ONNX tensor: 693 [X] Mul_246 [Mul] outputs: [693 -> (1, 20, 72, 128)[FLOAT]], [X] Parsing node: Add_247 [Add] [X] Searching for input: 692 [X] Searching for input: 693 [X] Add_247 [Add] inputs: [692 -> (1, 20, 72, 128)[FLOAT]], [693 -> (1, 20, 72, 128)[FLOAT]], [X] Registering layer: Add_247 for ONNX node: Add_247 [X] Registering tensor: r2o_25 for ONNX tensor: r2o [X] Add_247 [Add] outputs: [r2o -> (1, 20, 72, 128)[FLOAT]], [X] Parsing node: Concat_248 [Concat] [X] Searching for input: 677 [X] Searching for input: r2o [X] Concat_248 [Concat] inputs: [677 -> (1, 20, 72, 128)[FLOAT]], [r2o -> (1, 20, 72, 128)[FLOAT]], [X] Registering layer: Concat_248 for ONNX node: Concat_248 [X] Registering tensor: 695 for ONNX tensor: 695 [X] Concat_248 [Concat] outputs: [695 -> (1, 40, 72, 128)[FLOAT]], [X] Parsing node: Resize_250 [Resize] [X] Searching for input: 695 [X] Searching for input: 386 [X] Searching for input: 985 [X] Resize_250 [Resize] inputs: [695 -> (1, 40, 72, 128)[FLOAT]], [386 -> (0)[FLOAT]], [985 -> (4)[FLOAT]], [X] Registering layer: Resize_250 for ONNX node: Resize_250 [X] Running resize layer with: Transformation mode: pytorch_half_pixel Resize mode: linear [X] Registering tensor: 700 for ONNX tensor: 700 [X] Resize_250 [Resize] outputs: [700 -> (1, 40, 144, 256)[FLOAT]], [X] Parsing node: Shape_251 [Shape] [X] Searching for input: 602 [X] Shape_251 [Shape] inputs: [602 -> (1, 3, 144, 256)[FLOAT]], [X] Registering layer: Shape_251 for ONNX node: Shape_251 [X] Registering tensor: 701 for ONNX tensor: 701 [X] Shape_251 [Shape] outputs: [701 -> (4)[INT32]], [X] Parsing node: Slice_255 [Slice] [X] Searching for input: 701 [X] Searching for input: 630 [X] Searching for input: 631 [X] Searching for input: 632 [X] Slice_255 [Slice] inputs: [701 -> (4)[INT32]], [630 -> (1)[INT32]], [631 -> (1)[INT32]], [632 -> (1)[INT32]], [X] Registering layer: Slice_255 for ONNX node: Slice_255 [X] Registering tensor: 705 for ONNX tensor: 705 [X] Slice_255 [Slice] outputs: [705 -> (2)[INT32]], [X] Parsing node: Slice_258 [Slice] [X] Searching for input: 700 [X] Searching for input: 634 [X] Searching for input: 705 [X] Searching for input: 635 [X] Slice_258 [Slice] inputs: [700 -> (1, 40, 144, 256)[FLOAT]], [634 -> (2)[INT32]], [705 -> (2)[INT32]], [635 -> (2)[INT32]], [X] Registering layer: Slice_258 for ONNX node: Slice_258 [X] Registering tensor: 708 for ONNX tensor: 708 [X] Slice_258 [Slice] outputs: [708 -> (1, 40, 144, 256)[FLOAT]], [X] Parsing node: Concat_259 [Concat] [X] Searching for input: 708 [X] Searching for input: 403 [X] Searching for input: 602 [X] Concat_259 [Concat] inputs: [708 -> (1, 40, 144, 256)[FLOAT]], [403 -> (1, 16, 144, 256)[FLOAT]], [602 -> (1, 3, 144, 256)[FLOAT]], [X] Registering layer: Concat_259 for ONNX node: Concat_259 [X] Registering tensor: 709 for ONNX tensor: 709 [X] Concat_259 [Concat] outputs: [709 -> (1, 59, 144, 256)[FLOAT]], [X] Parsing node: Conv_260 [Conv] [X] Searching for input: 709 [X] Searching for input: 971 [X] Searching for input: 972 [X] Conv_260 [Conv] inputs: [709 -> (1, 59, 144, 256)[FLOAT]], [971 -> (32, 59, 3, 3)[FLOAT]], [972 -> (32)[FLOAT]], [X] Convolution input dimensions: (1, 59, 144, 256) [X] Registering layer: Conv_260 for ONNX node: Conv_260 [X] Using kernel: (3, 3), strides: (1, 1), prepadding: (1, 1), postpadding: (1, 1), dilations: (1, 1), numOutputs: 32 [X] Convolution output dimensions: (1, 32, 144, 256) [X] Registering tensor: 970 for ONNX tensor: 970 [X] Conv_260 [Conv] outputs: [970 -> (1, 32, 144, 256)[FLOAT]], [X] Parsing node: Relu_261 [Relu] [X] Searching for input: 970 [X] Relu_261 [Relu] inputs: [970 -> (1, 32, 144, 256)[FLOAT]], [X] Registering layer: Relu_261 for ONNX node: Relu_261 [X] Registering tensor: 712 for ONNX tensor: 712 [X] Relu_261 [Relu] outputs: [712 -> (1, 32, 144, 256)[FLOAT]], [X] Parsing node: Split_262 [Split] [X] Searching for input: 712 [X] Split_262 [Split] inputs: [712 -> (1, 32, 144, 256)[FLOAT]], [X] Registering layer: Split_262 for ONNX node: Split_262 [X] Registering layer: Split_262_32 for