^[[A^[[A(media_app_deployment) sai (master *) scripts $ ./onnx_torch2trt.sh ../models/text_recognition/latin.onnx ../models/text_recognition latin.onnx &&&& RUNNING TensorRT.trtexec # /home/sai/ncai/TensorRT-7.0.0.11/bin/trtexec --fp16 --explicitBatch --workspace=2048 --onnx=../models/text_recognition/latin.onnx --saveEngine=../models/text_recognition/latin.onnx.trt --minShapes='input':1x1x64x16 --optShapes='input':1x1x64x320 --maxShapes='input':2x1x64x3200 --verbose [11/26/2020-16:41:45] [I] === Model Options === [11/26/2020-16:41:45] [I] Format: ONNX [11/26/2020-16:41:45] [I] Model: ../models/text_recognition/latin.onnx [11/26/2020-16:41:45] [I] Output: [11/26/2020-16:41:45] [I] === Build Options === [11/26/2020-16:41:45] [I] Max batch: explicit [11/26/2020-16:41:45] [I] Workspace: 2048 MB [11/26/2020-16:41:45] [I] minTiming: 1 [11/26/2020-16:41:45] [I] avgTiming: 8 [11/26/2020-16:41:45] [I] Precision: FP16 [11/26/2020-16:41:45] [I] Calibration: [11/26/2020-16:41:45] [I] Safe mode: Disabled [11/26/2020-16:41:45] [I] Save engine: ../models/text_recognition/latin.onnx.trt [11/26/2020-16:41:45] [I] Load engine: [11/26/2020-16:41:45] [I] Inputs format: fp32:CHW [11/26/2020-16:41:45] [I] Outputs format: fp32:CHW [11/26/2020-16:41:45] [I] Input build shape: input=1x1x64x16+1x1x64x320+2x1x64x3200 [11/26/2020-16:41:45] [I] === System Options === [11/26/2020-16:41:45] [I] Device: 0 [11/26/2020-16:41:45] [I] DLACore: [11/26/2020-16:41:45] [I] Plugins: [11/26/2020-16:41:45] [I] === Inference Options === [11/26/2020-16:41:45] [I] Batch: Explicit [11/26/2020-16:41:45] [I] Iterations: 10 [11/26/2020-16:41:45] [I] Duration: 3s (+ 200ms warm up) [11/26/2020-16:41:45] [I] Sleep time: 0ms [11/26/2020-16:41:45] [I] Streams: 1 [11/26/2020-16:41:45] [I] ExposeDMA: Disabled [11/26/2020-16:41:45] [I] Spin-wait: Disabled [11/26/2020-16:41:45] [I] Multithreading: Disabled [11/26/2020-16:41:45] [I] CUDA Graph: Disabled [11/26/2020-16:41:45] [I] Skip inference: Disabled [11/26/2020-16:41:45] [I] Inputs: [11/26/2020-16:41:45] [I] === Reporting Options === [11/26/2020-16:41:45] [I] Verbose: Enabled [11/26/2020-16:41:45] [I] Averages: 10 inferences [11/26/2020-16:41:45] [I] Percentile: 99 [11/26/2020-16:41:45] [I] Dump output: Disabled [11/26/2020-16:41:45] [I] Profile: Disabled [11/26/2020-16:41:45] [I] Export timing to JSON file: [11/26/2020-16:41:45] [I] Export output to JSON file: [11/26/2020-16:41:45] [I] Export profile to JSON file: [11/26/2020-16:41:45] [I] [11/26/2020-16:41:45] [V] [TRT] Plugin creator registration succeeded - ::GridAnchor_TRT [11/26/2020-16:41:45] [V] [TRT] Plugin creator registration succeeded - ::NMS_TRT [11/26/2020-16:41:45] [V] [TRT] Plugin creator registration succeeded - ::Reorg_TRT [11/26/2020-16:41:45] [V] [TRT] Plugin creator registration succeeded - ::Region_TRT [11/26/2020-16:41:45] [V] [TRT] Plugin creator registration succeeded - ::PriorBox_TRT [11/26/2020-16:41:45] [V] [TRT] Plugin creator registration succeeded - ::Normalize_TRT [11/26/2020-16:41:45] [V] [TRT] Plugin creator registration succeeded - ::RPROI_TRT [11/26/2020-16:41:45] [V] [TRT] Plugin creator registration succeeded - ::BatchedNMS_TRT [11/26/2020-16:41:45] [V] [TRT] Plugin creator registration succeeded - ::FlattenConcat_TRT [11/26/2020-16:41:45] [V] [TRT] Plugin creator registration succeeded - ::CropAndResize [11/26/2020-16:41:45] [V] [TRT] Plugin creator registration succeeded - ::Proposal [11/26/2020-16:41:45] [V] [TRT] Plugin creator registration succeeded - ::BatchTilePlugin_TRT [11/26/2020-16:41:45] [V] [TRT] Plugin creator registration succeeded - ::DetectionLayer_TRT [11/26/2020-16:41:45] [V] [TRT] Plugin creator registration succeeded - ::ProposalLayer_TRT [11/26/2020-16:41:45] [V] [TRT] Plugin creator registration succeeded - ::PyramidROIAlign_TRT [11/26/2020-16:41:45] [V] [TRT] Plugin creator registration succeeded - ::ResizeNearest_TRT [11/26/2020-16:41:45] [V] [TRT] Plugin creator registration succeeded - ::SpecialSlice_TRT [11/26/2020-16:41:45] [V] [TRT] Plugin creator registration succeeded - ::InstanceNormalization_TRT [11/26/2020-16:41:45] [V] [TRT] Plugin creator registration succeeded - ::GenerateDetection_TRT [11/26/2020-16:41:45] [V] [TRT] Plugin creator registration succeeded - ::MultilevelProposeROI_TRT [11/26/2020-16:41:45] [V] [TRT] Plugin creator registration succeeded - ::MultilevelCropAndResize_TRT ---------------------------------------------------------------- Input filename: ../models/text_recognition/latin.onnx ONNX IR version: 0.0.6 Opset version: 11 Producer name: pytorch Producer version: 1.7 Domain: Model version: 0 Doc string: ---------------------------------------------------------------- [11/26/2020-16:41:46] [V] [TRT] Plugin creator registration succeeded - ONNXTRT_NAMESPACE::GridAnchor_TRT [11/26/2020-16:41:46] [V] [TRT] Plugin creator registration succeeded - ONNXTRT_NAMESPACE::NMS_TRT [11/26/2020-16:41:46] [V] [TRT] Plugin creator registration succeeded - ONNXTRT_NAMESPACE::Reorg_TRT [11/26/2020-16:41:46] [V] [TRT] Plugin creator registration succeeded - ONNXTRT_NAMESPACE::Region_TRT [11/26/2020-16:41:46] [V] [TRT] Plugin creator registration succeeded - ONNXTRT_NAMESPACE::PriorBox_TRT [11/26/2020-16:41:46] [V] [TRT] Plugin creator registration succeeded - ONNXTRT_NAMESPACE::Normalize_TRT [11/26/2020-16:41:46] [V] [TRT] Plugin creator registration succeeded - ONNXTRT_NAMESPACE::RPROI_TRT [11/26/2020-16:41:46] [V] [TRT] Plugin creator registration succeeded - ONNXTRT_NAMESPACE::BatchedNMS_TRT [11/26/2020-16:41:46] [V] [TRT] Plugin creator registration succeeded - ONNXTRT_NAMESPACE::FlattenConcat_TRT [11/26/2020-16:41:46] [V] [TRT] Plugin creator registration succeeded - ONNXTRT_NAMESPACE::CropAndResize [11/26/2020-16:41:46] [V] [TRT] Plugin creator registration succeeded - ONNXTRT_NAMESPACE::Proposal [11/26/2020-16:41:46] [V] [TRT] Plugin creator registration succeeded - ONNXTRT_NAMESPACE::BatchTilePlugin_TRT [11/26/2020-16:41:46] [V] [TRT] Plugin creator registration succeeded - ONNXTRT_NAMESPACE::DetectionLayer_TRT [11/26/2020-16:41:46] [V] [TRT] Plugin creator registration succeeded - ONNXTRT_NAMESPACE::ProposalLayer_TRT [11/26/2020-16:41:46] [V] [TRT] Plugin creator registration succeeded - ONNXTRT_NAMESPACE::PyramidROIAlign_TRT [11/26/2020-16:41:46] [V] [TRT] Plugin creator registration succeeded - ONNXTRT_NAMESPACE::ResizeNearest_TRT [11/26/2020-16:41:46] [V] [TRT] Plugin creator registration succeeded - ONNXTRT_NAMESPACE::SpecialSlice_TRT [11/26/2020-16:41:46] [V] [TRT] Plugin creator registration succeeded - ONNXTRT_NAMESPACE::InstanceNormalization_TRT [11/26/2020-16:41:46] [V] [TRT] Plugin creator registration succeeded - ONNXTRT_NAMESPACE::GenerateDetection_TRT [11/26/2020-16:41:46] [V] [TRT] Plugin creator registration succeeded - ONNXTRT_NAMESPACE::MultilevelProposeROI_TRT [11/26/2020-16:41:46] [V] [TRT] Plugin creator registration succeeded - ONNXTRT_NAMESPACE::MultilevelCropAndResize_TRT [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:204: Adding network input: input with dtype: float32, dimensions: (-1, 1, 64, -1) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: input for ONNX tensor: input [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:90: Importing initializer: 617 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:90: Importing initializer: 618 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:90: Importing initializer: 620 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:90: Importing initializer: 621 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:90: Importing initializer: 623 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:90: Importing initializer: 624 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:90: Importing initializer: 626 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:90: Importing initializer: 627 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:90: Importing initializer: 629 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:90: Importing initializer: 630 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:90: Importing initializer: 632 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:90: Importing initializer: 633 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:90: Importing initializer: 635 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:90: Importing initializer: 636 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:90: Importing initializer: 638 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:90: Importing initializer: 639 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:90: Importing initializer: 641 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:90: Importing initializer: 642 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:90: Importing initializer: 644 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:90: Importing initializer: 645 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:90: Importing initializer: 647 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:90: Importing initializer: 648 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:90: Importing initializer: 650 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:90: Importing initializer: 651 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:90: Importing initializer: 653 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:90: Importing initializer: 654 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:90: Importing initializer: 656 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:90: Importing initializer: 657 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:90: Importing initializer: 659 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:90: Importing initializer: 660 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:90: Importing initializer: 662 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:90: Importing initializer: 663 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:90: Importing initializer: 665 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:90: Importing initializer: 666 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:90: Importing initializer: 668 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:90: Importing initializer: 669 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:90: Importing initializer: 671 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:90: Importing initializer: 672 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:90: Importing initializer: 674 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:90: Importing initializer: 675 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:90: Importing initializer: 677 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:90: Importing initializer: 678 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:90: Importing initializer: 680 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:90: Importing initializer: 681 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:90: Importing initializer: 683 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:90: Importing initializer: 684 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:90: Importing initializer: 686 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:90: Importing initializer: 687 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:90: Importing initializer: 689 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:90: Importing initializer: 690 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:90: Importing initializer: 692 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:90: Importing initializer: 693 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:90: Importing initializer: 695 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:90: Importing initializer: 696 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:90: Importing initializer: 698 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:90: Importing initializer: 699 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:90: Importing initializer: 701 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:90: Importing initializer: 702 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:90: Importing initializer: 704 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:90: Importing initializer: 705 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:90: Importing initializer: 707 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:90: Importing initializer: 708 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:90: Importing initializer: 710 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:90: Importing initializer: 711 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:90: Importing initializer: 712 [11/26/2020-16:41:46] [W] [TRT] /home/sai/TensorRT/parsers/onnx/onnx2trt_utils.cpp:235: Your ONNX model has been generated with INT64 weights, while TensorRT does not natively support INT64. Attempting to cast down to INT32. [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:90: Importing initializer: 713 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:90: Importing initializer: 754 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:90: Importing initializer: 755 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:90: Importing initializer: 756 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:90: Importing initializer: 757 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:90: Importing initializer: 758 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:90: Importing initializer: 759 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:90: Importing initializer: 800 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:90: Importing initializer: 801 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:90: Importing initializer: 802 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:90: Importing initializer: 803 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:90: Importing initializer: 804 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:90: Importing initializer: Prediction.bias [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:90: Importing initializer: SequenceModeling.0.linear.bias [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:90: Importing initializer: SequenceModeling.1.linear.bias [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Conv_0 [Conv] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: input [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 617 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 618 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Conv_0 [Conv] inputs: [input -> (-1, 1, 64, -1)], [617 -> (32, 1, 3, 3)], [618 -> (32)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:444: Convolution input dimensions: (-1, 1, 64, -1) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:526: Using kernel: (3, 3), strides: (1, 1), padding: (1, 1), dilations: (1, 1), numOutputs: 32 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:527: Convolution output dimensions: (-1, 32, 64, -1) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Conv_0 for ONNX node: Conv_0 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 616 for ONNX tensor: 616 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Conv_0 [Conv] outputs: [616 -> (-1, 32, 64, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Relu_1 [Relu] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 616 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Relu_1 [Relu] inputs: [616 -> (-1, 32, 64, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Relu_1 for ONNX node: Relu_1 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 217 for ONNX tensor: 217 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Relu_1 [Relu] outputs: [217 -> (-1, 32, 64, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Conv_2 [Conv] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 217 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 620 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 621 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Conv_2 [Conv] inputs: [217 -> (-1, 32, 64, -1)], [620 -> (64, 32, 3, 3)], [621 -> (64)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:444: Convolution input dimensions: (-1, 32, 64, -1) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:526: Using kernel: (3, 3), strides: (1, 1), padding: (1, 1), dilations: (1, 1), numOutputs: 64 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:527: Convolution output dimensions: (-1, 64, 64, -1) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Conv_2 for ONNX node: Conv_2 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 619 for ONNX tensor: 619 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Conv_2 [Conv] outputs: [619 -> (-1, 64, 64, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Relu_3 [Relu] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 619 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Relu_3 [Relu] inputs: [619 -> (-1, 64, 64, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Relu_3 for ONNX node: Relu_3 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 220 for ONNX tensor: 220 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Relu_3 [Relu] outputs: [220 -> (-1, 64, 64, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: MaxPool_4 [MaxPool] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 220 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: MaxPool_4 [MaxPool] inputs: [220 -> (-1, 64, 64, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: MaxPool_4 for ONNX node: MaxPool_4 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 221 for ONNX tensor: 221 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: MaxPool_4 [MaxPool] outputs: [221 -> (-1, 64, 32, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Conv_5 [Conv] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 221 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 623 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 624 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Conv_5 [Conv] inputs: [221 -> (-1, 64, 32, -1)], [623 -> (128, 64, 3, 3)], [624 -> (128)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:444: Convolution input dimensions: (-1, 64, 32, -1) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:526: Using kernel: (3, 3), strides: (1, 1), padding: (1, 1), dilations: (1, 1), numOutputs: 128 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:527: Convolution output dimensions: (-1, 128, 32, -1) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Conv_5 for ONNX node: Conv_5 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 622 for ONNX tensor: 622 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Conv_5 [Conv] outputs: [622 -> (-1, 128, 32, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Relu_6 [Relu] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 622 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Relu_6 [Relu] inputs: [622 -> (-1, 128, 32, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Relu_6 for ONNX node: Relu_6 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 224 for ONNX tensor: 224 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Relu_6 [Relu] outputs: [224 -> (-1, 128, 32, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Conv_7 [Conv] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 224 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 626 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 627 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Conv_7 [Conv] inputs: [224 -> (-1, 128, 32, -1)], [626 -> (128, 128, 3, 3)], [627 -> (128)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:444: Convolution input dimensions: (-1, 128, 32, -1) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:526: Using kernel: (3, 3), strides: (1, 1), padding: (1, 1), dilations: (1, 1), numOutputs: 128 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:527: Convolution output dimensions: (-1, 128, 32, -1) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Conv_7 for ONNX node: Conv_7 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 625 for ONNX tensor: 625 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Conv_7 [Conv] outputs: [625 -> (-1, 128, 32, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Conv_8 [Conv] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 221 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 629 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 630 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Conv_8 [Conv] inputs: [221 -> (-1, 64, 32, -1)], [629 -> (128, 64, 1, 1)], [630 -> (128)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:444: Convolution input dimensions: (-1, 64, 32, -1) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:526: Using kernel: (1, 1), strides: (1, 1), padding: (0, 0), dilations: (1, 1), numOutputs: 128 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:527: Convolution output dimensions: (-1, 128, 32, -1) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Conv_8 for ONNX node: Conv_8 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 628 for ONNX tensor: 628 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Conv_8 [Conv] outputs: [628 -> (-1, 128, 32, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Add_9 [Add] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 625 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 628 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Add_9 [Add] inputs: [625 -> (-1, 128, 32, -1)], [628 -> (-1, 128, 32, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Add_9 for ONNX node: Add_9 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 229 for ONNX tensor: 229 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Add_9 [Add] outputs: [229 -> (-1, 128, 32, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Relu_10 [Relu] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 229 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Relu_10 [Relu] inputs: [229 -> (-1, 128, 32, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Relu_10 for ONNX node: Relu_10 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 230 for ONNX tensor: 230 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Relu_10 [Relu] outputs: [230 -> (-1, 128, 32, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Conv_11 [Conv] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 230 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 632 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 633 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Conv_11 [Conv] inputs: [230 -> (-1, 128, 32, -1)], [632 -> (128, 128, 3, 3)], [633 -> (128)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:444: Convolution input dimensions: (-1, 128, 32, -1) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:526: Using kernel: (3, 3), strides: (1, 1), padding: (1, 1), dilations: (1, 1), numOutputs: 128 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:527: Convolution output dimensions: (-1, 128, 32, -1) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Conv_11 for ONNX node: Conv_11 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 631 for ONNX tensor: 631 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Conv_11 [Conv] outputs: [631 -> (-1, 128, 32, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Relu_12 [Relu] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 631 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Relu_12 [Relu] inputs: [631 -> (-1, 128, 32, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Relu_12 for ONNX node: Relu_12 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 233 for ONNX tensor: 233 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Relu_12 [Relu] outputs: [233 -> (-1, 128, 32, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: MaxPool_13 [MaxPool] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 233 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: MaxPool_13 [MaxPool] inputs: [233 -> (-1, 128, 32, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: MaxPool_13 for ONNX node: MaxPool_13 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 234 for ONNX tensor: 234 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: MaxPool_13 [MaxPool] outputs: [234 -> (-1, 128, 16, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Conv_14 [Conv] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 234 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 635 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 636 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Conv_14 [Conv] inputs: [234 -> (-1, 128, 16, -1)], [635 -> (256, 128, 3, 3)], [636 -> (256)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:444: Convolution input dimensions: (-1, 128, 16, -1) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:526: Using kernel: (3, 3), strides: (1, 1), padding: (1, 1), dilations: (1, 1), numOutputs: 256 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:527: Convolution output dimensions: (-1, 256, 16, -1) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Conv_14 for ONNX node: Conv_14 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 634 for ONNX tensor: 634 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Conv_14 [Conv] outputs: [634 -> (-1, 256, 16, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Relu_15 [Relu] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 634 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Relu_15 [Relu] inputs: [634 -> (-1, 256, 16, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Relu_15 for ONNX node: Relu_15 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 237 for ONNX tensor: 237 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Relu_15 [Relu] outputs: [237 -> (-1, 256, 16, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Conv_16 [Conv] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 237 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 638 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 639 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Conv_16 [Conv] inputs: [237 -> (-1, 256, 16, -1)], [638 -> (256, 256, 3, 3)], [639 -> (256)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:444: Convolution input dimensions: (-1, 256, 16, -1) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:526: Using kernel: (3, 3), strides: (1, 1), padding: (1, 1), dilations: (1, 1), numOutputs: 256 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:527: Convolution output dimensions: (-1, 256, 16, -1) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Conv_16 for ONNX node: Conv_16 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 637 for ONNX tensor: 637 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Conv_16 [Conv] outputs: [637 -> (-1, 256, 16, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Conv_17 [Conv] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 234 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 641 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 642 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Conv_17 [Conv] inputs: [234 -> (-1, 128, 16, -1)], [641 -> (256, 128, 1, 1)], [642 -> (256)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:444: Convolution input dimensions: (-1, 128, 16, -1) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:526: Using kernel: (1, 1), strides: (1, 1), padding: (0, 0), dilations: (1, 1), numOutputs: 256 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:527: Convolution output dimensions: (-1, 256, 16, -1) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Conv_17 for ONNX node: Conv_17 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 640 for ONNX tensor: 640 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Conv_17 [Conv] outputs: [640 -> (-1, 256, 16, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Add_18 [Add] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 637 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 640 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Add_18 [Add] inputs: [637 -> (-1, 256, 16, -1)], [640 -> (-1, 256, 16, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Add_18 for ONNX node: Add_18 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 242 for ONNX tensor: 242 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Add_18 [Add] outputs: [242 -> (-1, 256, 16, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Relu_19 [Relu] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 242 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Relu_19 [Relu] inputs: [242 -> (-1, 256, 16, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Relu_19 for ONNX node: Relu_19 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 243 for ONNX tensor: 243 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Relu_19 [Relu] outputs: [243 -> (-1, 256, 16, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Conv_20 [Conv] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 243 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 644 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 645 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Conv_20 [Conv] inputs: [243 -> (-1, 256, 16, -1)], [644 -> (256, 256, 3, 3)], [645 -> (256)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:444: Convolution input dimensions: (-1, 256, 16, -1) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:526: Using kernel: (3, 3), strides: (1, 1), padding: (1, 1), dilations: (1, 1), numOutputs: 256 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:527: Convolution output dimensions: (-1, 256, 16, -1) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Conv_20 for ONNX node: Conv_20 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 643 for ONNX tensor: 643 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Conv_20 [Conv] outputs: [643 -> (-1, 256, 16, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Relu_21 [Relu] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 643 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Relu_21 [Relu] inputs: [643 -> (-1, 256, 16, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Relu_21 for ONNX node: Relu_21 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 246 for ONNX tensor: 246 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Relu_21 [Relu] outputs: [246 -> (-1, 256, 16, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Conv_22 [Conv] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 246 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 647 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 648 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Conv_22 [Conv] inputs: [246 -> (-1, 256, 16, -1)], [647 -> (256, 256, 3, 3)], [648 -> (256)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:444: Convolution input dimensions: (-1, 256, 16, -1) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:526: Using kernel: (3, 3), strides: (1, 1), padding: (1, 1), dilations: (1, 1), numOutputs: 256 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:527: Convolution output dimensions: (-1, 256, 16, -1) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Conv_22 for ONNX node: Conv_22 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 646 for ONNX tensor: 646 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Conv_22 [Conv] outputs: [646 -> (-1, 256, 16, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Add_23 [Add] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 646 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 243 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Add_23 [Add] inputs: [646 -> (-1, 256, 16, -1)], [243 -> (-1, 256, 16, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Add_23 for ONNX node: Add_23 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 249 for ONNX tensor: 249 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Add_23 [Add] outputs: [249 -> (-1, 256, 16, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Relu_24 [Relu] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 249 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Relu_24 [Relu] inputs: [249 -> (-1, 256, 16, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Relu_24 for ONNX node: Relu_24 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 250 for ONNX tensor: 250 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Relu_24 [Relu] outputs: [250 -> (-1, 256, 16, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Conv_25 [Conv] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 250 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 650 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 651 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Conv_25 [Conv] inputs: [250 -> (-1, 256, 16, -1)], [650 -> (256, 256, 3, 3)], [651 -> (256)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:444: Convolution input dimensions: (-1, 256, 16, -1) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:526: Using kernel: (3, 3), strides: (1, 1), padding: (1, 1), dilations: (1, 1), numOutputs: 256 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:527: Convolution output dimensions: (-1, 256, 16, -1) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Conv_25 for ONNX node: Conv_25 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 649 for ONNX tensor: 649 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Conv_25 [Conv] outputs: [649 -> (-1, 256, 16, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Relu_26 [Relu] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 649 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Relu_26 [Relu] inputs: [649 -> (-1, 256, 16, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Relu_26 for ONNX node: Relu_26 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 253 for ONNX tensor: 253 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Relu_26 [Relu] outputs: [253 -> (-1, 256, 16, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: MaxPool_27 [MaxPool] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 253 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: MaxPool_27 [MaxPool] inputs: [253 -> (-1, 256, 16, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: MaxPool_27 for ONNX node: MaxPool_27 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 254 for ONNX tensor: 254 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: MaxPool_27 [MaxPool] outputs: [254 -> (-1, 256, 8, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Conv_28 [Conv] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 254 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 653 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 654 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Conv_28 [Conv] inputs: [254 -> (-1, 256, 8, -1)], [653 -> (512, 256, 3, 3)], [654 -> (512)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:444: Convolution input dimensions: (-1, 256, 8, -1) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:526: Using kernel: (3, 3), strides: (1, 1), padding: (1, 1), dilations: (1, 1), numOutputs: 512 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:527: Convolution output dimensions: (-1, 512, 8, -1) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Conv_28 for ONNX node: Conv_28 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 652 for ONNX tensor: 652 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Conv_28 [Conv] outputs: [652 -> (-1, 512, 8, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Relu_29 [Relu] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 652 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Relu_29 [Relu] inputs: [652 -> (-1, 512, 8, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Relu_29 for ONNX node: Relu_29 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 257 for ONNX tensor: 257 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Relu_29 [Relu] outputs: [257 -> (-1, 512, 8, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Conv_30 [Conv] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 257 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 656 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 657 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Conv_30 [Conv] inputs: [257 -> (-1, 512, 8, -1)], [656 -> (512, 512, 3, 3)], [657 -> (512)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:444: Convolution input dimensions: (-1, 512, 8, -1) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:526: Using kernel: (3, 3), strides: (1, 1), padding: (1, 1), dilations: (1, 1), numOutputs: 512 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:527: Convolution output dimensions: (-1, 512, 8, -1) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Conv_30 for ONNX node: Conv_30 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 655 for ONNX tensor: 655 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Conv_30 [Conv] outputs: [655 -> (-1, 512, 8, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Conv_31 [Conv] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 254 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 659 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 660 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Conv_31 [Conv] inputs: [254 -> (-1, 256, 8, -1)], [659 -> (512, 256, 1, 1)], [660 -> (512)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:444: Convolution input dimensions: (-1, 256, 8, -1) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:526: Using kernel: (1, 1), strides: (1, 1), padding: (0, 0), dilations: (1, 1), numOutputs: 512 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:527: Convolution output dimensions: (-1, 512, 8, -1) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Conv_31 for ONNX node: Conv_31 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 658 for ONNX tensor: 658 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Conv_31 [Conv] outputs: [658 -> (-1, 512, 8, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Add_32 [Add] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 655 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 658 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Add_32 [Add] inputs: [655 -> (-1, 512, 8, -1)], [658 -> (-1, 512, 8, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Add_32 for ONNX node: Add_32 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 262 for ONNX tensor: 262 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Add_32 [Add] outputs: [262 -> (-1, 512, 8, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Relu_33 [Relu] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 262 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Relu_33 [Relu] inputs: [262 -> (-1, 512, 8, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Relu_33 for ONNX node: Relu_33 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 263 for ONNX tensor: 263 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Relu_33 [Relu] outputs: [263 -> (-1, 512, 8, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Conv_34 [Conv] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 263 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 662 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 663 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Conv_34 [Conv] inputs: [263 -> (-1, 512, 8, -1)], [662 -> (512, 512, 3, 3)], [663 -> (512)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:444: Convolution input dimensions: (-1, 512, 8, -1) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:526: Using kernel: (3, 3), strides: (1, 1), padding: (1, 1), dilations: (1, 1), numOutputs: 512 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:527: Convolution output dimensions: (-1, 512, 8, -1) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Conv_34 for ONNX node: Conv_34 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 661 for ONNX tensor: 661 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Conv_34 [Conv] outputs: [661 -> (-1, 512, 8, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Relu_35 [Relu] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 661 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Relu_35 [Relu] inputs: [661 -> (-1, 512, 8, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Relu_35 for ONNX node: Relu_35 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 266 for ONNX tensor: 266 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Relu_35 [Relu] outputs: [266 -> (-1, 512, 8, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Conv_36 [Conv] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 266 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 665 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 666 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Conv_36 [Conv] inputs: [266 -> (-1, 512, 8, -1)], [665 -> (512, 512, 3, 3)], [666 -> (512)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:444: Convolution input dimensions: (-1, 512, 8, -1) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:526: Using kernel: (3, 3), strides: (1, 1), padding: (1, 1), dilations: (1, 1), numOutputs: 512 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:527: Convolution output dimensions: (-1, 512, 8, -1) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Conv_36 for ONNX node: Conv_36 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 664 for ONNX tensor: 664 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Conv_36 [Conv] outputs: [664 -> (-1, 512, 8, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Add_37 [Add] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 664 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 263 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Add_37 [Add] inputs: [664 -> (-1, 512, 8, -1)], [263 -> (-1, 512, 8, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Add_37 for ONNX node: Add_37 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 269 for ONNX tensor: 269 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Add_37 [Add] outputs: [269 -> (-1, 512, 8, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Relu_38 [Relu] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 269 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Relu_38 [Relu] inputs: [269 -> (-1, 512, 8, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Relu_38 for ONNX node: Relu_38 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 270 for ONNX tensor: 270 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Relu_38 [Relu] outputs: [270 -> (-1, 512, 8, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Conv_39 [Conv] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 270 