ONNX node: Split_262 [X] Registering tensor: 713 for ONNX tensor: 713 [X] Registering tensor: 714 for ONNX tensor: 714 [X] Split_262 [Split] outputs: [713 -> (1, 16, 144, 256)[FLOAT]], [714 -> (1, 16, 144, 256)[FLOAT]], [X] Parsing node: Shape_263 [Shape] [X] Searching for input: 714 [X] Shape_263 [Shape] inputs: [714 -> (1, 16, 144, 256)[FLOAT]], [X] Registering layer: Shape_263 for ONNX node: Shape_263 [X] Registering tensor: 715 for ONNX tensor: 715 [X] Shape_263 [Shape] outputs: [715 -> (4)[INT32]], [X] Parsing node: Expand_264 [Expand] [X] Searching for input: r1i [X] Searching for input: 715 [X] Expand_264 [Expand] inputs: [r1i -> (1, 1, 1, 1)[FLOAT]], [715 -> (4)[INT32]], [X] Registering layer: Expand_264 for ONNX node: Expand_264 [X] Registering tensor: 716 for ONNX tensor: 716 [X] Expand_264 [Expand] outputs: [716 -> (1, 16, 144, 256)[FLOAT]], [X] Parsing node: Concat_265 [Concat] [X] Searching for input: 714 [X] Searching for input: 716 [X] Concat_265 [Concat] inputs: [714 -> (1, 16, 144, 256)[FLOAT]], [716 -> (1, 16, 144, 256)[FLOAT]], [X] Registering layer: Concat_265 for ONNX node: Concat_265 [X] Registering tensor: 717 for ONNX tensor: 717 [X] Concat_265 [Concat] outputs: [717 -> (1, 32, 144, 256)[FLOAT]], [X] Parsing node: Conv_266 [Conv] [X] Searching for input: 717 [X] Searching for input: decoder.decode1.gru.ih.0.weight [X] Searching for input: decoder.decode1.gru.ih.0.bias [X] Conv_266 [Conv] inputs: [717 -> (1, 32, 144, 256)[FLOAT]], [decoder.decode1.gru.ih.0.weight -> (32, 32, 3, 3)[FLOAT]], [decoder.decode1.gru.ih.0.bias -> (32)[FLOAT]], [X] Convolution input dimensions: (1, 32, 144, 256) [X] Registering layer: Conv_266 for ONNX node: Conv_266 [X] Using kernel: (3, 3), strides: (1, 1), prepadding: (1, 1), postpadding: (1, 1), dilations: (1, 1), numOutputs: 32 [X] Convolution output dimensions: (1, 32, 144, 256) [X] Registering tensor: 718 for ONNX tensor: 718 [X] Conv_266 [Conv] outputs: [718 -> (1, 32, 144, 256)[FLOAT]], [X] Parsing node: Sigmoid_267 [Sigmoid] [X] Searching for input: 718 [X] Sigmoid_267 [Sigmoid] inputs: [718 -> (1, 32, 144, 256)[FLOAT]], [X] Registering layer: Sigmoid_267 for ONNX node: Sigmoid_267 [X] Registering tensor: 719 for ONNX tensor: 719 [X] Sigmoid_267 [Sigmoid] outputs: [719 -> (1, 32, 144, 256)[FLOAT]], [X] Parsing node: Split_268 [Split] [X] Searching for input: 719 [X] Split_268 [Split] inputs: [719 -> (1, 32, 144, 256)[FLOAT]], [X] Registering layer: Split_268 for ONNX node: Split_268 [X] Registering layer: Split_268_35 for ONNX node: Split_268 [X] Registering tensor: 720 for ONNX tensor: 720 [X] Registering tensor: 721 for ONNX tensor: 721 [X] Split_268 [Split] outputs: [720 -> (1, 16, 144, 256)[FLOAT]], [721 -> (1, 16, 144, 256)[FLOAT]], [X] Parsing node: Mul_269 [Mul] [X] Searching for input: 720 [X] Searching for input: 716 [X] Mul_269 [Mul] inputs: [720 -> (1, 16, 144, 256)[FLOAT]], [716 -> (1, 16, 144, 256)[FLOAT]], [X] Registering layer: Mul_269 for ONNX node: Mul_269 [X] Registering tensor: 722 for ONNX tensor: 722 [X] Mul_269 [Mul] outputs: [722 -> (1, 16, 144, 256)[FLOAT]], [X] Parsing node: Concat_270 [Concat] [X] Searching for input: 714 [X] Searching for input: 722 [X] Concat_270 [Concat] inputs: [714 -> (1, 16, 144, 256)[FLOAT]], [722 -> (1, 16, 144, 256)[FLOAT]], [X] Registering layer: Concat_270 for ONNX node: Concat_270 [X] Registering tensor: 723 for ONNX tensor: 723 [X] Concat_270 [Concat] outputs: [723 -> (1, 32, 144, 256)[FLOAT]], [X] Parsing node: Conv_271 [Conv] [X] Searching for input: 723 [X] Searching for input: decoder.decode1.gru.hh.0.weight [X] Searching for input: decoder.decode1.gru.hh.0.bias [X] Conv_271 [Conv] inputs: [723 -> (1, 32, 144, 256)[FLOAT]], [decoder.decode1.gru.hh.0.weight -> (16, 32, 3, 3)[FLOAT]], [decoder.decode1.gru.hh.0.bias -> (16)[FLOAT]], [X] Convolution input dimensions: (1, 32, 144, 256) [X] Registering layer: Conv_271 for ONNX node: Conv_271 [X] Using kernel: (3, 3), strides: (1, 1), prepadding: (1, 1), postpadding: (1, 1), dilations: (1, 1), numOutputs: 16 [X] Convolution output dimensions: (1, 16, 144, 256) [X] Registering tensor: 724 for ONNX tensor: 724 [X] Conv_271 [Conv] outputs: [724 -> (1, 16, 144, 256)[FLOAT]], [X] Parsing node: Tanh_272 [Tanh] [X] Searching for input: 724 [X] Tanh_272 [Tanh] inputs: [724 -> (1, 16, 144, 256)[FLOAT]], [X] Registering layer: Tanh_272 for ONNX node: Tanh_272 [X] Registering tensor: 725 for ONNX tensor: 725 [X] Tanh_272 [Tanh] outputs: [725 -> (1, 16, 144, 256)[FLOAT]], [X] Parsing node: Sub_274 [Sub] [X] Searching for input: 990 [X] Searching for input: 721 [X] Sub_274 [Sub] inputs: [990 -> ()[FLOAT]], [721 -> (1, 16, 144, 256)[FLOAT]], [X] Registering layer: Sub_274 for ONNX node: Sub_274 [X] Registering tensor: 727 for ONNX tensor: 727 [X] Sub_274 [Sub] outputs: [727 -> (1, 16, 144, 256)[FLOAT]], [X] Parsing node: Mul_275 [Mul] [X] Searching for input: 727 [X] Searching for input: 716 [X] Mul_275 [Mul] inputs: [727 -> (1, 16, 144, 256)[FLOAT]], [716 -> (1, 16, 144, 256)[FLOAT]], [X] Registering layer: Mul_275 for ONNX node: Mul_275 [X] Registering tensor: 728 for ONNX tensor: 728 [X] Mul_275 [Mul] outputs: [728 -> (1, 16, 144, 256)[FLOAT]], [X] Parsing node: Mul_276 [Mul] [X] Searching for input: 721 [X] Searching for input: 725 [X] Mul_276 [Mul] inputs: [721 -> (1, 16, 144, 256)[FLOAT]], [725 -> (1, 16, 144, 256)[FLOAT]], [X] Registering layer: Mul_276 for ONNX node: Mul_276 [X] Registering tensor: 729 for ONNX tensor: 729 [X] Mul_276 [Mul] outputs: [729 -> (1, 16, 144, 256)[FLOAT]], [X] Parsing node: Add_277 [Add] [X] Searching for input: 728 [X] Searching for input: 729 [X] Add_277 [Add] inputs: [728 -> (1, 16, 144, 256)[FLOAT]], [729 -> (1, 16, 144, 256)[FLOAT]], [X] Registering layer: Add_277 for ONNX node: Add_277 [X] Registering tensor: r1o_36 for ONNX tensor: r1o [X] Add_277 [Add] outputs: [r1o -> (1, 16, 144, 256)[FLOAT]], [X] Parsing node: Concat_278 [Concat] [X] Searching for input: 713 [X] Searching for input: r1o [X] Concat_278 [Concat] inputs: [713 -> (1, 16, 144, 256)[FLOAT]], [r1o -> (1, 16, 144, 256)[FLOAT]], [X] Registering layer: Concat_278 for ONNX node: Concat_278 [X] Registering tensor: 731 for ONNX tensor: 731 [X] Concat_278 [Concat] outputs: [731 -> (1, 32, 144, 256)[FLOAT]], [X] Parsing node: Resize_280 [Resize] [X] Searching for input: 731 [X] Searching for input: 386 [X] Searching for input: 985 [X] Resize_280 [Resize] inputs: [731 -> (1, 32, 144, 256)[FLOAT]], [386 -> (0)[FLOAT]], [985 -> (4)[FLOAT]], [X] Registering layer: Resize_280 for ONNX node: Resize_280 [X] Running resize layer with: Transformation mode: pytorch_half_pixel Resize mode: linear [X] Registering tensor: 736 for ONNX tensor: 736 [X] Resize_280 [Resize] outputs: [736 -> (1, 32, 288, 512)[FLOAT]], [X] Parsing node: Shape_281 [Shape] [X] Searching for input: 389 [X] Shape_281 [Shape] inputs: [389 -> (1, 3, 288, 512)[FLOAT]], [X] Registering layer: Shape_281 for ONNX node: Shape_281 [X] Registering tensor: 737 for ONNX tensor: 737 [X] Shape_281 [Shape] outputs: [737 -> (4)[INT32]], [X] Parsing node: Slice_285 [Slice] [X] Searching for input: 737 [X] Searching for input: 630 [X] Searching for input: 631 [X] Searching for input: 632 [X] Slice_285 [Slice] inputs: [737 -> (4)[INT32]], [630 -> (1)[INT32]], [631 -> (1)[INT32]], [632 -> (1)[INT32]], [X] Registering layer: Slice_285 for ONNX node: Slice_285 [X] Registering tensor: 741 for ONNX tensor: 741 [X] Slice_285 [Slice] outputs: [741 -> (2)[INT32]], [X] Parsing node: Slice_288 [Slice] [X] Searching for input: 736 [X] Searching for input: 634 [X] Searching for input: 741 [X] Searching for input: 635 [X] Slice_288 [Slice] inputs: [736 -> (1, 32, 288, 512)[FLOAT]], [634 -> (2)[INT32]], [741 -> (2)[INT32]], [635 -> (2)[INT32]], [X] Registering layer: Slice_288 for ONNX node: Slice_288 [X] Registering tensor: 744 for ONNX tensor: 744 [X] Slice_288 [Slice] outputs: [744 -> (1, 32, 288, 512)[FLOAT]], [X] Parsing node: Concat_289 [Concat] [X] Searching for input: 744 [X] Searching for input: 389 [X] Concat_289 [Concat] inputs: [744 -> (1, 32, 288, 512)[FLOAT]], [389 -> (1, 3, 288, 