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 668 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 669 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Conv_39 [Conv] inputs: [270 -> (-1, 512, 8, -1)], [668 -> (512, 512, 3, 3)], [669 -> (512)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:444: Convolution input dimensions: (-1, 512, 8, -1) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:526: Using kernel: (3, 3), strides: (1, 1), padding: (1, 1), dilations: (1, 1), numOutputs: 512 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:527: Convolution output dimensions: (-1, 512, 8, -1) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Conv_39 for ONNX node: Conv_39 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 667 for ONNX tensor: 667 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Conv_39 [Conv] outputs: [667 -> (-1, 512, 8, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Relu_40 [Relu] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 667 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Relu_40 [Relu] inputs: [667 -> (-1, 512, 8, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Relu_40 for ONNX node: Relu_40 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 273 for ONNX tensor: 273 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Relu_40 [Relu] outputs: [273 -> (-1, 512, 8, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Conv_41 [Conv] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 273 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 671 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 672 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Conv_41 [Conv] inputs: [273 -> (-1, 512, 8, -1)], [671 -> (512, 512, 3, 3)], [672 -> (512)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:444: Convolution input dimensions: (-1, 512, 8, -1) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:526: Using kernel: (3, 3), strides: (1, 1), padding: (1, 1), dilations: (1, 1), numOutputs: 512 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:527: Convolution output dimensions: (-1, 512, 8, -1) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Conv_41 for ONNX node: Conv_41 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 670 for ONNX tensor: 670 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Conv_41 [Conv] outputs: [670 -> (-1, 512, 8, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Add_42 [Add] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 670 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 270 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Add_42 [Add] inputs: [670 -> (-1, 512, 8, -1)], [270 -> (-1, 512, 8, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Add_42 for ONNX node: Add_42 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 276 for ONNX tensor: 276 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Add_42 [Add] outputs: [276 -> (-1, 512, 8, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Relu_43 [Relu] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 276 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Relu_43 [Relu] inputs: [276 -> (-1, 512, 8, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Relu_43 for ONNX node: Relu_43 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 277 for ONNX tensor: 277 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Relu_43 [Relu] outputs: [277 -> (-1, 512, 8, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Conv_44 [Conv] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 277 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 674 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 675 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Conv_44 [Conv] inputs: [277 -> (-1, 512, 8, -1)], [674 -> (512, 512, 3, 3)], [675 -> (512)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:444: Convolution input dimensions: (-1, 512, 8, -1) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:526: Using kernel: (3, 3), strides: (1, 1), padding: (1, 1), dilations: (1, 1), numOutputs: 512 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:527: Convolution output dimensions: (-1, 512, 8, -1) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Conv_44 for ONNX node: Conv_44 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 673 for ONNX tensor: 673 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Conv_44 [Conv] outputs: [673 -> (-1, 512, 8, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Relu_45 [Relu] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 673 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Relu_45 [Relu] inputs: [673 -> (-1, 512, 8, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Relu_45 for ONNX node: Relu_45 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 280 for ONNX tensor: 280 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Relu_45 [Relu] outputs: [280 -> (-1, 512, 8, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Conv_46 [Conv] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 280 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 677 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 678 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Conv_46 [Conv] inputs: [280 -> (-1, 512, 8, -1)], [677 -> (512, 512, 3, 3)], [678 -> (512)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:444: Convolution input dimensions: (-1, 512, 8, -1) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:526: Using kernel: (3, 3), strides: (1, 1), padding: (1, 1), dilations: (1, 1), numOutputs: 512 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:527: Convolution output dimensions: (-1, 512, 8, -1) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Conv_46 for ONNX node: Conv_46 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 676 for ONNX tensor: 676 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Conv_46 [Conv] outputs: [676 -> (-1, 512, 8, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Add_47 [Add] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 676 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 277 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Add_47 [Add] inputs: [676 -> (-1, 512, 8, -1)], [277 -> (-1, 512, 8, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Add_47 for ONNX node: Add_47 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 283 for ONNX tensor: 283 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Add_47 [Add] outputs: [283 -> (-1, 512, 8, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Relu_48 [Relu] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 283 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Relu_48 [Relu] inputs: [283 -> (-1, 512, 8, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Relu_48 for ONNX node: Relu_48 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 284 for ONNX tensor: 284 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Relu_48 [Relu] outputs: [284 -> (-1, 512, 8, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Conv_49 [Conv] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 284 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 680 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 681 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Conv_49 [Conv] inputs: [284 -> (-1, 512, 8, -1)], [680 -> (512, 512, 3, 3)], [681 -> (512)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:444: Convolution input dimensions: (-1, 512, 8, -1) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:526: Using kernel: (3, 3), strides: (1, 1), padding: (1, 1), dilations: (1, 1), numOutputs: 512 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:527: Convolution output dimensions: (-1, 512, 8, -1) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Conv_49 for ONNX node: Conv_49 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 679 for ONNX tensor: 679 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Conv_49 [Conv] outputs: [679 -> (-1, 512, 8, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Relu_50 [Relu] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 679 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Relu_50 [Relu] inputs: [679 -> (-1, 512, 8, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Relu_50 for ONNX node: Relu_50 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 287 for ONNX tensor: 287 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Relu_50 [Relu] outputs: [287 -> (-1, 512, 8, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Conv_51 [Conv] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 287 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 683 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 684 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Conv_51 [Conv] inputs: [287 -> (-1, 512, 8, -1)], [683 -> (512, 512, 3, 3)], [684 -> (512)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:444: Convolution input dimensions: (-1, 512, 8, -1) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:526: Using kernel: (3, 3), strides: (1, 1), padding: (1, 1), dilations: (1, 1), numOutputs: 512 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:527: Convolution output dimensions: (-1, 512, 8, -1) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Conv_51 for ONNX node: Conv_51 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 682 for ONNX tensor: 682 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Conv_51 [Conv] outputs: [682 -> (-1, 512, 8, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Add_52 [Add] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 682 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 284 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Add_52 [Add] inputs: [682 -> (-1, 512, 8, -1)], [284 -> (-1, 512, 8, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Add_52 for ONNX node: Add_52 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 290 for ONNX tensor: 290 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Add_52 [Add] outputs: [290 -> (-1, 512, 8, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Relu_53 [Relu] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 290 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Relu_53 [Relu] inputs: [290 -> (-1, 512, 8, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Relu_53 for ONNX node: Relu_53 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 291 for ONNX tensor: 291 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Relu_53 [Relu] outputs: [291 -> (-1, 512, 8, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Conv_54 [Conv] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 291 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 686 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 687 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Conv_54 [Conv] inputs: [291 -> (-1, 512, 8, -1)], [686 -> (512, 512, 3, 3)], [687 -> (512)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:444: Convolution input dimensions: (-1, 512, 8, -1) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:526: Using kernel: (3, 3), strides: (1, 1), padding: (1, 1), dilations: (1, 1), numOutputs: 512 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:527: Convolution output dimensions: (-1, 512, 8, -1) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Conv_54 for ONNX node: Conv_54 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 685 for ONNX tensor: 685 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Conv_54 [Conv] outputs: [685 -> (-1, 512, 8, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Relu_55 [Relu] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 685 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Relu_55 [Relu] inputs: [685 -> (-1, 512, 8, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Relu_55 for ONNX node: Relu_55 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 294 for ONNX tensor: 294 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Relu_55 [Relu] outputs: [294 -> (-1, 512, 8, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Conv_56 [Conv] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 294 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 689 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 690 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Conv_56 [Conv] inputs: [294 -> (-1, 512, 8, -1)], [689 -> (512, 512, 3, 3)], [690 -> (512)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:444: Convolution input dimensions: (-1, 512, 8, -1) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:526: Using kernel: (3, 3), strides: (1, 1), padding: (1, 1), dilations: (1, 1), numOutputs: 512 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:527: Convolution output dimensions: (-1, 512, 8, -1) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Conv_56 for ONNX node: Conv_56 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 688 for ONNX tensor: 688 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Conv_56 [Conv] outputs: [688 -> (-1, 512, 8, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Relu_57 [Relu] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 688 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Relu_57 [Relu] inputs: [688 -> (-1, 512, 8, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Relu_57 for ONNX node: Relu_57 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 297 for ONNX tensor: 297 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Relu_57 [Relu] outputs: [297 -> (-1, 512, 8, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Conv_58 [Conv] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 297 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 692 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 693 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Conv_58 [Conv] inputs: [297 -> (-1, 512, 8, -1)], [692 -> (512, 512, 3, 3)], [693 -> (512)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:444: Convolution input dimensions: (-1, 512, 8, -1) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:526: Using kernel: (3, 3), strides: (1, 1), padding: (1, 1), dilations: (1, 1), numOutputs: 512 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:527: Convolution output dimensions: (-1, 512, 8, -1) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Conv_58 for ONNX node: Conv_58 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 691 for ONNX tensor: 691 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Conv_58 [Conv] outputs: [691 -> (-1, 512, 8, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Add_59 [Add] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 691 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 294 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Add_59 [Add] inputs: [691 -> (-1, 512, 8, -1)], [294 -> (-1, 512, 8, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Add_59 for ONNX node: Add_59 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 300 for ONNX tensor: 300 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Add_59 [Add] outputs: [300 -> (-1, 512, 8, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Relu_60 [Relu] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 300 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Relu_60 [Relu] inputs: [300 -> (-1, 512, 8, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Relu_60 for ONNX node: Relu_60 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 301 for ONNX tensor: 301 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Relu_60 [Relu] outputs: [301 -> (-1, 512, 8, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Conv_61 [Conv] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 301 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 695 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 696 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Conv_61 [Conv] inputs: [301 -> (-1, 512, 8, -1)], [695 -> (512, 512, 3, 3)], [696 -> (512)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:444: Convolution input dimensions: (-1, 512, 8, -1) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:526: Using kernel: (3, 3), strides: (1, 1), padding: (1, 1), dilations: (1, 1), numOutputs: 512 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:527: Convolution output dimensions: (-1, 512, 8, -1) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Conv_61 for ONNX node: Conv_61 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 694 for ONNX tensor: 694 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Conv_61 [Conv] outputs: [694 -> (-1, 512, 8, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Relu_62 [Relu] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 694 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Relu_62 [Relu] inputs: [694 -> (-1, 512, 8, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Relu_62 for ONNX node: Relu_62 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 304 for ONNX tensor: 304 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Relu_62 [Relu] outputs: [304 -> (-1, 512, 8, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Conv_63 [Conv] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 304 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 698 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 699 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Conv_63 [Conv] inputs: [304 -> (-1, 512, 8, -1)], [698 -> (512, 512, 3, 3)], [699 -> (512)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:444: Convolution input dimensions: (-1, 512, 8, -1) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:526: Using kernel: (3, 3), strides: (1, 1), padding: (1, 1), dilations: (1, 1), numOutputs: 512 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:527: Convolution output dimensions: (-1, 512, 8, -1) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Conv_63 for ONNX node: Conv_63 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 697 for ONNX tensor: 697 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Conv_63 [Conv] outputs: [697 -> (-1, 512, 8, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Add_64 [Add] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 697 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 301 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Add_64 [Add] inputs: [697 -> (-1, 512, 8, -1)], [301 -> (-1, 512, 8, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Add_64 for ONNX node: Add_64 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 307 for ONNX tensor: 307 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Add_64 [Add] outputs: [307 -> (-1, 512, 8, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Relu_65 [Relu] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 307 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Relu_65 [Relu] inputs: [307 -> (-1, 512, 8, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Relu_65 for ONNX node: Relu_65 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 308 for ONNX tensor: 308 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Relu_65 [Relu] outputs: [308 -> (-1, 512, 8, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Conv_66 [Conv] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 308 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 701 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 702 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Conv_66 [Conv] inputs: [308 -> (-1, 512, 8, -1)], [701 -> (512, 512, 3, 3)], [702 -> (512)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:444: Convolution input dimensions: (-1, 512, 8, -1) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:526: Using kernel: (3, 3), strides: (1, 1), padding: (1, 1), dilations: (1, 1), numOutputs: 512 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:527: Convolution output dimensions: (-1, 512, 8, -1) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Conv_66 for ONNX node: Conv_66 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 700 for ONNX tensor: 700 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Conv_66 [Conv] outputs: [700 -> (-1, 512, 8, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Relu_67 [Relu] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 700 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Relu_67 [Relu] inputs: [700 -> (-1, 512, 8, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Relu_67 for ONNX node: Relu_67 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 311 for ONNX tensor: 311 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Relu_67 [Relu] outputs: [311 -> (-1, 512, 8, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Conv_68 [Conv] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 311 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 704 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 705 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Conv_68 [Conv] inputs: [311 -> (-1, 512, 8, -1)], [704 -> (512, 512, 3, 3)], [705 -> (512)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:444: Convolution input dimensions: (-1, 512, 8, -1) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:526: Using kernel: (3, 3), strides: (1, 1), padding: (1, 1), dilations: (1, 1), numOutputs: 512 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:527: Convolution output dimensions: (-1, 512, 8, -1) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Conv_68 