512)[FLOAT]], [X] Registering layer: Concat_289 for ONNX node: Concat_289 [X] Registering tensor: 745 for ONNX tensor: 745 [X] Concat_289 [Concat] outputs: [745 -> (1, 35, 288, 512)[FLOAT]], [X] Parsing node: Conv_290 [Conv] [X] Searching for input: 745 [X] Searching for input: 974 [X] Searching for input: 975 [X] Conv_290 [Conv] inputs: [745 -> (1, 35, 288, 512)[FLOAT]], [974 -> (16, 35, 3, 3)[FLOAT]], [975 -> (16)[FLOAT]], [X] Convolution input dimensions: (1, 35, 288, 512) [X] Registering layer: Conv_290 for ONNX node: Conv_290 [X] Using kernel: (3, 3), strides: (1, 1), prepadding: (1, 1), postpadding: (1, 1), dilations: (1, 1), numOutputs: 16 [X] Convolution output dimensions: (1, 16, 288, 512) [X] Registering tensor: 973 for ONNX tensor: 973 [X] Conv_290 [Conv] outputs: [973 -> (1, 16, 288, 512)[FLOAT]], [X] Parsing node: Relu_291 [Relu] [X] Searching for input: 973 [X] Relu_291 [Relu] inputs: [973 -> (1, 16, 288, 512)[FLOAT]], [X] Registering layer: Relu_291 for ONNX node: Relu_291 [X] Registering tensor: 748 for ONNX tensor: 748 [X] Relu_291 [Relu] outputs: [748 -> (1, 16, 288, 512)[FLOAT]], [X] Parsing node: Conv_292 [Conv] [X] Searching for input: 748 [X] Searching for input: 977 [X] Searching for input: 978 [X] Conv_292 [Conv] inputs: [748 -> (1, 16, 288, 512)[FLOAT]], [977 -> (16, 16, 3, 3)[FLOAT]], [978 -> (16)[FLOAT]], [X] Convolution input dimensions: (1, 16, 288, 512) [X] Registering layer: Conv_292 for ONNX node: Conv_292 [X] Using kernel: (3, 3), strides: (1, 1), prepadding: (1, 1), postpadding: (1, 1), dilations: (1, 1), numOutputs: 16 [X] Convolution output dimensions: (1, 16, 288, 512) [X] Registering tensor: 976 for ONNX tensor: 976 [X] Conv_292 [Conv] outputs: [976 -> (1, 16, 288, 512)[FLOAT]], [X] Parsing node: Relu_293 [Relu] [X] Searching for input: 976 [X] Relu_293 [Relu] inputs: [976 -> (1, 16, 288, 512)[FLOAT]], [X] Registering layer: Relu_293 for ONNX node: Relu_293 [X] Registering tensor: 751 for ONNX tensor: 751 [X] Relu_293 [Relu] outputs: [751 -> (1, 16, 288, 512)[FLOAT]], [X] Parsing node: Conv_294 [Conv] [X] Searching for input: 751 [X] Searching for input: project_mat.conv.weight [X] Searching for input: project_mat.conv.bias [X] Conv_294 [Conv] inputs: [751 -> (1, 16, 288, 512)[FLOAT]], [project_mat.conv.weight -> (4, 16, 1, 1)[FLOAT]], [project_mat.conv.bias -> (4)[FLOAT]], [X] Convolution input dimensions: (1, 16, 288, 512) [X] Registering layer: Conv_294 for ONNX node: Conv_294 [X] Using kernel: (1, 1), strides: (1, 1), prepadding: (0, 0), postpadding: (0, 0), dilations: (1, 1), numOutputs: 4 [X] Convolution output dimensions: (1, 4, 288, 512) [X] Registering tensor: 752 for ONNX tensor: 752 [X] Conv_294 [Conv] outputs: [752 -> (1, 4, 288, 512)[FLOAT]], [X] Parsing node: Split_295 [Split] [X] Searching for input: 752 [X] Split_295 [Split] inputs: [752 -> (1, 4, 288, 512)[FLOAT]], [X] Registering layer: Split_295 for ONNX node: Split_295 [X] Registering layer: Split_295_43 for ONNX node: Split_295 [X] Registering tensor: 753 for ONNX tensor: 753 [X] Registering tensor: 754 for ONNX tensor: 754 [X] Split_295 [Split] outputs: [753 -> (1, 3, 288, 512)[FLOAT]], [754 -> (1, 1, 288, 512)[FLOAT]], [X] Parsing node: ReduceMean_296 [ReduceMean] [X] Searching for input: src [X] ReduceMean_296 [ReduceMean] inputs: [src -> (1, 3, 1440, 2560)[FLOAT]], [X] Registering layer: ReduceMean_296 for ONNX node: ReduceMean_296 [X] Registering tensor: 755 for ONNX tensor: 755 [X] ReduceMean_296 [ReduceMean] outputs: [755 -> (1, 1, 1440, 2560)[FLOAT]], [X] Parsing node: Concat_297 [Concat] [X] Searching for input: src [X] Searching for input: 755 [X] Concat_297 [Concat] inputs: [src -> (1, 3, 1440, 2560)[FLOAT]], [755 -> (1, 1, 1440, 2560)[FLOAT]], [X] Registering layer: Concat_297 for ONNX node: Concat_297 [X] Registering tensor: 756 for ONNX tensor: 756 [X] Concat_297 [Concat] outputs: [756 -> (1, 4, 1440, 2560)[FLOAT]], [X] Parsing node: ReduceMean_298 [ReduceMean] [X] Searching for input: 389 [X] ReduceMean_298 [ReduceMean] inputs: [389 -> (1, 3, 288, 512)[FLOAT]], [X] Registering