for ONNX node: Conv_68 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 703 for ONNX tensor: 703 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Conv_68 [Conv] outputs: [703 -> (-1, 512, 8, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Add_69 [Add] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 703 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 308 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Add_69 [Add] inputs: [703 -> (-1, 512, 8, -1)], [308 -> (-1, 512, 8, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Add_69 for ONNX node: Add_69 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 314 for ONNX tensor: 314 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Add_69 [Add] outputs: [314 -> (-1, 512, 8, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Relu_70 [Relu] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 314 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Relu_70 [Relu] inputs: [314 -> (-1, 512, 8, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Relu_70 for ONNX node: Relu_70 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 315 for ONNX tensor: 315 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Relu_70 [Relu] outputs: [315 -> (-1, 512, 8, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Conv_71 [Conv] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 315 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 707 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 708 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Conv_71 [Conv] inputs: [315 -> (-1, 512, 8, -1)], [707 -> (512, 512, 2, 2)], [708 -> (512)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:444: Convolution input dimensions: (-1, 512, 8, -1) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:526: Using kernel: (2, 2), strides: (2, 1), padding: (0, 1), dilations: (1, 1), numOutputs: 512 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:527: Convolution output dimensions: (-1, 512, 4, -1) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Conv_71 for ONNX node: Conv_71 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 706 for ONNX tensor: 706 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Conv_71 [Conv] outputs: [706 -> (-1, 512, 4, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Relu_72 [Relu] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 706 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Relu_72 [Relu] inputs: [706 -> (-1, 512, 4, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Relu_72 for ONNX node: Relu_72 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 318 for ONNX tensor: 318 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Relu_72 [Relu] outputs: [318 -> (-1, 512, 4, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Conv_73 [Conv] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 318 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 710 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 711 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Conv_73 [Conv] inputs: [318 -> (-1, 512, 4, -1)], [710 -> (512, 512, 2, 2)], [711 -> (512)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:444: Convolution input dimensions: (-1, 512, 4, -1) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:526: Using kernel: (2, 2), strides: (1, 1), padding: (0, 0), dilations: (1, 1), numOutputs: 512 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:527: Convolution output dimensions: (-1, 512, 3, -1) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Conv_73 for ONNX node: Conv_73 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 709 for ONNX tensor: 709 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Conv_73 [Conv] outputs: [709 -> (-1, 512, 3, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Relu_74 [Relu] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 709 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Relu_74 [Relu] inputs: [709 -> (-1, 512, 3, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Relu_74 for ONNX node: Relu_74 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 321 for ONNX tensor: 321 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Relu_74 [Relu] outputs: [321 -> (-1, 512, 3, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Transpose_75 [Transpose] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 321 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Transpose_75 [Transpose] inputs: [321 -> (-1, 512, 3, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Transpose_75 for ONNX node: Transpose_75 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 322 for ONNX tensor: 322 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Transpose_75 [Transpose] outputs: [322 -> (-1, -1, 512, 3)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: AveragePool_76 [AveragePool] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 322 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: AveragePool_76 [AveragePool] inputs: [322 -> (-1, -1, 512, 3)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: AveragePool_76 for ONNX node: AveragePool_76 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 323 for ONNX tensor: 323 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: AveragePool_76 [AveragePool] outputs: [323 -> (-1, -1, 512, 1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Shape_77 [Shape] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 323 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Shape_77 [Shape] inputs: [323 -> (-1, -1, 512, 1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Shape_77 for ONNX node: Shape_77 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 324 for ONNX tensor: 324 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Shape_77 [Shape] outputs: [324 -> (4)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Constant_78 [Constant] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Constant_78 [Constant] inputs: [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Constant_78 [Constant] outputs: [325 -> ()], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Gather_79 [Gather] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 324 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 325 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Gather_79 [Gather] inputs: [324 -> (4)], [325 -> ()], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:1015: Using Gather axis: 0 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Gather_79 for ONNX node: Gather_79 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 326 for ONNX tensor: 326 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Gather_79 [Gather] outputs: [326 -> ()], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Shape_80 [Shape] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 323 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Shape_80 [Shape] inputs: [323 -> (-1, -1, 512, 1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Shape_80 for ONNX node: Shape_80 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 327 for ONNX tensor: 327 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Shape_80 [Shape] outputs: [327 -> (4)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Constant_81 [Constant] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Constant_81 [Constant] inputs: [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Constant_81 [Constant] outputs: [328 -> ()], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Gather_82 [Gather] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 327 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 328 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Gather_82 [Gather] inputs: [327 -> (4)], [328 -> ()], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:1015: Using Gather axis: 0 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Gather_82 for ONNX node: Gather_82 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 329 for ONNX tensor: 329 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Gather_82 [Gather] outputs: [329 -> ()], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Shape_83 [Shape] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 323 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Shape_83 [Shape] inputs: [323 -> (-1, -1, 512, 1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Shape_83 for ONNX node: Shape_83 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 330 for ONNX tensor: 330 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Shape_83 [Shape] outputs: [330 -> (4)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Constant_84 [Constant] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Constant_84 [Constant] inputs: [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Constant_84 [Constant] outputs: [331 -> ()], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Gather_85 [Gather] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 330 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 331 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Gather_85 [Gather] inputs: [330 -> (4)], [331 -> ()], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:1015: Using Gather axis: 0 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Gather_85 for ONNX node: Gather_85 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 332 for ONNX tensor: 332 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Gather_85 [Gather] outputs: [332 -> ()], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Unsqueeze_86 [Unsqueeze] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 326 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Unsqueeze_86 [Unsqueeze] inputs: [326 -> ()], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/onnx2trt_utils.cpp:1679: Original shape: (), unsqueezing to: (1) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Unsqueeze_86 for ONNX node: Unsqueeze_86 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 333 for ONNX tensor: 333 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Unsqueeze_86 [Unsqueeze] outputs: [333 -> (1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Unsqueeze_87 [Unsqueeze] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 329 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Unsqueeze_87 [Unsqueeze] inputs: [329 -> ()], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/onnx2trt_utils.cpp:1679: Original shape: (), unsqueezing to: (1) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Unsqueeze_87 for ONNX node: Unsqueeze_87 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 334 for ONNX tensor: 334 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Unsqueeze_87 [Unsqueeze] outputs: [334 -> (1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Unsqueeze_88 [Unsqueeze] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 332 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Unsqueeze_88 [Unsqueeze] inputs: [332 -> ()], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/onnx2trt_utils.cpp:1679: Original shape: (), unsqueezing to: (1) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Unsqueeze_88 for ONNX node: Unsqueeze_88 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 335 for ONNX tensor: 335 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Unsqueeze_88 [Unsqueeze] outputs: [335 -> (1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Concat_89 [Concat] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 333 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 334 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 335 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Concat_89 [Concat] inputs: [333 -> (1)], [334 -> (1)], [335 -> (1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Concat_89 for ONNX node: Concat_89 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 336 for ONNX tensor: 336 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Concat_89 [Concat] outputs: [336 -> (3)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Reshape_90 [Reshape] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 323 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 336 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Reshape_90 [Reshape] inputs: [323 -> (-1, -1, 512, 1)], [336 -> (3)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Reshape_90 for ONNX node: Reshape_90 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 337 for ONNX tensor: 337 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Reshape_90 [Reshape] outputs: [337 -> (-1, -1, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Shape_91 [Shape] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 337 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Shape_91 [Shape] inputs: [337 -> (-1, -1, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Shape_91 for ONNX node: Shape_91 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 338 for ONNX tensor: 338 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Shape_91 [Shape] outputs: [338 -> (3)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Constant_92 [Constant] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Constant_92 [Constant] inputs: [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Constant_92 [Constant] outputs: [339 -> ()], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Gather_93 [Gather] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 338 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 339 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Gather_93 [Gather] inputs: [338 -> (3)], [339 -> ()], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:1015: Using Gather axis: 0 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Gather_93 for ONNX node: Gather_93 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 340 for ONNX tensor: 340 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Gather_93 [Gather] outputs: [340 -> ()], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Unsqueeze_94 [Unsqueeze] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 340 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Unsqueeze_94 [Unsqueeze] inputs: [340 -> ()], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/onnx2trt_utils.cpp:1679: Original shape: (), unsqueezing to: (1) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Unsqueeze_94 for ONNX node: Unsqueeze_94 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 344 for ONNX tensor: 344 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Unsqueeze_94 [Unsqueeze] outputs: [344 -> (1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Concat_95 [Concat] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 712 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 344 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 713 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Concat_95 [Concat] inputs: [712 -> (1)], [344 -> (1)], [713 -> (1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Concat_95 for ONNX node: Concat_95 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 346 for ONNX tensor: 346 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Concat_95 [Concat] outputs: [346 -> (3)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: ConstantOfShape_96 [ConstantOfShape] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 346 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: ConstantOfShape_96 [ConstantOfShape] inputs: [346 -> (3)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: ConstantOfShape_96 for ONNX node: ConstantOfShape_96 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 347 for ONNX tensor: 347 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: ConstantOfShape_96 [ConstantOfShape] outputs: [347 -> (-1, -1, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Transpose_97 [Transpose] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 337 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Transpose_97 [Transpose] inputs: [337 -> (-1, -1, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Transpose_97 for ONNX node: Transpose_97 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 348 for ONNX tensor: 348 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Transpose_97 [Transpose] outputs: [348 -> (-1, -1, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: LSTM_98 [LSTM] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 348 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 755 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 756 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 754 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 347 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 347 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: LSTM_98 [LSTM] inputs: [348 -> (-1, -1, -1)], [755 -> (2, 2048, 512)], [756 -> (2, 2048, 512)], [754 -> (2, 4096)], [optional input, not set], [347 -> (-1, -1, -1)], [347 -> (-1, -1, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:1838: Bias shape is: (2, 4096) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:1842: Reshaping bias to: (2, 2, 2048) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:1847: After reduction, bias shape is: (2, 1, 2048) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:1856: numDirectionsTensor shape: (1) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:1860: hiddenSizeTensor shape: (1) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:1862: batchSizeTensor shape: (1) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:1869: Gate output rank (equal to initial hidden/cell state rank): (3) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:1883: Initial hidden state shape: (-1, -1, -1) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:1886: Initial cell state shape: (-1, -1, -1) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:1888: Entering Loop [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/onnx2trt_utils.cpp:1679: Original shape: (_, _), unsqueezing to: (_, _, _) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/onnx2trt_utils.cpp:1679: Original shape: (_, _), unsqueezing to: (_, _, _) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:1921: Input shape: (-1, -1, -1) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:1925: Hidden state shape: (-1, -1, -1) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:1929: Cell state shape: (-1, -1, -1) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:1937: X(t) * W^T -> (-1, -1, 2048) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:1943: H(t-1) * R^T -> (-1, -1, 2048) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:1950: intermediate(t) -> (2, -1, 2048) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:2021: c(t) -> (-1, -1, -1) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:2030: C(t) -> (-1, -1, -1) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:2053: H(t) -> (-1, -1, -1) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:1207: Concatenated output shape: (-1, -1, -1) [11/26/2020-16:41:46] [E] [TRT] (Unnamed Layer* 163) [Slice]: slice size must be positive, size = [0,0,0] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:1210: Forward pass shape: () [11/26/2020-16:41:46] [E] [TRT] (Unnamed Layer* 164) [Slice]: slice size must be positive, size = [0,0,0] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:1216: Reverse pass shape: () [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: LSTM_98 for ONNX node: LSTM_98 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 465 for ONNX tensor: 465 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 466 for ONNX tensor: 466 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 467 for ONNX tensor: 467 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: LSTM_98 [LSTM] outputs: [465 -> (-1, -1, -1, -1)], [466 -> (-1, -1, -1)], [467 -> (-1, -1, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Transpose_99 [Transpose] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 465 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Transpose_99 [Transpose] inputs: [465 -> (-1, -1, -1, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Transpose_99 for ONNX node: Transpose_99 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 468 for ONNX tensor: 468 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Transpose_99 [Transpose] outputs: [468 -> (-1, -1, -1, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Constant_100 [Constant] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Constant_100 [Constant] inputs: [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Constant_100 [Constant] outputs: [469 -> (3)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Reshape_101 [Reshape] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 468 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 469 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Reshape_101 [Reshape] inputs: [468 -> (-1, -1, -1, -1)], [469 -> (3)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Reshape_101 for ONNX node: Reshape_101 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 470 for ONNX tensor: 470 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Reshape_101 [Reshape] outputs: [470 -> (-1, -1, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Transpose_102 [Transpose] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 470 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Transpose_102 [Transpose] inputs: [470 -> (-1, -1, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Transpose_102 for ONNX node: Transpose_102 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 471 for ONNX tensor: 471 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Transpose_102 [Transpose] outputs: [471 -> (-1, -1, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: MatMul_103 [MatMul] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 471 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 757 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: MatMul_103 [MatMul] inputs: [471 -> (-1, -1, -1)], [757 -> (1024, 512)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: MatMul_103 for ONNX node: MatMul_103 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 473 for ONNX tensor: 473 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: MatMul_103 [MatMul] outputs: [473 -> (-1, -1, 512)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Add_104 [Add] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 473 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: SequenceModeling.0.linear.bias [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Add_104 [Add] inputs: [473 -> (-1, -1, 512)], [SequenceModeling.0.linear.bias -> (512)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Add_104 for ONNX node: Add_104 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 474 for ONNX tensor: 474 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Add_104 [Add] outputs: [474 -> (-1, -1, 512)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Shape_105 [Shape] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 474 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Shape_105 [Shape] inputs: [474 -> (-1, -1, 512)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Shape_105 for ONNX node: Shape_105 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 475 for ONNX tensor: 475 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Shape_105 [Shape] outputs: [475 -> (3)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Constant_106 [Constant] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Constant_106 [Constant] inputs: [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Constant_106 [Constant] outputs: [476 -> ()], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Gather_107 [Gather] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 475 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 476 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Gather_107 [Gather] inputs: [475 -> (3)], [476 -> ()], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:1015: Using Gather axis: 0 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Gather_107 for ONNX node: Gather_107 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 477 for ONNX tensor: 477 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Gather_107 [Gather] outputs: [477 -> ()], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Unsqueeze_108 [Unsqueeze] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 477 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Unsqueeze_108 [Unsqueeze] inputs: [477 -> ()], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/onnx2trt_utils.cpp:1679: Original shape: (), unsqueezing to: (1) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Unsqueeze_108 for ONNX node: Unsqueeze_108 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 481 for ONNX tensor: 481 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Unsqueeze_108 [Unsqueeze] outputs: [481 -> (1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Concat_109 [Concat] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 758 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 481 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 759 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Concat_109 [Concat] inputs: [758 -> (1)], [481 -> (1)], [759 -> (1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Concat_109 for ONNX node: Concat_109 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 483 for ONNX tensor: 483 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Concat_109 [Concat] outputs: [483 -> (3)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: ConstantOfShape_110 [ConstantOfShape] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 483 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: ConstantOfShape_110 [ConstantOfShape] inputs: [483 -> (3)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: ConstantOfShape_110 for ONNX node: ConstantOfShape_110 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 484 for ONNX tensor: 484 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: ConstantOfShape_110 [ConstantOfShape] outputs: [484 -> (-1, -1, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Transpose_111 [Transpose] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 474 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Transpose_111 [Transpose] inputs: [474 -> (-1, -1, 512)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Transpose_111 for ONNX node: Transpose_111 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 485 for ONNX tensor: 485 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Transpose_111 [Transpose] outputs: [485 -> (-1, -1, 512)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: LSTM_112 [LSTM] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 485 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 801 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 802 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 800 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 484 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 484 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: LSTM_112 [LSTM] inputs: [485 -> (-1, -1, 512)], [801 -> (2, 2048, 512)], [802 -> (2, 2048, 512)], [800 -> (2, 4096)], [optional input, not set], [484 -> (-1, -1, -1)], [484 -> (-1, -1, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:1838: Bias shape is: (2, 4096) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:1842: Reshaping bias to: (2, 2, 2048) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:1847: After reduction, bias shape is: (2, 1, 2048) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:1856: numDirectionsTensor shape: (1) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:1860: hiddenSizeTensor shape: (1) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:1862: batchSizeTensor shape: (1) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:1869: Gate output rank (equal to initial hidden/cell state rank): (3) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:1883: Initial hidden state shape: (-1, -1, -1) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:1886: Initial cell state shape: (-1, -1, -1) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:1888: Entering Loop [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/onnx2trt_utils.cpp:1679: Original shape: (_, _), unsqueezing to: (_, _, _) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/onnx2trt_utils.cpp:1679: Original shape: (_, _), unsqueezing to: (_, _, _) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:1921: Input shape: (-1, -1, -1) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:1925: Hidden state shape: (-1, -1, -1) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:1929: Cell state shape: (-1, -1, -1) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:1937: X(t) * W^T -> (-1, -1, 2048) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:1943: H(t-1) * R^T -> (-1, -1, 2048) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:1950: intermediate(t) -> (2, -1, 2048) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:2021: c(t) -> (-1, -1, -1) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:2030: C(t) -> (-1, -1, -1) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:2053: H(t) -> (-1, -1, -1) [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:1207: Concatenated output shape: (-1, -1, -1) [11/26/2020-16:41:46] [E] [TRT] (Unnamed Layer* 251) [Slice]: slice size must be positive, size = [0,0,0] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:1210: Forward pass shape: () [11/26/2020-16:41:46] [E] [TRT] (Unnamed Layer* 252) [Slice]: slice size must be positive, size = [0,0,0] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/builtin_op_importers.cpp:1216: Reverse pass shape: () [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: LSTM_112 for ONNX node: LSTM_112 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 602 for ONNX tensor: 602 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 603 for ONNX tensor: 603 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 604 for ONNX tensor: 604 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: LSTM_112 [LSTM] outputs: [602 -> (-1, -1, -1, -1)], [603 -> (-1, -1, -1)], [604 -> (-1, -1, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Transpose_113 [Transpose] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 602 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Transpose_113 [Transpose] inputs: [602 -> (-1, -1, -1, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Transpose_113 for ONNX node: Transpose_113 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 605 for ONNX tensor: 605 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Transpose_113 [Transpose] outputs: [605 -> (-1, -1, -1, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Constant_114 [Constant] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Constant_114 [Constant] inputs: [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Constant_114 [Constant] outputs: [606 -> (3)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Reshape_115 [Reshape] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 605 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 606 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Reshape_115 [Reshape] inputs: [605 -> (-1, -1, -1, -1)], [606 -> (3)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Reshape_115 for ONNX node: Reshape_115 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 607 for ONNX tensor: 607 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Reshape_115 [Reshape] outputs: [607 -> (-1, -1, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Transpose_116 [Transpose] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 607 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Transpose_116 [Transpose] inputs: [607 -> (-1, -1, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Transpose_116 for ONNX node: Transpose_116 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 608 for ONNX tensor: 608 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Transpose_116 [Transpose] outputs: [608 -> (-1, -1, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: MatMul_117 [MatMul] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 608 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 803 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: MatMul_117 [MatMul] inputs: [608 -> (-1, -1, -1)], [803 -> (1024, 512)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: MatMul_117 for ONNX node: MatMul_117 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 610 for ONNX tensor: 610 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: MatMul_117 [MatMul] outputs: [610 -> (-1, -1, 512)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Add_118 [Add] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 610 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: SequenceModeling.1.linear.bias [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Add_118 [Add] inputs: [610 -> (-1, -1, 512)], [SequenceModeling.1.linear.bias -> (512)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Add_118 for ONNX node: Add_118 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 611 for ONNX tensor: 611 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Add_118 [Add] outputs: [611 -> (-1, -1, 512)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: MatMul_119 [MatMul] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 611 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 804 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: MatMul_119 [MatMul] inputs: [611 -> (-1, -1, 512)], [804 -> (512, 168)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: MatMul_119 for ONNX node: MatMul_119 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 613 for ONNX tensor: 613 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: MatMul_119 [MatMul] outputs: [613 -> (-1, -1, 168)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Add_120 [Add] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 613 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: Prediction.bias [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Add_120 [Add] inputs: [613 -> (-1, -1, 168)], [Prediction.bias -> (168)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Add_120 for ONNX node: Add_120 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: 614 for ONNX tensor: 614 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Add_120 [Add] outputs: [614 -> (-1, -1, 168)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:107: Parsing node: Softmax_121 [Softmax] [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:123: Searching for input: 614 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:129: Softmax_121 [Softmax] inputs: [614 -> (-1, -1, 168)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:132: Registering layer: Softmax_121 for ONNX node: Softmax_121 [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ImporterContext.hpp:107: Registering tensor: scores_1 for ONNX tensor: scores [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:181: Softmax_121 [Softmax] outputs: [scores -> (-1, -1, -1)], [11/26/2020-16:41:46] [V] [TRT] /home/sai/TensorRT/parsers/onnx/ModelImporter.cpp:513: Marking scores_1 as output: scores [11/26/2020-16:41:46] [W] [TRT] Calling isShapeTensor before the entire network is constructed may result in an inaccurate result. [11/26/2020-16:41:46] [W] [TRT] Calling isShapeTensor before the entire network is constructed may result in an inaccurate result. [11/26/2020-16:41:46] [W] [TRT] Calling isShapeTensor before the entire network is constructed may result in an inaccurate result. [11/26/2020-16:41:46] [W] [TRT] Calling isShapeTensor before the entire network is constructed may result in an inaccurate result. ----- Parsing of ONNX model ../models/text_recognition/latin.onnx is Done ---- [11/26/2020-16:41:46] [W] [TRT] Half2 support requested on hardware without native FP16 support, performance will be negatively affected. [11/26/2020-16:41:46] [V] [TRT] Applying generic optimizations to the graph for inference. [11/26/2020-16:41:46] [V] [TRT] Original: 197 layers [11/26/2020-16:41:46] [V] [TRT] After dead-layer removal: 197 layers [11/26/2020-16:41:46] [V] [TRT] After Myelin optimization: 141 layers [11/26/2020-16:41:46] [V] [TRT] After scale fusion: 141 layers [11/26/2020-16:41:46] [V] [TRT] Fusing ConstantOfShape_96 with (Unnamed Layer* 99) [Shuffle] [11/26/2020-16:41:46] [V] [TRT] Fusing (Unnamed Layer* 105) [Constant] with (Unnamed Layer* 106) [Shuffle] [11/26/2020-16:41:46] [V] [TRT] Fusing MatMul_103 with (Unnamed Layer* 174) [Shuffle] [11/26/2020-16:41:46] [V] [TRT] Fusing Add_104 with (Unnamed Layer* 177) [Shuffle] [11/26/2020-16:41:46] [V] [TRT] Fusing ConstantOfShape_110 with (Unnamed Layer* 187) [Shuffle] [11/26/2020-16:41:46] [V] [TRT] Fusing (Unnamed Layer* 193) [Constant] with (Unnamed Layer* 194) [Shuffle] [11/26/2020-16:41:46] [V] [TRT] Fusing MatMul_117 with (Unnamed Layer* 262) [Shuffle] [11/26/2020-16:41:46] [V] [TRT] Fusing Add_118 with (Unnamed Layer* 265) [Shuffle] [11/26/2020-16:41:46] [V] [TRT] Fusing MatMul_119 with (Unnamed Layer* 268) [Shuffle] [11/26/2020-16:41:46] [V] [TRT] Fusing Add_120 with (Unnamed Layer* 271) [Shuffle] [11/26/2020-16:41:46] [V] [TRT] Fusing Conv_0 with Relu_1 [11/26/2020-16:41:46] [V] [TRT] Fusing Conv_2 with Relu_3 [11/26/2020-16:41:46] [V] [TRT] Fusing Conv_8 with Add_9 [11/26/2020-16:41:46] [V] [TRT] Fusing Conv_5 with Relu_6 [11/26/2020-16:41:46] [V] [TRT] Fusing Conv_8 + Add_9 with Relu_10 [11/26/2020-16:41:46] [V] [TRT] Fusing Conv_11 with Relu_12 [11/26/2020-16:41:46] [V] [TRT] Fusing Conv_17 with Add_18 [11/26/2020-16:41:46] [V] [TRT] Fusing Conv_14 with Relu_15 [11/26/2020-16:41:46] [V] [TRT] Fusing Conv_17 + Add_18 with Relu_19 [11/26/2020-16:41:46] [V] [TRT] Fusing Conv_20 with Relu_21 [11/26/2020-16:41:46] [V] [TRT] Fusing Conv_22 with Add_23 [11/26/2020-16:41:46] [V] [TRT] Fusing Conv_22 + Add_23 with Relu_24 [11/26/2020-16:41:46] [V] [TRT] Fusing Conv_25 with Relu_26 [11/26/2020-16:41:46] [V] [TRT] Fusing Conv_31 with Add_32 [11/26/2020-16:41:46] [V] [TRT] Fusing Conv_28 with Relu_29 [11/26/2020-16:41:46] [V] [TRT] Fusing Conv_31 + Add_32 with Relu_33 [11/26/2020-16:41:46] [V] [TRT] Fusing Conv_34 with Relu_35 [11/26/2020-16:41:46] [V] [TRT] Fusing Conv_36 with Add_37 [11/26/2020-16:41:46] [V] [TRT] Fusing Conv_36 + Add_37 with Relu_38 [11/26/2020-16:41:46] [V] [TRT] Fusing Conv_39 with Relu_40 [11/26/2020-16:41:46] [V] [TRT] Fusing Conv_41 with Add_42 [11/26/2020-16:41:46] [V] [TRT] Fusing Conv_41 + Add_42 with Relu_43 [11/26/2020-16:41:46] [V] [TRT] Fusing Conv_44 with Relu_45 [11/26/2020-16:41:46] [V] [TRT] Fusing Conv_46 with Add_47 [11/26/2020-16:41:46] [V] [TRT] Fusing Conv_46 + Add_47 with Relu_48 [11/26/2020-16:41:46] [V] [TRT] Fusing Conv_49 with Relu_50 [11/26/2020-16:41:46] [V] [TRT] Fusing Conv_51 with Add_52 [11/26/2020-16:41:46] [V] [TRT] Fusing Conv_51 + Add_52 with Relu_53 [11/26/2020-16:41:46] [V] [TRT] Fusing Conv_54 with Relu_55 [11/26/2020-16:41:46] [V] [TRT] Fusing Conv_56 with Relu_57 [11/26/2020-16:41:46] [V] [TRT] Fusing Conv_58 with Add_59 [11/26/2020-16:41:46] [V] [TRT] Fusing Conv_58 + Add_59 with Relu_60 [11/26/2020-16:41:46] [V] [TRT] Fusing Conv_61 with Relu_62 [11/26/2020-16:41:46] [V] [TRT] Fusing Conv_63 with Add_64 [11/26/2020-16:41:46] [V] [TRT] Fusing Conv_63 + Add_64 with Relu_65 [11/26/2020-16:41:46] [V] [TRT] Fusing Conv_66 with Relu_67 [11/26/2020-16:41:46] [V] [TRT] Fusing Conv_68 with Add_69 [11/26/2020-16:41:46] [V] [TRT] Fusing Conv_68 + Add_69 with Relu_70 [11/26/2020-16:41:46] [V] [TRT] Fusing Conv_71 with Relu_72 [11/26/2020-16:41:46] [V] [TRT] Fusing Conv_73 with Relu_74 [11/26/2020-16:41:46] [V] [TRT] Fusing Reshape_90 with Transpose_97 [11/26/2020-16:41:46] [V] [TRT] Fusing Transpose_99 with Reshape_101 [11/26/2020-16:41:46] [V] [TRT] Fusing Transpose_99 + Reshape_101 with Transpose_102 [11/26/2020-16:41:46] [V] [TRT] Fusing Transpose_113 with Reshape_115 [11/26/2020-16:41:46] [V] [TRT] Fusing Transpose_113 + Reshape_115 with Transpose_116 [11/26/2020-16:41:46] [V] [TRT] After vertical fusions: 86 layers [11/26/2020-16:41:46] [V] [TRT] After final dead-layer removal: 82 layers [11/26/2020-16:41:46] [V] [TRT] After tensor merging: 82 layers [11/26/2020-16:41:46] [V] [TRT] Eliminating concatenation (Unnamed Layer* 167) [Concatenation] [11/26/2020-16:41:46] [V] [TRT] Generating copy for (Unnamed Layer* 165) [LoopOutput]_output to 465 [11/26/2020-16:41:46] [V] [TRT] Generating copy for (Unnamed Layer* 166) [LoopOutput]_output to 465 [11/26/2020-16:41:46] [V] [TRT] Eliminating concatenation (Unnamed Layer* 255) [Concatenation] [11/26/2020-16:41:46] [V] [TRT] Generating copy for (Unnamed Layer* 253) [LoopOutput]_output to 602 [11/26/2020-16:41:46] [V] [TRT] Generating copy for (Unnamed Layer* 254) [LoopOutput]_output to 602 [11/26/2020-16:41:46] [V] [TRT] After concat removal: 84 layers [11/26/2020-16:41:46] [V] [TRT] Graph construction and optimization completed in 0.0158246 seconds. [11/26/2020-16:41:47] [V] [TRT] Constructing optimization profile number 0 out of 1 *************** Autotuning format combination: -> Int32(1) *************** [11/26/2020-16:41:47] [V] [TRT] --------------- Timing Runner: (Unnamed Layer* 254) [LoopOutput][HostToDeviceLogicalLen] (ShapeHostToDevice) [11/26/2020-16:41:47] [V] [TRT] Tactic: 0 is the only option, timing skipped [11/26/2020-16:41:47] [V] [TRT] Fastest Tactic: 0 Time: 0 [11/26/2020-16:41:47] [V] [TRT] *************** Autotuning format combination: Int32(1) -> Int32() *************** [11/26/2020-16:41:47] [V] [TRT] --------------- Timing Runner: (Unnamed Layer* 254) [LoopOutput][ShuffleLogicalLen] (Shuffle) [11/26/2020-16:41:47] [V] [TRT] Tactic: 0 is the only option, timing skipped [11/26/2020-16:41:47] [V] [TRT] Fastest Tactic: 0 Time: 0 [11/26/2020-16:41:47] [V] [TRT] *************** Autotuning format combination: -> Int32(1) *************** [11/26/2020-16:41:47] [V] [TRT] --------------- Timing Runner: (Unnamed Layer* 207) [Iterator][HostToDeviceLogicalLen] (ShapeHostToDevice) [11/26/2020-16:41:47] [V] [TRT] Tactic: 0 is the only option, timing skipped [11/26/2020-16:41:47] [V] [TRT] Fastest Tactic: 0 Time: 0 [11/26/2020-16:41:47] [V] [TRT] *************** Autotuning format combination: Int32(1) -> Int32() *************** [11/26/2020-16:41:47] [V] [TRT] --------------- Timing Runner: (Unnamed Layer* 207) [Iterator][ShuffleLogicalLen] (Shuffle) [11/26/2020-16:41:47] [V] [TRT] Tactic: 0 is the only option, timing skipped [11/26/2020-16:41:47] [V] [TRT] Fastest Tactic: 0 