layer: ReduceMean_298 for ONNX node: ReduceMean_298 [X] Registering tensor: 757 for ONNX tensor: 757 [X] ReduceMean_298 [ReduceMean] outputs: [757 -> (1, 1, 288, 512)[FLOAT]], [X] Parsing node: Concat_299 [Concat] [X] Searching for input: 389 [X] Searching for input: 757 [X] Concat_299 [Concat] inputs: [389 -> (1, 3, 288, 512)[FLOAT]], [757 -> (1, 1, 288, 512)[FLOAT]], [X] Registering layer: Concat_299 for ONNX node: Concat_299 [X] Registering tensor: 758 for ONNX tensor: 758 [X] Concat_299 [Concat] outputs: [758 -> (1, 4, 288, 512)[FLOAT]], [X] Parsing node: Concat_300 [Concat] [X] Searching for input: 753 [X] Searching for input: 754 [X] Concat_300 [Concat] inputs: [753 -> (1, 3, 288, 512)[FLOAT]], [754 -> (1, 1, 288, 512)[FLOAT]], [X] Registering layer: Concat_300 for ONNX node: Concat_300 [X] Registering tensor: 759 for ONNX tensor: 759 [X] Concat_300 [Concat] outputs: [759 -> (1, 4, 288, 512)[FLOAT]], [X] Parsing node: Conv_301 [Conv] [X] Searching for input: 758 [X] Searching for input: refiner.box_filter.weight [X] Conv_301 [Conv] inputs: [758 -> (1, 4, 288, 512)[FLOAT]], [refiner.box_filter.weight -> (4, 1, 3, 3)[FLOAT]], [X] Convolution input dimensions: (1, 4, 288, 512) [X] Registering layer: Conv_301 for ONNX node: Conv_301 [X] Using kernel: (3, 3), strides: (1, 1), prepadding: (1, 1), postpadding: (1, 1), dilations: (1, 1), numOutputs: 4 [X] Convolution output dimensions: (1, 4, 288, 512) [X] Registering tensor: 760 for ONNX tensor: 760 [X] Conv_301 [Conv] outputs: [760 -> (1, 4, 288, 512)[FLOAT]], [X] Parsing node: Conv_302 [Conv] [X] Searching for input: 759 [X] Searching for input: refiner.box_filter.weight [X] Conv_302 [Conv] inputs: [759 -> (1, 4, 288, 512)[FLOAT]], [refiner.box_filter.weight -> (4, 1, 3, 3)[FLOAT]], [X] Convolution input dimensions: (1, 4, 288, 512) [X] Registering layer: Conv_302 for ONNX node: Conv_302 [X] Using kernel: (3, 3), strides: (1, 1), prepadding: (1, 1), postpadding: (1, 1), dilations: (1, 1), numOutputs: 4 [X] Convolution output dimensions: (1, 4, 288, 512) [X] Registering tensor: 761 for ONNX tensor: 761 [X] Conv_302 [Conv] outputs: [761 -> (1, 4, 288, 512)[FLOAT]], [X] Parsing node: Mul_303 [Mul] [X] Searching for input: 758 [X] Searching for input: 759 [X] Mul_303 [Mul] inputs: [758 -> (1, 4, 288, 512)[FLOAT]], [759 -> (1, 4, 288, 512)[FLOAT]], [X] Registering layer: Mul_303 for ONNX node: Mul_303 [X] Registering tensor: 762 for ONNX tensor: 762 [X] Mul_303 [Mul] outputs: [762 -> (1, 4, 288, 512)[FLOAT]], [X] Parsing node: Conv_304 [Conv] [X] Searching for input: 762 [X] Searching for input: refiner.box_filter.weight [X] Conv_304 [Conv] inputs: [762 -> (1, 4, 288, 512)[FLOAT]], [refiner.box_filter.weight -> (4, 1, 3, 3)[FLOAT]], [X] Convolution input dimensions: (1, 4, 288, 512) [X] Registering layer: Conv_304 for ONNX node: Conv_304 [X] Using kernel: (3, 3), strides: (1, 1), prepadding: (1, 1), postpadding: (1, 1), dilations: (1, 1), numOutputs: 4 [X] Convolution output dimensions: (1, 4, 288, 512) [X] Registering tensor: 763 for ONNX tensor: 763 [X] Conv_304 [Conv] outputs: [763 -> (1, 4, 288, 512)[FLOAT]], [X] Parsing node: Mul_305 [Mul] [X] Searching for input: 760 [X] Searching for input: 761 [X] Mul_305 [Mul] inputs: [760 -> (1, 4, 288, 512)[FLOAT]], [761 -> (1, 4, 288, 512)[FLOAT]], [X] Registering layer: Mul_305 for ONNX node: Mul_305 [X] Registering tensor: 764 for ONNX tensor: 764 [X] Mul_305 [Mul] outputs: [764 -> (1, 4, 288, 512)[FLOAT]], [X] Parsing node: Sub_306 [Sub] [X] Searching for input: 763 [X] Searching for input: 764 [X] Sub_306 [Sub] inputs: [763 -> (1, 4, 288, 512)[FLOAT]], [764 -> (1, 4, 288, 512)[FLOAT]], [X] Registering layer: Sub_306 for ONNX node: Sub_306 [X] Registering tensor: 765 for ONNX tensor: 765 [X] Sub_306 [Sub] outputs: [765 -> (1, 4, 288, 512)[FLOAT]], [X] Parsing node: Mul_307 [Mul] [X] Searching for input: 758 [X] Searching for input: 758 [X] Mul_307 [Mul] inputs: [758 -> (1, 4, 288, 512)[FLOAT]], [758 -> (1, 4, 288, 512)[FLOAT]], [X] Registering layer: Mul_307 for ONNX node: Mul_307 [X] Registering tensor: 766 for ONNX tensor: 766 [X] Mul_307 [Mul] outputs: [766 -> (1, 4, 288, 512)[FLOAT]], [X] Parsing node: Conv_308 [Conv] [X] Searching for input: 766 [X] Searching for input: refiner.box_filter.weight [X] Conv_308 [Conv] inputs: [766 -> (1, 4, 288, 512)[FLOAT]], [refiner.box_filter.weight -> (4, 1, 3, 3)[FLOAT]], [X] Convolution input dimensions: (1, 4, 288, 512) [X] Registering layer: Conv_308 for ONNX node: Conv_308 [X] Using kernel: (3, 3), strides: (1, 1), prepadding: (1, 1), postpadding: (1, 1), dilations: (1, 1), numOutputs: 4 [X] Convolution output dimensions: (1, 4, 288, 512) [X] Registering tensor: 767 for ONNX tensor: 767 [X] Conv_308 [Conv] outputs: [767 -> (1, 4, 288, 512)[FLOAT]], [X] Parsing node: Mul_309 [Mul] [X] Searching for input: 760 [X] Searching for input: 760 [X] Mul_309 [Mul] inputs: [760 -> (1, 4, 288, 512)[FLOAT]], [760 -> (1, 4, 288, 512)[FLOAT]], [X] Registering layer: Mul_309 for ONNX node: Mul_309 [X] Registering tensor: 768 for ONNX tensor: 768 [X] Mul_309 [Mul] outputs: [768 -> (1, 4, 288, 512)[FLOAT]], [X] Parsing node: Sub_310 [Sub] [X] Searching for input: 767 [X] Searching for input: 768 [X] Sub_310 [Sub] inputs: [767 -> (1, 4, 288, 512)[FLOAT]], [768 -> (1, 4, 288, 512)[FLOAT]], [X] Registering layer: Sub_310 for ONNX node: Sub_310 [X] Registering tensor: 769 for ONNX tensor: 769 [X] Sub_310 [Sub] outputs: [769 -> (1, 4, 288, 512)[FLOAT]], [X] Parsing node: Concat_311 [Concat] [X] Searching for input: 765 [X] Searching for input: 769 [X] Searching for input: 751 [X] Concat_311 [Concat] inputs: [765 -> (1, 4, 288, 512)[FLOAT]], [769 -> (1, 4, 288, 512)[FLOAT]], [751 -> (1, 16, 288, 512)[FLOAT]], [X] Registering layer: Concat_311 for ONNX node: Concat_311 [X] Registering tensor: 770 for ONNX tensor: 770 [X] Concat_311 [Concat] outputs: [770 -> (1, 24, 288, 512)[FLOAT]], [X] Parsing node: Conv_312 [Conv] [X] Searching for input: 770 [X] Searching for input: 980 [X] Searching for input: 981 [X] Conv_312 [Conv] inputs: [770 -> (1, 24, 288, 512)[FLOAT]], [980 -> (16, 24, 1, 1)[FLOAT]], [981 -> (16)[FLOAT]], [X] Convolution input dimensions: (1, 24, 288, 512) [X] Registering layer: Conv_312 for ONNX node: Conv_312 [X] Using kernel: (1, 1), strides: (1, 1), prepadding: (0, 0), postpadding: (0, 0), dilations: (1, 1), numOutputs: 16 [X] Convolution output dimensions: (1, 16, 288, 512) [X] Registering tensor: 979 for ONNX tensor: 979 [X] Conv_312 [Conv] outputs: [979 -> (1, 16, 288, 512)[FLOAT]], [X] Parsing node: Relu_313 [Relu] [X] Searching for input: 979 [X] Relu_313 [Relu] inputs: [979 -> (1, 16, 288, 512)[FLOAT]], [X] Registering layer: Relu_313 for ONNX node: Relu_313 [X] Registering tensor: 773 for ONNX tensor: 773 [X] Relu_313 [Relu] outputs: [773 -> (1, 16, 288, 512)[FLOAT]], [X] Parsing node: Conv_314 [Conv] [X] Searching for input: 773 [X] Searching for input: 983 [X] Searching for input: 984 [X] Conv_314 [Conv] inputs: [773 -> (1, 16, 288, 512)[FLOAT]], [983 -> (16, 16, 1, 1)[FLOAT]], [984 -> (16)[FLOAT]], [X] Convolution input dimensions: (1, 16, 288, 512) [X] Registering layer: Conv_314 for ONNX node: Conv_314 [X] Using kernel: (1, 1), strides: (1, 1), prepadding: (0, 0), postpadding: (0, 0), dilations: (1, 1), numOutputs: 16 [X] Convolution output dimensions: (1, 16, 288, 512) [X] Registering tensor: 982 for ONNX tensor: 982 [X] Conv_314 [Conv] outputs: [982 -> (1, 16, 288, 512)[FLOAT]], [X] Parsing node: Relu_315 [Relu] [X] Searching for input: 982 [X] Relu_315 [Relu] inputs: [982 -> (1, 16, 288, 512)[FLOAT]], [X] Registering layer: Relu_315 for ONNX node: Relu_315 [X] Registering tensor: 776 for ONNX tensor: 776 [X] Relu_315 [Relu] outputs: [776 -> (1, 16, 288, 512)[FLOAT]], [X] Parsing node: Conv_316 [Conv] [X] Searching for input: 776 [X] Searching for input: refiner.conv.6.weight [X] Searching for input: refiner.conv.6.bias [X] Conv_316 [Conv] inputs: [776 -> (1, 16, 288, 512)[FLOAT]], [refiner.conv.6.weight -> (4, 16, 1, 1)[FLOAT]], [refiner.conv.6.bias -> (4)[FLOAT]], [X] Convolution input dimensions: (1, 16, 288, 512) [X] Registering layer: Conv_316 for ONNX node: Conv_316 [X] Using kernel: (1, 1), strides: (1, 1), prepadding: (0, 0), postpadding: (0, 0), dilations: (1, 1), numOutputs: 4 [X] Convolution output dimensions: (1, 4, 288, 512) [X] Registering tensor: 777 for ONNX tensor: 777 [X] Conv_316 [Conv] outputs: [777 -> (1, 4, 288, 512)[FLOAT]], [X] Parsing node: Mul_317 [Mul] [X] Searching for input: 777 [X] Searching for input: 760 [X] Mul_317 [Mul] inputs: [777 -> (1, 4, 288, 512)[FLOAT]], [760 -> (1, 4, 288, 512)[FLOAT]], [X] Registering layer: Mul_317 for ONNX node: Mul_317 [X] Registering tensor: 778 for ONNX tensor: 778 [X] Mul_317 [Mul] outputs: [778 -> (1, 4, 288, 512)[FLOAT]], [X] Parsing node: Sub_318 [Sub] [X] Searching for input: 761 [X] Searching for input: 778 [X] Sub_318 [Sub] inputs: [761 -> (1, 4, 288, 512)[FLOAT]], [778 -> (1, 4, 288, 512)[FLOAT]], [X] Registering layer: Sub_318 for ONNX node: Sub_318 [X] Registering tensor: 779 for ONNX tensor: 779 [X] Sub_318 [Sub] outputs: [779 -> (1, 4, 288, 512)[FLOAT]], [X] Parsing node: Shape_321 [Shape] [X] Searching for input: 777 [X] Shape_321 [Shape] inputs: [777 -> (1, 4, 288, 512)[FLOAT]], [X] Registering layer: Shape_321 for ONNX node: Shape_321 [X] Registering tensor: 782 for ONNX tensor: 782 [X] Shape_321 [Shape] outputs: [782 -> (4)[INT32]], [X] Parsing node: Slice_325 [Slice] [X] Searching for input: 782 [X] Searching for input: 632 [X] Searching for input: 630 [X] Searching for input: 632 [X] Slice_325 [Slice] inputs: [782 -> (4)[INT32]], [632 -> (1)[INT32]], [630 -> (1)[INT32]], [632 -> (1)[INT32]], [X] Registering layer: Slice_325 for ONNX node: Slice_325 [X] Registering tensor: 786 for ONNX tensor: 786 [X] Slice_325 [Slice] outputs: [786 -> (2)[INT32]], [X] Parsing node: Concat_331 [Concat] [X] Searching for input: 786 [X] Searching for input: 815 [X] Concat_331 [Concat] inputs: [786 -> (2)[INT32]], [815 -> (2)[INT32]], [X] Registering layer: 815 for ONNX node: 815 [X] Registering layer: Concat_331 for ONNX node: Concat_331 [X] Registering tensor: 792 for ONNX tensor: 792 [X] Concat_331 [Concat] outputs: [792 -> (4)[INT32]], [X] Parsing node: Resize_332 [Resize] [X] Searching for input: 777 [X] Searching for input: 386 [X] Searching for input: 386 [X] Searching for input: 792 [X] Resize_332 [Resize] inputs: [777 -> (1, 4, 288, 512)[FLOAT]], [386 -> (0)[FLOAT]], [386 -> (0)[FLOAT]], [792 -> (4)[INT32]], [X] Registering layer: Resize_332 for ONNX node: Resize_332 [X] Registering tensor: 793 for ONNX tensor: 793 [X] Resize_332 [Resize] outputs: [793 -> (1, 4, 1440, 2560)[FLOAT]], [X] Parsing node: Shape_335 [Shape] [X] Searching for input: 779 [X] Shape_335 [Shape] inputs: [779 -> (1, 4, 288, 512)[FLOAT]], [X] Registering layer: Shape_335 for ONNX node: Shape_335 [X] Registering tensor: 796 for ONNX tensor: 796 [X] Shape_335 [Shape] outputs: [796 -> (4)[INT32]], [X] Parsing node: Slice_339 [Slice] [X] Searching for input: 796 [X] Searching for input: 632 [X] Searching for input: 630 [X] Searching for input: 632 [X] Slice_339 [Slice] inputs: [796 -> (4)[INT32]], [632 -> (1)[INT32]], [630 -> (1)[INT32]], [632 -> (1)[INT32]], [X] Registering layer: Slice_339 for ONNX node: Slice_339 [X] Registering tensor: 800 for ONNX tensor: 800 [X] Slice_339 [Slice] outputs: [800 -> (2)[INT32]], [X] Parsing node: Concat_345 [Concat] [X] Searching for input: 800 [X] Searching for input: 815 [X] Concat_345 [Concat] inputs: [800 -> (2)[INT32]], [815 -> (2)[INT32]], [X] Registering layer: Concat_345 for ONNX node: Concat_345 [X] Registering tensor: 806 for ONNX tensor: 806 [X] Concat_345 [Concat] outputs: [806 -> (4)[INT32]], [X] Parsing node: Resize_346 [Resize] [X] Searching for input: 779 [X] Searching for input: 386 [X] Searching for input: 386 [X] Searching for input: 806 [X] Resize_346 [Resize] inputs: [779 -> (1, 4, 288, 512)[FLOAT]], [386 -> (0)[FLOAT]], [386 -> (0)[FLOAT]], [806 -> (4)[INT32]], [X] Registering layer: Resize_346 for ONNX node: Resize_346 [X] Registering tensor: 807 for ONNX tensor: 807 [X] Resize_346 [Resize] outputs: [807 -> (1, 4, 1440, 2560)[FLOAT]], [X] Parsing node: Mul_347 [Mul] [X] Searching for input: 793 [X] Searching for input: 756 [X] Mul_347 [Mul] inputs: [793 -> (1, 4, 1440, 2560)[FLOAT]], [756 -> (1, 4, 1440, 2560)[FLOAT]], [X] Registering layer: Mul_347 for ONNX node: Mul_347 [X] Registering tensor: 808 for ONNX tensor: 808 [X] Mul_347 [Mul] outputs: [808 -> (1, 4, 1440, 2560)[FLOAT]], [X] Parsing node: Add_348 [Add] [X] Searching for input: 808 [X] Searching for input: 807 [X] Add_348 [Add] inputs: [808 -> (1, 4, 1440, 2560)[FLOAT]], [807 -> (1, 4, 1440, 2560)[FLOAT]], [X] Registering layer: Add_348 for ONNX node: Add_348 [X] Registering tensor: 809 for ONNX tensor: 809 [X] Add_348 [Add] outputs: [809 -> (1, 4, 1440, 2560)[FLOAT]], [X] Parsing node: Split_349 [Split] [X] Searching for input: 809 [X] Split_349 [Split] inputs: [809 -> (1, 4, 1440, 2560)[FLOAT]], [X] Registering layer: Split_349 for ONNX node: Split_349 [X] Registering layer: Split_349_44 for ONNX node: Split_349 [X] Registering tensor: 810 for ONNX tensor: 810 [X] Registering tensor: 811 for ONNX tensor: 811 [X] Split_349 [Split] outputs: [810 -> (1, 3, 1440, 2560)[FLOAT]], [811 -> (1, 1, 1440, 2560)[FLOAT]], [X] Parsing node: Add_350 [Add] [X] Searching for input: 810 [X] Searching for input: src [X] Add_350 [Add] inputs: [810 -> (1, 3, 1440, 2560)[FLOAT]], [src -> (1, 3, 1440, 2560)[FLOAT]], [X] Registering layer: Add_350 for ONNX node: Add_350 [X] Registering tensor: 812 for ONNX tensor: 812 [X] Add_350 [Add] outputs: [812 -> (1, 3, 1440, 2560)[FLOAT]], [X] Parsing node: Clip_351 [Clip] [X] Searching for input: 812 [X] Searching for input: 989 [X] Searching for input: 990 [X] Clip_351 [Clip] inputs: [812 -> (1, 3, 1440, 2560)[FLOAT]], [989 -> ()[FLOAT]], [990 -> ()[FLOAT]], [X] Registering layer: Clip_351 for ONNX node: Clip_351 [X] Registering tensor: fgr_45 for ONNX tensor: fgr [X] Clip_351 [Clip] outputs: [fgr -> (1, 3, 1440, 2560)[FLOAT]], [X] Parsing node: Clip_352 [Clip] [X] Searching for input: 811 [X] Searching for input: 989 [X] Searching for input: 990 [X] Clip_352 [Clip] inputs: [811 -> (1, 1, 1440, 2560)[FLOAT]], [989 -> ()[FLOAT]], [990 -> ()[FLOAT]], [X] Registering layer: Clip_352 for ONNX node: Clip_352 [X] Registering tensor: pha_46 for ONNX tensor: pha [X] Clip_352 [Clip] outputs: [pha -> (1, 1, 1440, 2560)[FLOAT]], [X] Marking fgr_45 as output: fgr [X] Marking pha_46 as output: pha [X] Marking r1o_36 as output: r1o [X] Marking r2o_25 as output: r2o [X] Marking r3o_14 as output: r3o [X] Marking r4o_3 as output: r4o [V] Setting TensorRT Optimization Profiles [V] Input tensor: src (dtype=DataType.FLOAT, shape=(1, 3, 1440, 2560)) | Setting input tensor shapes to: (min=[1, 3, 1440, 2560], opt=[1, 3, 1440, 2560], max=[1, 3, 1440, 2560]) [V] Input tensor: r1i (dtype=DataType.FLOAT, shape=(1, 1, 1, 1)) | Setting input tensor shapes to: (min=[1, 1, 1, 1], opt=[1, 1, 1, 1], max=[1, 1, 1, 1]) [V] Input tensor: r2i (dtype=DataType.FLOAT, shape=(1, 1, 1, 1)) | Setting input tensor shapes to: (min=[1, 1, 1, 1], opt=[1, 1, 1, 1], max=[1, 1, 1, 1]) [V] Input tensor: r3i (dtype=DataType.FLOAT, shape=(1, 1, 1, 1)) | Setting input tensor shapes to: (min=[1, 1, 1, 1], opt=[1, 1, 1, 1], max=[1, 1, 1, 1]) [V] Input tensor: r4i (dtype=DataType.FLOAT, shape=(1, 1, 1, 1)) | Setting input tensor shapes to: (min=[1, 1, 1, 1], opt=[1, 1, 1, 1], max=[1, 1, 1, 1]) [I] Configuring with profiles: [Profile().add('src', min=[1, 3, 1440, 2560], opt=[1, 3, 1440, 2560], max=[1, 3, 1440, 2560]).add('r1i', min=[1, 1, 1, 1], opt=[1, 1, 1, 1], max=[1, 1, 1, 1]).add('r2i', min=[1, 1, 1, 1], opt=[1, 1, 1, 1], max=[1, 1, 1, 1]).add('r3i', min=[1, 1, 1, 1], opt=[1, 1, 1, 1], max=[1, 1, 1, 1]).add('r4i', min=[1, 1, 1, 1], opt=[1, 1, 1, 1], max=[1, 1, 1, 1])] [E] 3: [builderConfig.cpp::getFlag::42] Error Code 3: API Usage Error (Parameter check failed at: optimizer/api/builderConfig.cpp::getFlag::42, condition: int(builderFlag) >= 0 && int(builderFlag) < EnumMax() ) Segmentation fault (core dumped) (venv) tech@Artax:/opt/metamirror/onnx_trt_utils$