Time: 0 [11/26/2020-16:41:47] [V] [TRT] *************** Autotuning format combination: -> Int32(1) *************** [11/26/2020-16:41:47] [V] [TRT] --------------- Timing Runner: (Unnamed Layer* 166) [LoopOutput][HostToDeviceLogicalLen] (ShapeHostToDevice) [11/26/2020-16:41:47] [V] [TRT] Tactic: 0 is the only option, timing skipped [11/26/2020-16:41:47] [V] [TRT] Fastest Tactic: 0 Time: 0 [11/26/2020-16:41:47] [V] [TRT] *************** Autotuning format combination: Int32(1) -> Int32() *************** [11/26/2020-16:41:47] [V] [TRT] --------------- Timing Runner: (Unnamed Layer* 166) [LoopOutput][ShuffleLogicalLen] (Shuffle) [11/26/2020-16:41:47] [V] [TRT] Tactic: 0 is the only option, timing skipped [11/26/2020-16:41:47] [V] [TRT] Fastest Tactic: 0 Time: 0 [11/26/2020-16:41:47] [V] [TRT] *************** Autotuning format combination: -> Int32(1) *************** [11/26/2020-16:41:47] [V] [TRT] --------------- Timing Runner: (Unnamed Layer* 119) [Iterator][HostToDeviceLogicalLen] (ShapeHostToDevice) [11/26/2020-16:41:47] [V] [TRT] Tactic: 0 is the only option, timing skipped [11/26/2020-16:41:47] [V] [TRT] Fastest Tactic: 0 Time: 0 [11/26/2020-16:41:47] [V] [TRT] *************** Autotuning format combination: Int32(1) -> Int32() *************** [11/26/2020-16:41:47] [V] [TRT] --------------- Timing Runner: (Unnamed Layer* 119) [Iterator][ShuffleLogicalLen] (Shuffle) [11/26/2020-16:41:47] [V] [TRT] Tactic: 0 is the only option, timing skipped [11/26/2020-16:41:47] [V] [TRT] Fastest Tactic: 0 Time: 0 [11/26/2020-16:41:47] [V] [TRT] *************** Autotuning format combination: -> Float(1,1,1) *************** [11/26/2020-16:41:47] [V] [TRT] *************** Autotuning format combination: -> Half(1,1,1) *************** [11/26/2020-16:41:47] [V] [TRT] *************** Autotuning format combination: -> Float(1,2048,4096) *************** [11/26/2020-16:41:47] [V] [TRT] *************** Autotuning format combination: -> Half(1,2048,4096) *************** [11/26/2020-16:41:47] [V] [TRT] *************** Autotuning format combination: -> Int32() *************** [11/26/2020-16:41:47] [V] [TRT] *************** Autotuning format combination: -> Float(1,512,524288) *************** [11/26/2020-16:41:47] [V] [TRT] *************** Autotuning format combination: -> Half(1,512,524288) *************** [11/26/2020-16:41:47] [V] [TRT] *************** Autotuning format combination: -> Float(1,512,512) *************** [11/26/2020-16:41:47] [V] [TRT] *************** Autotuning format combination: -> Half(1,512,512) *************** [11/26/2020-16:41:47] [V] [TRT] *************** Autotuning format combination: -> Float(1,1,1) *************** [11/26/2020-16:41:47] [V] [TRT] *************** Autotuning format combination: -> Half(1,1,1) *************** [11/26/2020-16:41:47] [V] [TRT] *************** Autotuning format combination: -> Float(1,2048,4096) *************** [11/26/2020-16:41:47] [V] [TRT] *************** Autotuning format combination: -> Half(1,2048,4096) *************** [11/26/2020-16:41:47] [V] [TRT] *************** Autotuning format combination: -> Int32() *************** [11/26/2020-16:41:47] [V] [TRT] *************** Autotuning format combination: -> Float(1,512,524288) *************** [11/26/2020-16:41:47] [V] [TRT] *************** Autotuning format combination: -> Half(1,512,524288) *************** [11/26/2020-16:41:47] [V] [TRT] *************** Autotuning format combination: -> Float(1,512,512) *************** [11/26/2020-16:41:47] [V] [TRT] *************** Autotuning format combination: -> Half(1,512,512) *************** [11/26/2020-16:41:47] [V] [TRT] *************** Autotuning format combination: -> Float(1,168,86016) *************** [11/26/2020-16:41:47] [V] [TRT] *************** Autotuning format combination: -> Half(1,168,86016) *************** [11/26/2020-16:41:47] [V] [TRT] *************** Autotuning format combination: -> Float(1,168,168) *************** [11/26/2020-16:41:47] [V] [TRT] *************** Autotuning format combination: -> Half(1,168,168) *************** [11/26/2020-16:41:47] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:41:47] [V] [TRT] Tactic: 0 time 0.005792 [11/26/2020-16:41:47] [V] [TRT] Fastest Tactic: 0 Time: 0.005792 [11/26/2020-16:41:47] [V] [TRT] *************** Autotuning format combination: Float(1,(# 3 (SHAPE input)),(* 64 (# 3 (SHAPE input))),(* 64 (# 3 (SHAPE input)))) -> Float(1,(# 3 (SHAPE input)),(* 64 (# 3 (SHAPE input))),(* 2048 (# 3 (SHAPE input)))) *************** [11/26/2020-16:41:47] [V] [TRT] --------------- Timing Runner: Conv_0 + Relu_1 (LegacySASSConvolution) [11/26/2020-16:41:47] [V] [TRT] Tactic: 0 time 0.015328 [11/26/2020-16:41:47] [V] [TRT] Tactic: 1 time 0.021504 [11/26/2020-16:41:47] [V] [TRT] Fastest Tactic: 0 Time: 0.015328 [11/26/2020-16:41:47] [V] [TRT] --------------- Timing Runner: Conv_0 + Relu_1 (FusedConvActConvolution) [11/26/2020-16:41:47] [V] [TRT] Tactic: 7 time 0.043808 [11/26/2020-16:41:47] [V] [TRT] Tactic: 29 time 0.050144 [11/26/2020-16:41:47] [V] [TRT] Tactic: 30 time 0.045632 [11/26/2020-16:41:47] [V] [TRT] Tactic: 43 time 0.049152 [11/26/2020-16:41:47] [V] [TRT] Tactic: 66 time 0.05104 [11/26/2020-16:41:47] [V] [TRT] Tactic: 90 time 0.048992 [11/26/2020-16:41:47] [V] [TRT] Tactic: 104 time 0.0472 [11/26/2020-16:41:47] [V] [TRT] Tactic: 130 time 0.042976 [11/26/2020-16:41:47] [V] [TRT] Tactic: 136 time 0.046976 [11/26/2020-16:41:47] [V] [TRT] Tactic: 144 time 0.050176 [11/26/2020-16:41:47] [V] [TRT] Tactic: 153 time 0.050176 [11/26/2020-16:41:47] [V] [TRT] Tactic: 156 time 0.04608 [11/26/2020-16:41:47] [V] [TRT] Fastest Tactic: 130 Time: 0.042976 [11/26/2020-16:41:47] [V] [TRT] --------------- Timing Runner: Conv_0 + Relu_1 (CaskConvolution) [11/26/2020-16:41:47] [V] [TRT] Conv_0 + Relu_1 (scudnn) Set Tactic Name: maxwell_scudnn_128x32_relu_medium_nn_v1 [11/26/2020-16:41:47] [V] [TRT] Tactic: 1062367460111450758 time 0.011936 [11/26/2020-16:41:47] [V] [TRT] Conv_0 + Relu_1 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148n_nt_v1 [11/26/2020-16:41:47] [V] [TRT] Tactic: 3827454225649558724 time 0.019456 [11/26/2020-16:41:47] [V] [TRT] Conv_0 + Relu_1 (scudnn) Set Tactic Name: maxwell_scudnn_128x64_relu_large_nn_v1 [11/26/2020-16:41:47] [V] [TRT] Tactic: 4337000649858996379 time 0.018432 [11/26/2020-16:41:47] [V] [TRT] Conv_0 + Relu_1 (scudnn) Set Tactic Name: maxwell_scudnn_128x128_relu_medium_nn_v1 [11/26/2020-16:41:47] [V] [TRT] Tactic: 4501471010995462441 time 0.027648 [11/26/2020-16:41:47] [V] [TRT] Conv_0 + Relu_1 (scudnn) Set Tactic Name: maxwell_scudnn_128x64_relu_small_nn_v1 [11/26/2020-16:41:47] [V] [TRT] Tactic: 5137655947464784826 time 0.018272 [11/26/2020-16:41:47] [V] [TRT] Conv_0 + Relu_1 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148m_nt_v1 [11/26/2020-16:41:47] [V] [TRT] Tactic: 5921334924264294896 time 0.019456 [11/26/2020-16:41:47] [V] [TRT] Conv_0 + Relu_1 (scudnn) Set Tactic Name: maxwell_scudnn_128x64_relu_medium_nn_v1 [11/26/2020-16:41:47] [V] [TRT] Tactic: 6645123197870846056 time 0.017408 [11/26/2020-16:41:47] [V] [TRT] Conv_0 + Relu_1 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148t_nt_v1 [11/26/2020-16:41:47] [V] [TRT] Tactic: 7852627285308570038 time 0.020352 [11/26/2020-16:41:47] [V] [TRT] Conv_0 + Relu_1 (scudnn) Set Tactic Name: maxwell_scudnn_128x128_relu_large_nn_v1 [11/26/2020-16:41:47] [V] [TRT] Tactic: -9137461792520977713 time 0.028672 [11/26/2020-16:41:47] [V] [TRT] Conv_0 + Relu_1 (scudnn) Set Tactic Name: maxwell_scudnn_128x32_relu_large_nn_v1 [11/26/2020-16:41:47] [V] [TRT] Tactic: -6092040395344634144 time 0.012736 [11/26/2020-16:41:47] [V] [TRT] Conv_0 + Relu_1 (scudnn) Set Tactic Name: maxwell_scudnn_128x32_relu_small_nn_v1 [11/26/2020-16:41:47] [V] [TRT] Tactic: -3456450830548107839 time 0.012288 [11/26/2020-16:41:47] [V] [TRT] Conv_0 + Relu_1 (scudnn) Set Tactic Name: maxwell_scudnn_128x128_relu_small_nn_v1 [11/26/2020-16:41:47] [V] [TRT] Tactic: -410470605513481746 time 0.027648 [11/26/2020-16:41:47] [V] [TRT] Fastest Tactic: 1062367460111450758 Time: 0.011936 [11/26/2020-16:41:47] [V] [TRT] --------------- Timing Runner: Conv_0 + Relu_1 (CudaConvolution) [11/26/2020-16:41:47] [V] [TRT] Tactic: 0 time 0.044736 [11/26/2020-16:41:47] [V] [TRT] Tactic: 1 time 0.039936 [11/26/2020-16:41:47] [V] [TRT] Tactic: 2 time 0.050176 [11/26/2020-16:41:47] [V] [TRT] Tactic: 5 time 0.677888 [11/26/2020-16:41:47] [V] [TRT] Tactic: 6 time 0.045056 [11/26/2020-16:41:47] [V] [TRT] Fastest Tactic: 1 Time: 0.039936 [11/26/2020-16:41:47] [V] [TRT] --------------- Timing Runner: Conv_0 + Relu_1 (CudaDepthwiseConvolution) [11/26/2020-16:41:47] [V] [TRT] CudaDepthwiseConvolution has no valid tactics for this config, skipping [11/26/2020-16:41:47] [V] [TRT] >>>>>>>>>>>>>>> Chose Runner Type: CaskConvolution Tactic: 1062367460111450758 [11/26/2020-16:41:47] [V] [TRT] Conv_0 + Relu_1 (scudnn) Set Tactic Name: maxwell_scudnn_128x32_relu_medium_nn_v1 [11/26/2020-16:41:47] [V] [TRT] [11/26/2020-16:41:47] [V] [TRT] *************** Autotuning format combination: Half(1,(# 3 (SHAPE input)),(* 64 (# 3 (SHAPE input))),(* 64 (# 3 (SHAPE input)))) -> Half(1,(# 3 (SHAPE input)),(* 64 (# 3 (SHAPE input))),(* 2048 (# 3 (SHAPE input)))) *************** [11/26/2020-16:41:47] [V] [TRT] --------------- Timing Runner: Conv_0 + Relu_1 (FusedConvActConvolution) [11/26/2020-16:41:47] [V] [TRT] FusedConvActConvolution has no valid tactics for this config, skipping [11/26/2020-16:41:47] [V] [TRT] --------------- Timing Runner: Conv_0 + Relu_1 (CaskConvolution) [11/26/2020-16:41:47] [V] [TRT] CaskConvolution has no valid tactics for this config, skipping [11/26/2020-16:41:47] [V] [TRT] --------------- Timing Runner: Conv_0 + Relu_1 (CudaConvolution) [11/26/2020-16:41:47] [V] [TRT] Tactic: 0 time 0.04096 [11/26/2020-16:41:47] [V] [TRT] Tactic: 1 time 0.037888 [11/26/2020-16:41:47] [V] [TRT] Tactic: 2 time 0.045056 [11/26/2020-16:41:47] [V] [TRT] Tactic: 5 time 0.72192 [11/26/2020-16:41:47] [V] [TRT] Tactic: 6 time 0.045056 [11/26/2020-16:41:47] [V] [TRT] Fastest Tactic: 1 Time: 0.037888 [11/26/2020-16:41:47] [V] [TRT] --------------- Timing Runner: Conv_0 + Relu_1 (CudaDepthwiseConvolution) [11/26/2020-16:41:47] [V] [TRT] CudaDepthwiseConvolution has no valid tactics for this config, skipping [11/26/2020-16:41:47] [V] [TRT] >>>>>>>>>>>>>>> Chose Runner Type: CudaConvolution Tactic: 1 [11/26/2020-16:41:47] [V] [TRT] [11/26/2020-16:41:47] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:41:47] [V] [TRT] Tactic: 0 time 0.005056 [11/26/2020-16:41:47] [V] [TRT] Fastest Tactic: 0 Time: 0.005056 [11/26/2020-16:41:47] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:41:47] [V] [TRT] Tactic: 0 time 0.003776 [11/26/2020-16:41:47] [V] [TRT] Fastest Tactic: 0 Time: 0.003776 [11/26/2020-16:41:47] [V] [TRT] *************** Autotuning format combination: Float(1,2048,4096) -> Float(1,2048,2048) *************** [11/26/2020-16:41:47] [V] [TRT] --------------- Timing Runner: (Unnamed Layer* 107) [Reduce] (Reduce) [11/26/2020-16:41:47] [V] [TRT] Tactic: 6 time 0.009088 [11/26/2020-16:41:47] [V] [TRT] Fastest Tactic: 6 Time: 0.009088 [11/26/2020-16:41:47] [V] [TRT] *************** Autotuning format combination: Half(1,2048,4096) -> Half(1,2048,2048) *************** [11/26/2020-16:41:47] [V] [TRT] --------------- Timing Runner: (Unnamed Layer* 107) [Reduce] (Reduce) [11/26/2020-16:41:47] [V] [TRT] Tactic: 6 time 0.009216 [11/26/2020-16:41:47] [V] [TRT] Fastest Tactic: 6 Time: 0.009216 [11/26/2020-16:41:47] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:41:47] [V] [TRT] Tactic: 0 time 0.004096 [11/26/2020-16:41:47] [V] [TRT] Fastest Tactic: 0 Time: 0.004096 [11/26/2020-16:41:47] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:41:47] [V] [TRT] Tactic: 0 time 0.003904 [11/26/2020-16:41:47] [V] [TRT] Fastest Tactic: 0 Time: 0.003904 [11/26/2020-16:41:47] [V] [TRT] *************** Autotuning format combination: Float(1,2048,4096) -> Float(1,2048,2048) *************** [11/26/2020-16:41:47] [V] [TRT] --------------- Timing Runner: (Unnamed Layer* 195) [Reduce] (Reduce) [11/26/2020-16:41:47] [V] [TRT] Tactic: 6 time 0.009824 [11/26/2020-16:41:47] [V] [TRT] Fastest Tactic: 6 Time: 0.009824 [11/26/2020-16:41:47] [V] [TRT] *************** Autotuning format combination: Half(1,2048,4096) -> Half(1,2048,2048) *************** [11/26/2020-16:41:47] [V] [TRT] --------------- Timing Runner: (Unnamed Layer* 195) [Reduce] (Reduce) [11/26/2020-16:41:47] [V] [TRT] Tactic: 6 time 0.009152 [11/26/2020-16:41:47] [V] [TRT] Fastest Tactic: 6 Time: 0.009152 [11/26/2020-16:41:48] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:41:48] [V] [TRT] Tactic: 0 time 0.026464 [11/26/2020-16:41:48] [V] [TRT] Fastest Tactic: 0 Time: 0.026464 [11/26/2020-16:41:48] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:41:48] [V] [TRT] Tactic: 0 time 0.018336 [11/26/2020-16:41:48] [V] [TRT] Fastest Tactic: 0 Time: 0.018336 [11/26/2020-16:41:48] [V] [TRT] *************** Autotuning format combination: Float(1,(# 3 (SHAPE input)),(* 64 (# 3 (SHAPE input))),(* 2048 (# 3 (SHAPE input)))) -> Float(1,(# 3 (SHAPE input)),(* 64 (# 3 (SHAPE input))),(* 4096 (# 3 (SHAPE input)))) *************** [11/26/2020-16:41:48] [V] [TRT] --------------- Timing Runner: Conv_2 + Relu_3 (LegacySASSConvolution) [11/26/2020-16:41:48] [V] [TRT] Tactic: 0 time 0.118464 [11/26/2020-16:41:48] [V] [TRT] Tactic: 1 time 0.069632 [11/26/2020-16:41:48] [V] [TRT] Fastest Tactic: 1 Time: 0.069632 [11/26/2020-16:41:48] [V] [TRT] --------------- Timing Runner: Conv_2 + Relu_3 (FusedConvActConvolution) [11/26/2020-16:41:48] [V] [TRT] Tactic: 7 time 0.136192 [11/26/2020-16:41:48] [V] [TRT] Tactic: 10 time 0.120768 [11/26/2020-16:41:48] [V] [TRT] Tactic: 14 time 0.116736 [11/26/2020-16:41:48] [V] [TRT] Tactic: 15 time 0.12144 [11/26/2020-16:41:48] [V] [TRT] Tactic: 25 time 0.115328 [11/26/2020-16:41:48] [V] [TRT] Tactic: 26 time 0.180544 [11/26/2020-16:41:48] [V] [TRT] Tactic: 29 time 0.125792 [11/26/2020-16:41:48] [V] [TRT] Tactic: 30 time 0.13312 [11/26/2020-16:41:48] [V] [TRT] Tactic: 33 time 0.136064 [11/26/2020-16:41:48] [V] [TRT] Tactic: 36 time 0.197536 [11/26/2020-16:41:48] [V] [TRT] Tactic: 39 time 0.14432 [11/26/2020-16:41:48] [V] [TRT] Tactic: 41 time 0.124064 [11/26/2020-16:41:48] [V] [TRT] Tactic: 42 time 0.277504 [11/26/2020-16:41:48] [V] [TRT] Tactic: 43 time 0.198656 [11/26/2020-16:41:48] [V] [TRT] Tactic: 45 time 0.116736 [11/26/2020-16:41:48] [V] [TRT] Tactic: 47 time 0.11264 [11/26/2020-16:41:48] [V] [TRT] Tactic: 52 time 0.190464 [11/26/2020-16:41:48] [V] [TRT] Tactic: 54 time 0.121856 [11/26/2020-16:41:48] [V] [TRT] Tactic: 56 time 0.186368 [11/26/2020-16:41:48] [V] [TRT] Tactic: 66 time 0.134144 [11/26/2020-16:41:48] [V] [TRT] Tactic: 76 time 0.123904 [11/26/2020-16:41:48] [V] [TRT] Tactic: 90 time 0.130048 [11/26/2020-16:41:48] [V] [TRT] Tactic: 93 time 0.123904 [11/26/2020-16:41:48] [V] [TRT] Tactic: 98 time 0.132096 [11/26/2020-16:41:48] [V] [TRT] Tactic: 104 time 0.131072 [11/26/2020-16:41:48] [V] [TRT] Tactic: 110 time 0.1536 [11/26/2020-16:41:48] [V] [TRT] Tactic: 119 time 0.1504 [11/26/2020-16:41:48] [V] [TRT] Tactic: 121 time 0.119808 [11/26/2020-16:41:48] [V] [TRT] Tactic: 130 time 0.131072 [11/26/2020-16:41:48] [V] [TRT] Tactic: 134 time 0.148256 [11/26/2020-16:41:48] [V] [TRT] Tactic: 136 time 0.134144 [11/26/2020-16:41:48] [V] [TRT] Tactic: 137 time 0.125952 [11/26/2020-16:41:48] [V] [TRT] Tactic: 139 time 0.121632 [11/26/2020-16:41:48] [V] [TRT] Tactic: 144 time 0.134144 [11/26/2020-16:41:48] [V] [TRT] Tactic: 149 time 0.169984 [11/26/2020-16:41:48] [V] [TRT] Tactic: 151 time 0.137216 [11/26/2020-16:41:48] [V] [TRT] Tactic: 152 time 0.118784 [11/26/2020-16:41:48] [V] [TRT] Tactic: 153 time 0.135168 [11/26/2020-16:41:48] [V] [TRT] Tactic: 156 time 0.125952 [11/26/2020-16:41:48] [V] [TRT] Tactic: 159 time 0.113664 [11/26/2020-16:41:48] [V] [TRT] Tactic: 162 time 0.139264 [11/26/2020-16:41:48] [V] [TRT] Tactic: 164 time 0.11264 [11/26/2020-16:41:48] [V] [TRT] Fastest Tactic: 47 Time: 0.11264 [11/26/2020-16:41:48] [V] [TRT] --------------- Timing Runner: Conv_2 + Relu_3 (CaskConvolution) [11/26/2020-16:41:48] [V] [TRT] Conv_2 + Relu_3 (scudnn) Set Tactic Name: maxwell_scudnn_128x32_relu_medium_nn_v1 [11/26/2020-16:41:48] [V] [TRT] Tactic: 1062367460111450758 time 0.106496 [11/26/2020-16:41:48] [V] [TRT] Conv_2 + Relu_3 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148n_nt_v1 [11/26/2020-16:41:48] [V] [TRT] Tactic: 3827454225649558724 time 0.062432 [11/26/2020-16:41:48] [V] [TRT] Conv_2 + Relu_3 (scudnn) Set Tactic Name: maxwell_scudnn_128x64_relu_large_nn_v1 [11/26/2020-16:41:48] [V] [TRT] Tactic: 4337000649858996379 time 0.093184 [11/26/2020-16:41:48] [V] [TRT] Conv_2 + Relu_3 (scudnn) Set Tactic Name: maxwell_scudnn_128x128_relu_medium_nn_v1 [11/26/2020-16:41:48] [V] [TRT] Tactic: 4501471010995462441 time 0.159744 [11/26/2020-16:41:48] [V] [TRT] Conv_2 + Relu_3 (scudnn) Set Tactic Name: maxwell_scudnn_128x64_relu_small_nn_v1 [11/26/2020-16:41:48] [V] [TRT] Tactic: 5137655947464784826 time 0.089088 [11/26/2020-16:41:48] [V] [TRT] Conv_2 + Relu_3 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148m_nt_v1 [11/26/2020-16:41:48] [V] [TRT] Tactic: 5921334924264294896 time 0.063488 [11/26/2020-16:41:48] [V] [TRT] Conv_2 + Relu_3 (scudnn) Set Tactic Name: maxwell_scudnn_128x64_relu_medium_nn_v1 [11/26/2020-16:41:48] [V] [TRT] Tactic: 6645123197870846056 time 0.09216 [11/26/2020-16:41:48] [V] [TRT] Conv_2 + Relu_3 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148t_nt_v1 [11/26/2020-16:41:48] [V] [TRT] Tactic: 7852627285308570038 time 0.06656 [11/26/2020-16:41:48] [V] [TRT] Conv_2 + Relu_3 (scudnn) Set Tactic Name: maxwell_scudnn_128x128_relu_large_nn_v1 [11/26/2020-16:41:48] [V] [TRT] Tactic: -9137461792520977713 time 0.159712 [11/26/2020-16:41:48] [V] [TRT] Conv_2 + Relu_3 (scudnn) Set Tactic Name: maxwell_scudnn_128x32_relu_large_nn_v1 [11/26/2020-16:41:48] [V] [TRT] Tactic: -6092040395344634144 time 0.109568 [11/26/2020-16:41:48] [V] [TRT] Conv_2 + Relu_3 (scudnn) Set Tactic Name: maxwell_scudnn_128x32_relu_small_nn_v1 [11/26/2020-16:41:48] [V] [TRT] Tactic: -3456450830548107839 time 0.101376 [11/26/2020-16:41:48] [V] [TRT] Conv_2 + Relu_3 (scudnn) Set Tactic Name: maxwell_scudnn_128x128_relu_small_nn_v1 [11/26/2020-16:41:48] [V] [TRT] Tactic: -410470605513481746 time 0.154624 [11/26/2020-16:41:48] [V] [TRT] Fastest Tactic: 3827454225649558724 Time: 0.062432 [11/26/2020-16:41:48] [V] [TRT] --------------- Timing Runner: Conv_2 + Relu_3 (CudaConvolution) [11/26/2020-16:41:48] [V] [TRT] Tactic: 0 time 0.246784 [11/26/2020-16:41:48] [V] [TRT] Tactic: 1 time 0.185344 [11/26/2020-16:41:48] [V] [TRT] Tactic: 2 time 0.34304 [11/26/2020-16:41:48] [V] [TRT] Tactic: 5 time 1.50323 [11/26/2020-16:41:48] [V] [TRT] Tactic: 6 time 0.145408 [11/26/2020-16:41:48] [V] [TRT] Fastest Tactic: 6 Time: 0.145408 [11/26/2020-16:41:48] [V] [TRT] --------------- Timing Runner: Conv_2 + Relu_3 (CudaDepthwiseConvolution) [11/26/2020-16:41:48] [V] [TRT] CudaDepthwiseConvolution has no valid tactics for this config, skipping [11/26/2020-16:41:48] [V] [TRT] >>>>>>>>>>>>>>> Chose Runner Type: CaskConvolution Tactic: 3827454225649558724 [11/26/2020-16:41:48] [V] [TRT] Conv_2 + Relu_3 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148n_nt_v1 [11/26/2020-16:41:48] [V] [TRT] [11/26/2020-16:41:48] [V] [TRT] *************** Autotuning format combination: Half(1,(# 3 (SHAPE input)),(* 64 (# 3 (SHAPE input))),(* 2048 (# 3 (SHAPE input)))) -> Half(1,(# 3 (SHAPE input)),(* 64 (# 3 (SHAPE input))),(* 4096 (# 3 (SHAPE input)))) *************** [11/26/2020-16:41:48] [V] [TRT] --------------- Timing Runner: Conv_2 + Relu_3 (FusedConvActConvolution) [11/26/2020-16:41:48] [V] [TRT] FusedConvActConvolution has no valid tactics for this config, skipping [11/26/2020-16:41:48] [V] [TRT] --------------- Timing Runner: Conv_2 + Relu_3 (CaskConvolution) [11/26/2020-16:41:48] [V] [TRT] CaskConvolution has no valid tactics for this config, skipping [11/26/2020-16:41:48] [V] [TRT] --------------- Timing Runner: Conv_2 + Relu_3 (CudaConvolution) [11/26/2020-16:41:48] [V] [TRT] Tactic: 0 time 0.2304 [11/26/2020-16:41:48] [V] [TRT] Tactic: 1 time 0.175104 [11/26/2020-16:41:48] [V] [TRT] Tactic: 2 time 0.280576 [11/26/2020-16:41:48] [V] [TRT] Tactic: 5 time 1.62506 [11/26/2020-16:41:48] [V] [TRT] Tactic: 6 time 0.119808 [11/26/2020-16:41:48] [V] [TRT] Fastest Tactic: 6 Time: 0.119808 [11/26/2020-16:41:48] [V] [TRT] --------------- Timing Runner: Conv_2 + Relu_3 (CudaDepthwiseConvolution) [11/26/2020-16:41:48] [V] [TRT] CudaDepthwiseConvolution has no valid tactics for this config, skipping [11/26/2020-16:41:48] [V] [TRT] >>>>>>>>>>>>>>> Chose Runner Type: CudaConvolution Tactic: 6 [11/26/2020-16:41:48] [V] [TRT] [11/26/2020-16:41:48] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:41:48] [V] [TRT] Tactic: 0 time 0.047104 [11/26/2020-16:41:48] [V] [TRT] Fastest Tactic: 0 Time: 0.047104 [11/26/2020-16:41:48] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:41:48] [V] [TRT] Tactic: 0 time 0.031744 [11/26/2020-16:41:48] [V] [TRT] Fastest Tactic: 0 Time: 0.031744 [11/26/2020-16:41:48] [V] [TRT] *************** Autotuning format combination: Float(1,(# 3 (SHAPE input)),(* 64 (# 3 (SHAPE input))),(* 4096 (# 3 (SHAPE input)))) -> Float(1,(+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) 1),(MUL_ADD 32 (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) 32),(MUL_ADD 2048 (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) 2048)) *************** [11/26/2020-16:41:48] [V] [TRT] --------------- Timing Runner: MaxPool_4 (Pooling) [11/26/2020-16:41:48] [V] [TRT] Tactic: -1 time 0.024512 [11/26/2020-16:41:48] [V] [TRT] Fastest Tactic: -1 Time: 0.024512 [11/26/2020-16:41:48] [V] [TRT] --------------- Timing Runner: MaxPool_4 (TiledPooling) [11/26/2020-16:41:48] [V] [TRT] Tactic: 5505281 time 0.02624 [11/26/2020-16:41:48] [V] [TRT] Tactic: 5570817 time 0.02144 [11/26/2020-16:41:48] [V] [TRT] Tactic: 5636353 time 0.021504 [11/26/2020-16:41:48] [V] [TRT] Tactic: 5701889 time 0.021504 [11/26/2020-16:41:48] [V] [TRT] Tactic: 5767425 time 0.021504 [11/26/2020-16:41:48] [V] [TRT] Tactic: 5832961 time 0.022144 [11/26/2020-16:41:48] [V] [TRT] Tactic: 5898497 time 0.021376 [11/26/2020-16:41:48] [V] [TRT] Tactic: 5964033 time 0.021728 [11/26/2020-16:41:48] [V] [TRT] Tactic: 6029569 time 0.030752 [11/26/2020-16:41:48] [V] [TRT] Tactic: 6095105 time 0.022112 [11/26/2020-16:41:48] [V] [TRT] Tactic: 6160641 time 0.022208 [11/26/2020-16:41:48] [V] [TRT] Tactic: 6226177 time 0.022272 [11/26/2020-16:41:48] [V] [TRT] Tactic: 6291713 time 0.02128 [11/26/2020-16:41:48] [V] [TRT] Tactic: 6357249 time 0.02192 [11/26/2020-16:41:48] [V] [TRT] Tactic: 6422785 time 0.022048 [11/26/2020-16:41:48] [V] [TRT] Tactic: 6488321 time 0.021312 [11/26/2020-16:41:48] [V] [TRT] Fastest Tactic: 6291713 Time: 0.02128 [11/26/2020-16:41:48] [V] [TRT] >>>>>>>>>>>>>>> Chose Runner Type: TiledPooling Tactic: 6291713 [11/26/2020-16:41:48] [V] [TRT] [11/26/2020-16:41:48] [V] [TRT] *************** Autotuning format combination: Half(1,(# 3 (SHAPE input)),(* 64 (# 3 (SHAPE input))),(* 4096 (# 3 (SHAPE input)))) -> Half(1,(+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) 1),(MUL_ADD 32 (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) 32),(MUL_ADD 2048 (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) 2048)) *************** [11/26/2020-16:41:48] [V] [TRT] --------------- Timing Runner: MaxPool_4 (Pooling) [11/26/2020-16:41:48] [V] [TRT] Tactic: -1 time 0.016384 [11/26/2020-16:41:48] [V] [TRT] Fastest Tactic: -1 Time: 0.016384 [11/26/2020-16:41:48] [V] [TRT] --------------- Timing Runner: MaxPool_4 (TiledPooling) [11/26/2020-16:41:48] [V] [TRT] TiledPooling has no valid tactics for this config, skipping [11/26/2020-16:41:48] [V] [TRT] >>>>>>>>>>>>>>> Chose Runner Type: Pooling Tactic: -1 [11/26/2020-16:41:48] [V] [TRT] [11/26/2020-16:41:48] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:41:48] [V] [TRT] Tactic: 0 time 0.015296 [11/26/2020-16:41:48] [V] [TRT] Fastest Tactic: 0 Time: 0.015296 [11/26/2020-16:41:48] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:41:48] [V] [TRT] Tactic: 0 time 0.012032 [11/26/2020-16:41:48] [V] [TRT] Fastest Tactic: 0 Time: 0.012032 [11/26/2020-16:41:48] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:41:48] [V] [TRT] Tactic: 0 time 0.014336 [11/26/2020-16:41:48] [V] [TRT] Fastest Tactic: 0 Time: 0.014336 [11/26/2020-16:41:48] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:41:48] [V] [TRT] Tactic: 0 time 0.012096 [11/26/2020-16:41:48] [V] [TRT] Fastest Tactic: 0 Time: 0.012096 [11/26/2020-16:41:48] [V] [TRT] *************** Autotuning format combination: Float(1,(+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) 1),(MUL_ADD 32 (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) 32),(MUL_ADD 2048 (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) 2048)) -> Float(1,(+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) 1),(MUL_ADD 32 (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) 32),(MUL_ADD 4096 (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) 4096)) *************** [11/26/2020-16:41:48] [V] [TRT] --------------- Timing Runner: Conv_5 + Relu_6 (LegacySASSConvolution) [11/26/2020-16:41:48] [V] [TRT] Tactic: 0 time 0.106304 [11/26/2020-16:41:48] [V] [TRT] Tactic: 1 time 0.059104 [11/26/2020-16:41:48] [V] [TRT] Fastest Tactic: 1 Time: 0.059104 [11/26/2020-16:41:48] [V] [TRT] --------------- Timing Runner: Conv_5 + Relu_6 (FusedConvActConvolution) [11/26/2020-16:41:48] [V] [TRT] Tactic: 7 time 0.119808 [11/26/2020-16:41:48] [V] [TRT] Tactic: 10 time 0.110528 [11/26/2020-16:41:48] [V] [TRT] Tactic: 14 time 0.099328 [11/26/2020-16:41:48] [V] [TRT] Tactic: 15 time 0.107488 [11/26/2020-16:41:48] [V] [TRT] Tactic: 25 time 0.095392 [11/26/2020-16:41:48] [V] [TRT] Tactic: 26 time 0.148704 [11/26/2020-16:41:48] [V] [TRT] Tactic: 29 time 0.112512 [11/26/2020-16:41:48] [V] [TRT] Tactic: 30 time 0.113664 [11/26/2020-16:41:48] [V] [TRT] Tactic: 33 time 0.128928 [11/26/2020-16:41:48] [V] [TRT] Tactic: 36 time 0.177024 [11/26/2020-16:41:48] [V] [TRT] Tactic: 39 time 0.113792 [11/26/2020-16:41:49] [V] [TRT] Tactic: 41 time 0.114368 [11/26/2020-16:41:49] [V] [TRT] Tactic: 42 time 0.164864 [11/26/2020-16:41:49] [V] [TRT] Tactic: 43 time 0.132768 [11/26/2020-16:41:49] [V] [TRT] Tactic: 45 time 0.098304 [11/26/2020-16:41:49] [V] [TRT] Tactic: 47 time 0.10512 [11/26/2020-16:41:49] [V] [TRT] Tactic: 52 time 0.11568 [11/26/2020-16:41:49] [V] [TRT] Tactic: 54 time 0.103392 [11/26/2020-16:41:49] [V] [TRT] Tactic: 56 time 0.14016 [11/26/2020-16:41:49] [V] [TRT] Tactic: 66 time 0.125888 [11/26/2020-16:41:49] [V] [TRT] Tactic: 76 time 0.106336 [11/26/2020-16:41:49] [V] [TRT] Tactic: 90 time 0.117696 [11/26/2020-16:41:49] [V] [TRT] Tactic: 93 time 0.11264 [11/26/2020-16:41:49] [V] [TRT] Tactic: 98 time 0.117664 [11/26/2020-16:41:49] [V] [TRT] Tactic: 104 time 0.116736 [11/26/2020-16:41:49] [V] [TRT] Tactic: 110 time 0.123904 [11/26/2020-16:41:49] [V] [TRT] Tactic: 119 time 0.122816 [11/26/2020-16:41:49] [V] [TRT] Tactic: 121 time 0.1024 [11/26/2020-16:41:49] [V] [TRT] Tactic: 130 time 0.115616 [11/26/2020-16:41:49] [V] [TRT] Tactic: 134 time 0.121504 [11/26/2020-16:41:49] [V] [TRT] Tactic: 136 time 0.120768 [11/26/2020-16:41:49] [V] [TRT] Tactic: 137 time 0.107328 [11/26/2020-16:41:49] [V] [TRT] Tactic: 139 time 0.1024 [11/26/2020-16:41:49] [V] [TRT] Tactic: 144 time 0.125856 [11/26/2020-16:41:49] [V] [TRT] Tactic: 149 time 0.11776 [11/26/2020-16:41:49] [V] [TRT] Tactic: 151 time 0.116384 [11/26/2020-16:41:49] [V] [TRT] Tactic: 152 time 0.106208 [11/26/2020-16:41:49] [V] [TRT] Tactic: 153 time 0.12592 [11/26/2020-16:41:49] [V] [TRT] Tactic: 156 time 0.113312 [11/26/2020-16:41:49] [V] [TRT] Tactic: 159 time 0.099296 [11/26/2020-16:41:49] [V] [TRT] Tactic: 162 time 0.119808 [11/26/2020-16:41:49] [V] [TRT] Tactic: 164 time 0.096256 [11/26/2020-16:41:49] [V] [TRT] Fastest Tactic: 25 Time: 0.095392 [11/26/2020-16:41:49] [V] [TRT] --------------- Timing Runner: Conv_5 + Relu_6 (CaskConvolution) [11/26/2020-16:41:49] [V] [TRT] Conv_5 + Relu_6 (scudnn) Set Tactic Name: maxwell_scudnn_128x32_relu_medium_nn_v1 [11/26/2020-16:41:49] [V] [TRT] Tactic: 1062367460111450758 time 0.106208 [11/26/2020-16:41:49] [V] [TRT] Conv_5 + Relu_6 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148n_nt_v1 [11/26/2020-16:41:49] [V] [TRT] Tactic: 3827454225649558724 time 0.050176 [11/26/2020-16:41:49] [V] [TRT] Conv_5 + Relu_6 (scudnn) Set Tactic Name: maxwell_scudnn_128x64_relu_large_nn_v1 [11/26/2020-16:41:49] [V] [TRT] Tactic: 4337000649858996379 time 0.089088 [11/26/2020-16:41:49] [V] [TRT] Conv_5 + Relu_6 (scudnn) Set Tactic Name: maxwell_scudnn_128x128_relu_medium_nn_v1 [11/26/2020-16:41:49] [V] [TRT] Tactic: 4501471010995462441 time 0.103392 [11/26/2020-16:41:49] [V] [TRT] Conv_5 + Relu_6 (scudnn) Set Tactic Name: maxwell_scudnn_128x64_relu_small_nn_v1 [11/26/2020-16:41:49] [V] [TRT] Tactic: 5137655947464784826 time 0.086016 [11/26/2020-16:41:49] [V] [TRT] Conv_5 + Relu_6 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148m_nt_v1 [11/26/2020-16:41:49] [V] [TRT] Tactic: 5921334924264294896 time 0.060416 [11/26/2020-16:41:49] [V] [TRT] Conv_5 + Relu_6 (scudnn) Set Tactic Name: maxwell_scudnn_128x64_relu_medium_nn_v1 [11/26/2020-16:41:49] [V] [TRT] Tactic: 6645123197870846056 time 0.0888 [11/26/2020-16:41:49] [V] [TRT] Conv_5 + Relu_6 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148t_nt_v1 [11/26/2020-16:41:49] [V] [TRT] Tactic: 7852627285308570038 time 0.054272 [11/26/2020-16:41:49] [V] [TRT] Conv_5 + Relu_6 (scudnn) Set Tactic Name: maxwell_scudnn_128x128_relu_large_nn_v1 [11/26/2020-16:41:49] [V] [TRT] Tactic: -9137461792520977713 time 0.10384 [11/26/2020-16:41:49] [V] [TRT] Conv_5 + Relu_6 (scudnn) Set Tactic Name: maxwell_scudnn_128x32_relu_large_nn_v1 [11/26/2020-16:41:49] [V] [TRT] Tactic: -6092040395344634144 time 0.106496 [11/26/2020-16:41:49] [V] [TRT] Conv_5 + Relu_6 (scudnn) Set Tactic Name: maxwell_scudnn_128x32_relu_small_nn_v1 [11/26/2020-16:41:49] [V] [TRT] Tactic: -3456450830548107839 time 0.100352 [11/26/2020-16:41:49] [V] [TRT] Conv_5 + Relu_6 (scudnn) Set Tactic Name: maxwell_scudnn_128x128_relu_small_nn_v1 [11/26/2020-16:41:49] [V] [TRT] Tactic: -410470605513481746 time 0.101376 [11/26/2020-16:41:49] [V] [TRT] Fastest Tactic: 3827454225649558724 Time: 0.050176 [11/26/2020-16:41:49] [V] [TRT] --------------- Timing Runner: Conv_5 + Relu_6 (CudaConvolution) [11/26/2020-16:41:49] [V] [TRT] Tactic: 0 time 0.154624 [11/26/2020-16:41:49] [V] [TRT] Tactic: 1 time 0.131584 [11/26/2020-16:41:49] [V] [TRT] Tactic: 2 time 0.206848 [11/26/2020-16:41:49] [V] [TRT] Tactic: 5 time 1.54317 [11/26/2020-16:41:49] [V] [TRT] Tactic: 6 time 0.080896 [11/26/2020-16:41:49] [V] [TRT] Fastest Tactic: 6 Time: 0.080896 [11/26/2020-16:41:49] [V] [TRT] --------------- Timing Runner: Conv_5 + Relu_6 (CudaDepthwiseConvolution) [11/26/2020-16:41:49] [V] [TRT] CudaDepthwiseConvolution has no valid tactics for this config, skipping [11/26/2020-16:41:49] [V] [TRT] >>>>>>>>>>>>>>> Chose Runner Type: CaskConvolution Tactic: 3827454225649558724 [11/26/2020-16:41:49] [V] [TRT] Conv_5 + Relu_6 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148n_nt_v1 [11/26/2020-16:41:49] [V] [TRT] [11/26/2020-16:41:49] [V] [TRT] *************** Autotuning format combination: Half(1,(+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) 1),(MUL_ADD 32 (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) 32),(MUL_ADD 2048 (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) 2048)) -> Half(1,(+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) 1),(MUL_ADD 32 (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) 32),(MUL_ADD 4096 (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) 4096)) *************** [11/26/2020-16:41:49] [V] [TRT] --------------- Timing Runner: Conv_5 + Relu_6 (FusedConvActConvolution) [11/26/2020-16:41:49] [V] [TRT] FusedConvActConvolution has no valid tactics for this config, skipping [11/26/2020-16:41:49] [V] [TRT] --------------- Timing Runner: Conv_5 + Relu_6 (CaskConvolution) [11/26/2020-16:41:49] [V] [TRT] CaskConvolution has no valid tactics for this config, skipping [11/26/2020-16:41:49] [V] [TRT] --------------- Timing Runner: Conv_5 + Relu_6 (CudaConvolution) [11/26/2020-16:41:49] [V] [TRT] Tactic: 0 time 0.146432 [11/26/2020-16:41:49] [V] [TRT] Tactic: 1 time 0.131072 [11/26/2020-16:41:49] [V] [TRT] Tactic: 2 time 0.196608 [11/26/2020-16:41:49] [V] [TRT] Tactic: 5 time 1.76829 [11/26/2020-16:41:49] [V] [TRT] Tactic: 6 time 0.077824 [11/26/2020-16:41:49] [V] [TRT] Fastest Tactic: 6 Time: 0.077824 [11/26/2020-16:41:49] [V] [TRT] --------------- Timing Runner: Conv_5 + Relu_6 (CudaDepthwiseConvolution) [11/26/2020-16:41:49] [V] [TRT] CudaDepthwiseConvolution has no valid tactics for this config, skipping [11/26/2020-16:41:49] [V] [TRT] >>>>>>>>>>>>>>> Chose Runner Type: CudaConvolution Tactic: 6 [11/26/2020-16:41:49] [V] [TRT] [11/26/2020-16:41:49] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:41:49] [V] [TRT] Tactic: 0 time 0.026432 [11/26/2020-16:41:49] [V] [TRT] Fastest Tactic: 0 Time: 0.026432 [11/26/2020-16:41:49] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:41:49] [V] [TRT] Tactic: 0 time 0.018208 [11/26/2020-16:41:49] [V] [TRT] Fastest Tactic: 0 Time: 0.018208 [11/26/2020-16:41:49] [V] [TRT] *************** Autotuning format combination: Float(1,(+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) 1),(MUL_ADD 32 (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) 32),(MUL_ADD 4096 (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) 4096)) -> Float(1,(+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) 1),(MUL_ADD 32 (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) 32),(MUL_ADD 4096 (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) 4096)) *************** [11/26/2020-16:41:49] [V] [TRT] --------------- Timing Runner: Conv_7 (LegacySASSConvolution) [11/26/2020-16:41:49] [V] [TRT] Tactic: 0 time 0.19968 [11/26/2020-16:41:49] [V] [TRT] Tactic: 1 time 0.1 [11/26/2020-16:41:49] [V] [TRT] Fastest Tactic: 1 Time: 0.1 [11/26/2020-16:41:49] [V] [TRT] --------------- Timing Runner: Conv_7 (FusedConvActConvolution) [11/26/2020-16:41:49] [V] [TRT] Tactic: 7 time 0.204736 [11/26/2020-16:41:49] [V] [TRT] Tactic: 10 time 0.194528 [11/26/2020-16:41:49] [V] [TRT] Tactic: 14 time 0.17168 [11/26/2020-16:41:49] [V] [TRT] Tactic: 15 time 0.186144 [11/26/2020-16:41:49] [V] [TRT] Tactic: 25 time 0.171968 [11/26/2020-16:41:49] [V] [TRT] Tactic: 26 time 0.249856 [11/26/2020-16:41:49] [V] [TRT] Tactic: 29 time 0.182272 [11/26/2020-16:41:49] [V] [TRT] Tactic: 30 time 0.18928 [11/26/2020-16:41:49] [V] [TRT] Tactic: 33 time 0.23648 [11/26/2020-16:41:49] [V] [TRT] Tactic: 36 time 0.314144 [11/26/2020-16:41:49] [V] [TRT] Tactic: 39 time 0.198656 [11/26/2020-16:41:49] [V] [TRT] Tactic: 41 time 0.197472 [11/26/2020-16:41:49] [V] [TRT] Tactic: 42 time 0.29696 [11/26/2020-16:41:49] [V] [TRT] Tactic: 43 time 0.223072 [11/26/2020-16:41:49] [V] [TRT] Tactic: 45 time 0.175104 [11/26/2020-16:41:49] [V] [TRT] Tactic: 47 time 0.188416 [11/26/2020-16:41:49] [V] [TRT] Tactic: 52 time 0.198656 [11/26/2020-16:41:49] [V] [TRT] Tactic: 54 time 0.18432 [11/26/2020-16:41:49] [V] [TRT] Tactic: 56 time 0.248832 [11/26/2020-16:41:49] [V] [TRT] Tactic: 66 time 0.205824 [11/26/2020-16:41:49] [V] [TRT] Tactic: 76 time 0.187392 [11/26/2020-16:41:49] [V] [TRT] Tactic: 90 time 0.19968 [11/26/2020-16:41:49] [V] [TRT] Tactic: 93 time 0.19456 [11/26/2020-16:41:49] [V] [TRT] Tactic: 98 time 0.208896 [11/26/2020-16:41:49] [V] [TRT] Tactic: 104 time 0.194048 [11/26/2020-16:41:49] [V] [TRT] Tactic: 110 time 0.207872 [11/26/2020-16:41:49] [V] [TRT] Tactic: 119 time 0.210944 [11/26/2020-16:41:49] [V] [TRT] Tactic: 121 time 0.18112 [11/26/2020-16:41:49] [V] [TRT] Tactic: 130 time 0.194496 [11/26/2020-16:41:49] [V] [TRT] Tactic: 134 time 0.203776 [11/26/2020-16:41:49] [V] [TRT] Tactic: 136 time 0.19968 [11/26/2020-16:41:49] [V] [TRT] Tactic: 137 time 0.189184 [11/26/2020-16:41:49] [V] [TRT] Tactic: 139 time 0.182272 [11/26/2020-16:41:49] [V] [TRT] Tactic: 144 time 0.206592 [11/26/2020-16:41:49] [V] [TRT] Tactic: 149 time 0.212928 [11/26/2020-16:41:49] [V] [TRT] Tactic: 151 time 0.19968 [11/26/2020-16:41:49] [V] [TRT] Tactic: 152 time 0.187392 [11/26/2020-16:41:49] [V] [TRT] Tactic: 153 time 0.205824 [11/26/2020-16:41:49] [V] [TRT] Tactic: 156 time 0.181248 [11/26/2020-16:41:49] [V] [TRT] Tactic: 159 time 0.178944 [11/26/2020-16:41:49] [V] [TRT] Tactic: 162 time 0.201728 [11/26/2020-16:41:49] [V] [TRT] Tactic: 164 time 0.173056 [11/26/2020-16:41:49] [V] [TRT] Fastest Tactic: 14 Time: 0.17168 [11/26/2020-16:41:49] [V] [TRT] --------------- Timing Runner: Conv_7 (CaskConvolution) [11/26/2020-16:41:49] [V] [TRT] Conv_7 (scudnn) Set Tactic Name: maxwell_scudnn_128x32_relu_medium_nn_v1 [11/26/2020-16:41:49] [V] [TRT] Tactic: 1062367460111450758 time 0.187392 [11/26/2020-16:41:49] [V] [TRT] Conv_7 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148n_nt_v1 [11/26/2020-16:41:49] [V] [TRT] Tactic: 3827454225649558724 time 0.093184 [11/26/2020-16:41:49] [V] [TRT] Conv_7 (scudnn) Set Tactic Name: maxwell_scudnn_128x64_relu_large_nn_v1 [11/26/2020-16:41:49] [V] [TRT] Tactic: 4337000649858996379 time 0.165888 [11/26/2020-16:41:49] [V] [TRT] Conv_7 (scudnn) Set Tactic Name: maxwell_scudnn_128x128_relu_medium_nn_v1 [11/26/2020-16:41:49] [V] [TRT] Tactic: 4501471010995462441 time 0.19456 [11/26/2020-16:41:49] [V] [TRT] Conv_7 (scudnn) Set Tactic Name: maxwell_scudnn_128x64_relu_small_nn_v1 [11/26/2020-16:41:49] [V] [TRT] Tactic: 5137655947464784826 time 0.15872 [11/26/2020-16:41:49] [V] [TRT] Conv_7 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148m_nt_v1 [11/26/2020-16:41:49] [V] [TRT] Tactic: 5921334924264294896 time 0.088064 [11/26/2020-16:41:49] [V] [TRT] Conv_7 (scudnn) Set Tactic Name: maxwell_scudnn_128x64_relu_medium_nn_v1 [11/26/2020-16:41:49] [V] [TRT] Tactic: 6645123197870846056 time 0.162784 [11/26/2020-16:41:49] [V] [TRT] Conv_7 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148t_nt_v1 [11/26/2020-16:41:49] [V] [TRT] Tactic: 7852627285308570038 time 0.094208 [11/26/2020-16:41:49] [V] [TRT] Conv_7 (scudnn) Set Tactic Name: maxwell_scudnn_128x128_relu_large_nn_v1 [11/26/2020-16:41:49] [V] [TRT] Tactic: -9137461792520977713 time 0.197632 [11/26/2020-16:41:49] [V] [TRT] Conv_7 (scudnn) Set Tactic Name: maxwell_scudnn_128x32_relu_large_nn_v1 [11/26/2020-16:41:49] [V] [TRT] Tactic: -6092040395344634144 time 0.193536 [11/26/2020-16:41:49] [V] [TRT] Conv_7 (scudnn) Set Tactic Name: maxwell_scudnn_128x32_relu_small_nn_v1 [11/26/2020-16:41:49] [V] [TRT] Tactic: -3456450830548107839 time 0.177152 [11/26/2020-16:41:49] [V] [TRT] Conv_7 (scudnn) Set Tactic Name: maxwell_scudnn_128x128_relu_small_nn_v1 [11/26/2020-16:41:49] [V] [TRT] Tactic: -410470605513481746 time 0.191488 [11/26/2020-16:41:49] [V] [TRT] Fastest Tactic: 5921334924264294896 Time: 0.088064 [11/26/2020-16:41:49] [V] [TRT] --------------- Timing Runner: Conv_7 (CudaConvolution) [11/26/2020-16:41:49] [V] [TRT] Tactic: 0 time 0.260096 [11/26/2020-16:41:49] [V] [TRT] Tactic: 1 time 0.219136 [11/26/2020-16:41:49] [V] [TRT] Tactic: 2 time 0.324608 [11/26/2020-16:41:49] [V] [TRT] Tactic: 5 time 2.85594 [11/26/2020-16:41:49] [V] [TRT] Tactic: 6 time 0.10944 [11/26/2020-16:41:49] [V] [TRT] Fastest Tactic: 6 Time: 0.10944 [11/26/2020-16:41:49] [V] [TRT] --------------- Timing Runner: Conv_7 (CudaDepthwiseConvolution) [11/26/2020-16:41:49] [V] [TRT] CudaDepthwiseConvolution has no valid tactics for this config, skipping [11/26/2020-16:41:49] [V] [TRT] >>>>>>>>>>>>>>> Chose Runner Type: CaskConvolution Tactic: 5921334924264294896 [11/26/2020-16:41:49] [V] [TRT] Conv_7 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148m_nt_v1 [11/26/2020-16:41:49] [V] [TRT] [11/26/2020-16:41:49] [V] [TRT] *************** Autotuning format combination: Half(1,(+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) 1),(MUL_ADD 32 (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) 32),(MUL_ADD 4096 (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) 4096)) -> Half(1,(+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) 1),(MUL_ADD 32 (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) 32),(MUL_ADD 4096 (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) 4096)) *************** [11/26/2020-16:41:49] [V] [TRT] --------------- Timing Runner: Conv_7 (FusedConvActConvolution) [11/26/2020-16:41:49] [V] [TRT] FusedConvActConvolution has no valid tactics for this config, skipping [11/26/2020-16:41:49] [V] [TRT] --------------- Timing Runner: Conv_7 (CaskConvolution) [11/26/2020-16:41:49] [V] [TRT] CaskConvolution has no valid tactics for this config, skipping [11/26/2020-16:41:49] [V] [TRT] --------------- Timing Runner: Conv_7 (CudaConvolution) [11/26/2020-16:41:49] [V] [TRT] Tactic: 0 time 0.249856 [11/26/2020-16:41:49] [V] [TRT] Tactic: 1 time 0.222208 [11/26/2020-16:41:49] [V] [TRT] Tactic: 2 time 0.304128 [11/26/2020-16:41:49] [V] [TRT] Tactic: 5 time 3.09043 [11/26/2020-16:41:49] [V] [TRT] Tactic: 6 time 0.10752 [11/26/2020-16:41:49] [V] [TRT] Fastest Tactic: 6 Time: 0.10752 [11/26/2020-16:41:49] [V] [TRT] --------------- Timing Runner: Conv_7 (CudaDepthwiseConvolution) [11/26/2020-16:41:49] [V] [TRT] CudaDepthwiseConvolution has no valid tactics for this config, skipping [11/26/2020-16:41:49] [V] [TRT] >>>>>>>>>>>>>>> Chose Runner Type: CudaConvolution Tactic: 6 [11/26/2020-16:41:49] [V] [TRT] [11/26/2020-16:41:49] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:41:49] [V] [TRT] Tactic: 0 time 0.015136 [11/26/2020-16:41:49] [V] [TRT] Fastest Tactic: 0 Time: 0.015136 [11/26/2020-16:41:49] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:41:49] [V] [TRT] Tactic: 0 time 0.011552 [11/26/2020-16:41:49] [V] [TRT] Fastest Tactic: 0 Time: 0.011552 [11/26/2020-16:41:49] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:41:49] [V] [TRT] Tactic: 0 time 0.027648 [11/26/2020-16:41:49] [V] [TRT] Fastest Tactic: 0 Time: 0.027648 [11/26/2020-16:41:49] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:41:49] [V] [TRT] Tactic: 0 time 0.018368 [11/26/2020-16:41:49] [V] [TRT] Fastest Tactic: 0 Time: 0.018368 [11/26/2020-16:41:49] [V] [TRT] *************** Autotuning format combination: Float(1,(+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) 1),(MUL_ADD 32 (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) 32),(MUL_ADD 2048 (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) 2048)), Float(1,(+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) 1),(MUL_ADD 32 (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) 32),(MUL_ADD 4096 (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) 4096)) -> Float(1,(+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) 1),(MUL_ADD 32 (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) 32),(MUL_ADD 4096 (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) 4096)) *************** [11/26/2020-16:41:49] [V] [TRT] --------------- Timing Runner: Conv_8 + Add_9 + Relu_10 (LegacySASSConvolution) [11/26/2020-16:41:49] [V] [TRT] Tactic: 0 time 0.026624 [11/26/2020-16:41:49] [V] [TRT] Fastest Tactic: 0 Time: 0.026624 [11/26/2020-16:41:49] [V] [TRT] --------------- Timing Runner: Conv_8 + Add_9 + Relu_10 (CaskConvolution) [11/26/2020-16:41:49] [V] [TRT] Conv_8 + Add_9 + Relu_10 (scudnn) Set Tactic Name: maxwell_scudnn_128x32_relu_medium_nn_v1 [11/26/2020-16:41:49] [V] [TRT] Tactic: 1062367460111450758 time 0.028672 [11/26/2020-16:41:49] [V] [TRT] Conv_8 + Add_9 + Relu_10 (scudnn) Set Tactic Name: maxwell_scudnn_128x128_relu_medium_nn_v1 [11/26/2020-16:41:49] [V] [TRT] Tactic: 4501471010995462441 time 0.0256 [11/26/2020-16:41:49] [V] [TRT] Conv_8 + Add_9 + Relu_10 (scudnn) Set Tactic Name: maxwell_scudnn_128x64_relu_small_nn_v1 [11/26/2020-16:41:49] [V] [TRT] Tactic: 5137655947464784826 time 0.025984 [11/26/2020-16:41:49] [V] [TRT] Conv_8 + Add_9 + Relu_10 (scudnn) Set Tactic Name: maxwell_scudnn_128x128_relu_interior_nn_v1 [11/26/2020-16:41:49] [V] [TRT] Tactic: 5326823351883942011 time 0.0256 [11/26/2020-16:41:49] [V] [TRT] Conv_8 + Add_9 + Relu_10 (scudnn) Set Tactic Name: maxwell_scudnn_128x64_relu_medium_nn_v1 [11/26/2020-16:41:49] [V] [TRT] Tactic: 6645123197870846056 time 0.026624 [11/26/2020-16:41:49] [V] [TRT] Conv_8 + Add_9 + Relu_10 (scudnn) Set Tactic Name: maxwell_scudnn_128x32_relu_interior_nn_v1 [11/26/2020-16:41:49] [V] [TRT] Tactic: -6576203419454146580 time 0.027424 [11/26/2020-16:41:49] [V] [TRT] Conv_8 + Add_9 + Relu_10 (scudnn) Set Tactic Name: maxwell_scudnn_128x32_relu_small_nn_v1 [11/26/2020-16:41:49] [V] [TRT] Tactic: -3456450830548107839 time 0.028608 [11/26/2020-16:41:49] [V] [TRT] Conv_8 + Add_9 + Relu_10 (scudnn) Set Tactic Name: maxwell_scudnn_128x128_relu_small_nn_v1 [11/26/2020-16:41:49] [V] [TRT] Tactic: -410470605513481746 time 0.026624 [11/26/2020-16:41:49] [V] [TRT] Conv_8 + Add_9 + Relu_10 (scudnn) Set Tactic Name: maxwell_scudnn_128x64_relu_interior_nn_v1 [11/26/2020-16:41:49] [V] [TRT] Tactic: -37215280111360163 time 0.026528 [11/26/2020-16:41:49] [V] [TRT] Fastest Tactic: 4501471010995462441 Time: 0.0256 [11/26/2020-16:41:49] [V] [TRT] --------------- Timing Runner: Conv_8 + Add_9 + Relu_10 (CudaConvolution) [11/26/2020-16:41:49] [V] [TRT] Tactic: 0 time 0.091136 [11/26/2020-16:41:49] [V] [TRT] Tactic: 1 time 0.084992 [11/26/2020-16:41:49] [V] [TRT] Tactic: 2 time 0.119808 [11/26/2020-16:41:49] [V] [TRT] Tactic: 5 time 0.217088 [11/26/2020-16:41:49] [V] [TRT] Fastest Tactic: 1 Time: 0.084992 [11/26/2020-16:41:49] [V] [TRT] >>>>>>>>>>>>>>> Chose Runner Type: CaskConvolution Tactic: 4501471010995462441 [11/26/2020-16:41:49] [V] [TRT] Conv_8 + Add_9 + Relu_10 (scudnn) Set Tactic Name: maxwell_scudnn_128x128_relu_medium_nn_v1 [11/26/2020-16:41:49] [V] [TRT] [11/26/2020-16:41:49] [V] [TRT] *************** Autotuning format combination: Half(1,(+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) 1),(MUL_ADD 32 (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) 32),(MUL_ADD 2048 (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) 2048)), Half(1,(+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) 1),(MUL_ADD 32 (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) 32),(MUL_ADD 4096 (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) 4096)) -> Half(1,(+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) 1),(MUL_ADD 32 (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) 32),(MUL_ADD 4096 (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) 4096)) *************** [11/26/2020-16:41:49] [V] [TRT] --------------- Timing Runner: Conv_8 + Add_9 + Relu_10 (CaskConvolution) [11/26/2020-16:41:49] [V] [TRT] CaskConvolution has no valid tactics for this config, skipping [11/26/2020-16:41:49] [V] [TRT] --------------- Timing Runner: Conv_8 + Add_9 + Relu_10 (CudaConvolution) [11/26/2020-16:41:49] [V] [TRT] Tactic: 0 time 0.057344 [11/26/2020-16:41:49] [V] [TRT] Tactic: 1 time 0.057344 [11/26/2020-16:41:49] [V] [TRT] Tactic: 2 time 0.084992 [11/26/2020-16:41:49] [V] [TRT] Tactic: 5 time 0.18944 [11/26/2020-16:41:49] [V] [TRT] Fastest Tactic: 0 Time: 0.057344 [11/26/2020-16:41:49] [V] [TRT] >>>>>>>>>>>>>>> Chose Runner Type: CudaConvolution Tactic: 0 [11/26/2020-16:41:49] [V] [TRT] [11/26/2020-16:41:49] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:41:50] [V] [TRT] Tactic: 0 time 0.026496 [11/26/2020-16:41:50] [V] [TRT] Fastest Tactic: 0 Time: 0.026496 [11/26/2020-16:41:50] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:41:50] [V] [TRT] Tactic: 0 time 0.01824 [11/26/2020-16:41:50] [V] [TRT] Fastest Tactic: 0 Time: 0.01824 [11/26/2020-16:41:50] [V] [TRT] *************** Autotuning format combination: Float(1,(+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) 1),(MUL_ADD 32 (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) 32),(MUL_ADD 4096 (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) 4096)) -> Float(1,(+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) 1),(MUL_ADD 32 (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) 32),(MUL_ADD 4096 (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) 4096)) *************** [11/26/2020-16:41:50] [V] [TRT] --------------- Timing Runner: Conv_11 + Relu_12 (LegacySASSConvolution) [11/26/2020-16:41:50] [V] [TRT] Tactic: 0 time 0.198656 [11/26/2020-16:41:50] [V] [TRT] Tactic: 1 time 0.100352 [11/26/2020-16:41:50] [V] [TRT] Fastest Tactic: 1 Time: 0.100352 [11/26/2020-16:41:50] [V] [TRT] --------------- Timing Runner: Conv_11 + Relu_12 (FusedConvActConvolution) [11/26/2020-16:41:50] [V] [TRT] Tactic: 7 time 0.204064 [11/26/2020-16:41:50] [V] [TRT] Tactic: 10 time 0.185344 [11/26/2020-16:41:50] [V] [TRT] Tactic: 14 time 0.169984 [11/26/2020-16:41:50] [V] [TRT] Tactic: 15 time 0.1872 [11/26/2020-16:41:50] [V] [TRT] Tactic: 25 time 0.17408 [11/26/2020-16:41:50] [V] [TRT] Tactic: 26 time 0.251904 [11/26/2020-16:41:50] [V] [TRT] Tactic: 29 time 0.183296 [11/26/2020-16:41:50] [V] [TRT] Tactic: 30 time 0.191488 [11/26/2020-16:41:50] [V] [TRT] Tactic: 33 time 0.23216 [11/26/2020-16:41:50] [V] [TRT] Tactic: 36 time 0.314368 [11/26/2020-16:41:50] [V] [TRT] Tactic: 39 time 0.197632 [11/26/2020-16:41:50] [V] [TRT] Tactic: 41 time 0.198496 [11/26/2020-16:41:50] [V] [TRT] Tactic: 42 time 0.29696 [11/26/2020-16:41:50] [V] [TRT] Tactic: 43 time 0.222208 [11/26/2020-16:41:50] [V] [TRT] Tactic: 45 time 0.17408 [11/26/2020-16:41:50] [V] [TRT] Tactic: 47 time 0.18944 [11/26/2020-16:41:50] [V] [TRT] Tactic: 52 time 0.197536 [11/26/2020-16:41:50] [V] [TRT] Tactic: 54 time 0.183296 [11/26/2020-16:41:50] [V] [TRT] Tactic: 56 time 0.25088 [11/26/2020-16:41:50] [V] [TRT] Tactic: 66 time 0.208608 [11/26/2020-16:41:50] [V] [TRT] Tactic: 76 time 0.18944 [11/26/2020-16:41:50] [V] [TRT] Tactic: 90 time 0.197632 [11/26/2020-16:41:50] [V] [TRT] Tactic: 93 time 0.196608 [11/26/2020-16:41:50] [V] [TRT] Tactic: 98 time 0.208896 [11/26/2020-16:41:50] [V] [TRT] Tactic: 104 time 0.192512 [11/26/2020-16:41:50] [V] [TRT] Tactic: 110 time 0.208896 [11/26/2020-16:41:50] [V] [TRT] Tactic: 119 time 0.211264 [11/26/2020-16:41:50] [V] [TRT] Tactic: 121 time 0.176128 [11/26/2020-16:41:50] [V] [TRT] Tactic: 130 time 0.194432 [11/26/2020-16:41:50] [V] [TRT] Tactic: 134 time 0.205056 [11/26/2020-16:41:50] [V] [TRT] Tactic: 136 time 0.200704 [11/26/2020-16:41:50] [V] [TRT] Tactic: 137 time 0.18944 [11/26/2020-16:41:50] [V] [TRT] Tactic: 139 time 0.181248 [11/26/2020-16:41:50] [V] [TRT] Tactic: 144 time 0.206848 [11/26/2020-16:41:50] [V] [TRT] Tactic: 149 time 0.22016 [11/26/2020-16:41:50] [V] [TRT] Tactic: 151 time 0.201728 [11/26/2020-16:41:50] [V] [TRT] Tactic: 152 time 0.187392 [11/26/2020-16:41:50] [V] [TRT] Tactic: 153 time 0.205888 [11/26/2020-16:41:50] [V] [TRT] Tactic: 156 time 0.1792 [11/26/2020-16:41:50] [V] [TRT] Tactic: 159 time 0.182272 [11/26/2020-16:41:50] [V] [TRT] Tactic: 162 time 0.201728 [11/26/2020-16:41:50] [V] [TRT] Tactic: 164 time 0.175104 [11/26/2020-16:41:50] [V] [TRT] Fastest Tactic: 14 Time: 0.169984 [11/26/2020-16:41:50] [V] [TRT] --------------- Timing Runner: Conv_11 + Relu_12 (CaskConvolution) [11/26/2020-16:41:50] [V] [TRT] Conv_11 + Relu_12 (scudnn) Set Tactic Name: maxwell_scudnn_128x32_relu_medium_nn_v1 [11/26/2020-16:41:50] [V] [TRT] Tactic: 1062367460111450758 time 0.188416 [11/26/2020-16:41:50] [V] [TRT] Conv_11 + Relu_12 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148n_nt_v1 [11/26/2020-16:41:50] [V] [TRT] Tactic: 3827454225649558724 time 0.083968 [11/26/2020-16:41:50] [V] [TRT] Conv_11 + Relu_12 (scudnn) Set Tactic Name: maxwell_scudnn_128x64_relu_large_nn_v1 [11/26/2020-16:41:50] [V] [TRT] Tactic: 4337000649858996379 time 0.165888 [11/26/2020-16:41:50] [V] [TRT] Conv_11 + Relu_12 (scudnn) Set Tactic Name: maxwell_scudnn_128x128_relu_medium_nn_v1 [11/26/2020-16:41:50] [V] [TRT] Tactic: 4501471010995462441 time 0.195584 [11/26/2020-16:41:50] [V] [TRT] Conv_11 + Relu_12 (scudnn) Set Tactic Name: maxwell_scudnn_128x64_relu_small_nn_v1 [11/26/2020-16:41:50] [V] [TRT] Tactic: 5137655947464784826 time 0.15872 [11/26/2020-16:41:50] [V] [TRT] Conv_11 + Relu_12 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148m_nt_v1 [11/26/2020-16:41:50] [V] [TRT] Tactic: 5921334924264294896 time 0.088064 [11/26/2020-16:41:50] [V] [TRT] Conv_11 + Relu_12 (scudnn) Set Tactic Name: maxwell_scudnn_128x64_relu_medium_nn_v1 [11/26/2020-16:41:50] [V] [TRT] Tactic: 6645123197870846056 time 0.16384 [11/26/2020-16:41:50] [V] [TRT] Conv_11 + Relu_12 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148t_nt_v1 [11/26/2020-16:41:50] [V] [TRT] Tactic: 7852627285308570038 time 0.092704 [11/26/2020-16:41:50] [V] [TRT] Conv_11 + Relu_12 (scudnn) Set Tactic Name: maxwell_scudnn_128x128_relu_large_nn_v1 [11/26/2020-16:41:50] [V] [TRT] Tactic: -9137461792520977713 time 0.197632 [11/26/2020-16:41:50] [V] [TRT] Conv_11 + Relu_12 (scudnn) Set Tactic Name: maxwell_scudnn_128x32_relu_large_nn_v1 [11/26/2020-16:41:50] [V] [TRT] Tactic: -6092040395344634144 time 0.192512 [11/26/2020-16:41:50] [V] [TRT] Conv_11 + Relu_12 (scudnn) Set Tactic Name: maxwell_scudnn_128x32_relu_small_nn_v1 [11/26/2020-16:41:50] [V] [TRT] Tactic: -3456450830548107839 time 0.177152 [11/26/2020-16:41:50] [V] [TRT] Conv_11 + Relu_12 (scudnn) Set Tactic Name: maxwell_scudnn_128x128_relu_small_nn_v1 [11/26/2020-16:41:50] [V] [TRT] Tactic: -410470605513481746 time 0.190464 [11/26/2020-16:41:50] [V] [TRT] Fastest Tactic: 3827454225649558724 Time: 0.083968 [11/26/2020-16:41:50] [V] [TRT] --------------- Timing Runner: Conv_11 + Relu_12 (CudaConvolution) [11/26/2020-16:41:50] [V] [TRT] Tactic: 0 time 0.268288 [11/26/2020-16:41:50] [V] [TRT] Tactic: 1 time 0.228352 [11/26/2020-16:41:50] [V] [TRT] Tactic: 2 time 0.326656 [11/26/2020-16:41:50] [V] [TRT] Tactic: 5 time 2.86822 [11/26/2020-16:41:50] [V] [TRT] Tactic: 6 time 0.115712 [11/26/2020-16:41:50] [V] [TRT] Fastest Tactic: 6 Time: 0.115712 [11/26/2020-16:41:50] [V] [TRT] --------------- Timing Runner: Conv_11 + Relu_12 (CudaDepthwiseConvolution) [11/26/2020-16:41:50] [V] [TRT] CudaDepthwiseConvolution has no valid tactics for this config, skipping [11/26/2020-16:41:50] [V] [TRT] >>>>>>>>>>>>>>> Chose Runner Type: CaskConvolution Tactic: 3827454225649558724 [11/26/2020-16:41:50] [V] [TRT] Conv_11 + Relu_12 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148n_nt_v1 [11/26/2020-16:41:50] [V] [TRT] [11/26/2020-16:41:50] [V] [TRT] *************** Autotuning format combination: Half(1,(+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) 1),(MUL_ADD 32 (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) 32),(MUL_ADD 4096 (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) 4096)) -> Half(1,(+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) 1),(MUL_ADD 32 (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) 32),(MUL_ADD 4096 (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) 4096)) *************** [11/26/2020-16:41:50] [V] [TRT] --------------- Timing Runner: Conv_11 + Relu_12 (FusedConvActConvolution) [11/26/2020-16:41:50] [V] [TRT] FusedConvActConvolution has no valid tactics for this config, skipping [11/26/2020-16:41:50] [V] [TRT] --------------- Timing Runner: Conv_11 + Relu_12 (CaskConvolution) [11/26/2020-16:41:50] [V] [TRT] CaskConvolution has no valid tactics for this config, skipping [11/26/2020-16:41:50] [V] [TRT] --------------- Timing Runner: Conv_11 + Relu_12 (CudaConvolution) [11/26/2020-16:41:50] [V] [TRT] Tactic: 0 time 0.257024 [11/26/2020-16:41:50] [V] [TRT] Tactic: 1 time 0.226304 [11/26/2020-16:41:50] [V] [TRT] Tactic: 2 time 0.308224 [11/26/2020-16:41:50] [V] [TRT] Tactic: 5 time 3.11398 [11/26/2020-16:41:50] [V] [TRT] Tactic: 6 time 0.113664 [11/26/2020-16:41:50] [V] [TRT] Fastest Tactic: 6 Time: 0.113664 [11/26/2020-16:41:50] [V] [TRT] --------------- Timing Runner: Conv_11 + Relu_12 (CudaDepthwiseConvolution) [11/26/2020-16:41:50] [V] [TRT] CudaDepthwiseConvolution has no valid tactics for this config, skipping [11/26/2020-16:41:50] [V] [TRT] >>>>>>>>>>>>>>> Chose Runner Type: CudaConvolution Tactic: 6 [11/26/2020-16:41:50] [V] [TRT] [11/26/2020-16:41:50] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:41:50] [V] [TRT] Tactic: 0 time 0.027456 [11/26/2020-16:41:50] [V] [TRT] Fastest Tactic: 0 Time: 0.027456 [11/26/2020-16:41:50] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:41:50] [V] [TRT] Tactic: 0 time 0.018368 [11/26/2020-16:41:50] [V] [TRT] Fastest Tactic: 0 Time: 0.018368 [11/26/2020-16:41:50] [V] [TRT] *************** Autotuning format combination: Float(1,(+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) 1),(MUL_ADD 32 (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) 32),(MUL_ADD 4096 (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) 4096)) -> Float(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 1),(MUL_ADD 16 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 2048 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2048)) *************** [11/26/2020-16:41:50] [V] [TRT] --------------- Timing Runner: MaxPool_13 (Pooling) [11/26/2020-16:41:50] [V] [TRT] Tactic: -1 time 0.013312 [11/26/2020-16:41:50] [V] [TRT] Fastest Tactic: -1 Time: 0.013312 [11/26/2020-16:41:50] [V] [TRT] --------------- Timing Runner: MaxPool_13 (TiledPooling) [11/26/2020-16:41:50] [V] [TRT] Tactic: 5505281 time 0.01616 [11/26/2020-16:41:50] [V] [TRT] Tactic: 5570817 time 0.014272 [11/26/2020-16:41:50] [V] [TRT] Tactic: 5636353 time 0.013312 [11/26/2020-16:41:50] [V] [TRT] Tactic: 5701889 time 0.012096 [11/26/2020-16:41:50] [V] [TRT] Tactic: 5767425 time 0.012224 [11/26/2020-16:41:50] [V] [TRT] Tactic: 5832961 time 0.013312 [11/26/2020-16:41:50] [V] [TRT] Tactic: 5898497 time 0.012096 [11/26/2020-16:41:50] [V] [TRT] Tactic: 5964033 time 0.012128 [11/26/2020-16:41:50] [V] [TRT] Tactic: 6029569 time 0.014336 [11/26/2020-16:41:50] [V] [TRT] Tactic: 6095105 time 0.012128 [11/26/2020-16:41:50] [V] [TRT] Tactic: 6160641 time 0.012288 [11/26/2020-16:41:50] [V] [TRT] Tactic: 6226177 time 0.012288 [11/26/2020-16:41:50] [V] [TRT] Tactic: 6291713 time 0.012 [11/26/2020-16:41:50] [V] [TRT] Tactic: 6357249 time 0.01392 [11/26/2020-16:41:50] [V] [TRT] Tactic: 6422785 time 0.012064 [11/26/2020-16:41:50] [V] [TRT] Tactic: 6488321 time 0.015552 [11/26/2020-16:41:50] [V] [TRT] Fastest Tactic: 6291713 Time: 0.012 [11/26/2020-16:41:50] [V] [TRT] >>>>>>>>>>>>>>> Chose Runner Type: TiledPooling Tactic: 6291713 [11/26/2020-16:41:50] [V] [TRT] [11/26/2020-16:41:50] [V] [TRT] *************** Autotuning format combination: Half(1,(+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) 1),(MUL_ADD 32 (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) 32),(MUL_ADD 4096 (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) 4096)) -> Half(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 1),(MUL_ADD 16 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 2048 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2048)) *************** [11/26/2020-16:41:50] [V] [TRT] --------------- Timing Runner: MaxPool_13 (Pooling) [11/26/2020-16:41:50] [V] [TRT] Tactic: -1 time 0.01024 [11/26/2020-16:41:50] [V] [TRT] Fastest Tactic: -1 Time: 0.01024 [11/26/2020-16:41:50] [V] [TRT] --------------- Timing Runner: MaxPool_13 (TiledPooling) [11/26/2020-16:41:50] [V] [TRT] TiledPooling has no valid tactics for this config, skipping [11/26/2020-16:41:50] [V] [TRT] >>>>>>>>>>>>>>> Chose Runner Type: Pooling Tactic: -1 [11/26/2020-16:41:50] [V] [TRT] [11/26/2020-16:41:50] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:41:50] [V] [TRT] Tactic: 0 time 0.009216 [11/26/2020-16:41:50] [V] [TRT] Fastest Tactic: 0 Time: 0.009216 [11/26/2020-16:41:50] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:41:50] [V] [TRT] Tactic: 0 time 0.00784 [11/26/2020-16:41:50] [V] [TRT] Fastest Tactic: 0 Time: 0.00784 [11/26/2020-16:41:50] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:41:50] [V] [TRT] Tactic: 0 time 0.009216 [11/26/2020-16:41:50] [V] [TRT] Fastest Tactic: 0 Time: 0.009216 [11/26/2020-16:41:50] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:41:50] [V] [TRT] Tactic: 0 time 0.007104 [11/26/2020-16:41:50] [V] [TRT] Fastest Tactic: 0 Time: 0.007104 [11/26/2020-16:41:50] [V] [TRT] *************** Autotuning format combination: Float(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 1),(MUL_ADD 16 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 2048 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2048)) -> Float(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 1),(MUL_ADD 16 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 4096)) *************** [11/26/2020-16:41:50] [V] [TRT] --------------- Timing Runner: Conv_14 + Relu_15 (LegacySASSConvolution) [11/26/2020-16:41:50] [V] [TRT] Tactic: 0 time 0.144 [11/26/2020-16:41:50] [V] [TRT] Tactic: 1 time 0.059392 [11/26/2020-16:41:50] [V] [TRT] Fastest Tactic: 1 Time: 0.059392 [11/26/2020-16:41:50] [V] [TRT] --------------- Timing Runner: Conv_14 + Relu_15 (FusedConvActConvolution) [11/26/2020-16:41:50] [V] [TRT] Tactic: 7 time 0.146432 [11/26/2020-16:41:50] [V] [TRT] Tactic: 10 time 0.1024 [11/26/2020-16:41:50] [V] [TRT] Tactic: 14 time 0.09216 [11/26/2020-16:41:50] [V] [TRT] Tactic: 15 time 0.10528 [11/26/2020-16:41:51] [V] [TRT] Tactic: 25 time 0.095232 [11/26/2020-16:41:51] [V] [TRT] Tactic: 26 time 0.144288 [11/26/2020-16:41:51] [V] [TRT] Tactic: 29 time 0.136128 [11/26/2020-16:41:51] [V] [TRT] Tactic: 30 time 0.150528 [11/26/2020-16:41:51] [V] [TRT] Tactic: 33 time 0.150368 [11/26/2020-16:41:51] [V] [TRT] Tactic: 36 time 0.15872 [11/26/2020-16:41:51] [V] [TRT] Tactic: 39 time 0.104384 [11/26/2020-16:41:51] [V] [TRT] Tactic: 41 time 0.139264 [11/26/2020-16:41:51] [V] [TRT] Tactic: 42 time 0.188416 [11/26/2020-16:41:51] [V] [TRT] Tactic: 43 time 0.159744 [11/26/2020-16:41:51] [V] [TRT] Tactic: 45 time 0.09216 [11/26/2020-16:41:51] [V] [TRT] Tactic: 47 time 0.109568 [11/26/2020-16:41:51] [V] [TRT] Tactic: 52 time 0.113952 [11/26/2020-16:41:51] [V] [TRT] Tactic: 54 time 0.112192 [11/26/2020-16:41:51] [V] [TRT] Tactic: 56 time 0.126976 [11/26/2020-16:41:51] [V] [TRT] Tactic: 66 time 0.146432 [11/26/2020-16:41:51] [V] [TRT] Tactic: 76 time 0.101376 [11/26/2020-16:41:51] [V] [TRT] Tactic: 90 time 0.12288 [11/26/2020-16:41:51] [V] [TRT] Tactic: 93 time 0.118784 [11/26/2020-16:41:51] [V] [TRT] Tactic: 98 time 0.1408 [11/26/2020-16:41:51] [V] [TRT] Tactic: 104 time 0.124928 [11/26/2020-16:41:51] [V] [TRT] Tactic: 110 time 0.115712 [11/26/2020-16:41:51] [V] [TRT] Tactic: 119 time 0.124928 [11/26/2020-16:41:51] [V] [TRT] Tactic: 121 time 0.098304 [11/26/2020-16:41:51] [V] [TRT] Tactic: 130 time 0.13824 [11/26/2020-16:41:51] [V] [TRT] Tactic: 134 time 0.130048 [11/26/2020-16:41:51] [V] [TRT] Tactic: 136 time 0.155392 [11/26/2020-16:41:51] [V] [TRT] Tactic: 137 time 0.121184 [11/26/2020-16:41:51] [V] [TRT] Tactic: 139 time 0.110592 [11/26/2020-16:41:51] [V] [TRT] Tactic: 144 time 0.14848 [11/26/2020-16:41:51] [V] [TRT] Tactic: 149 time 0.10912 [11/26/2020-16:41:51] [V] [TRT] Tactic: 151 time 0.120864 [11/26/2020-16:41:51] [V] [TRT] Tactic: 152 time 0.11776 [11/26/2020-16:41:51] [V] [TRT] Tactic: 153 time 0.160768 [11/26/2020-16:41:51] [V] [TRT] Tactic: 156 time 0.125952 [11/26/2020-16:41:51] [V] [TRT] Tactic: 159 time 0.111616 [11/26/2020-16:41:51] [V] [TRT] Tactic: 162 time 0.1176 [11/26/2020-16:41:51] [V] [TRT] Tactic: 164 time 0.090496 [11/26/2020-16:41:51] [V] [TRT] Fastest Tactic: 164 Time: 0.090496 [11/26/2020-16:41:51] [V] [TRT] --------------- Timing Runner: Conv_14 + Relu_15 (CaskConvolution) [11/26/2020-16:41:51] [V] [TRT] Conv_14 + Relu_15 (scudnn) Set Tactic Name: maxwell_scudnn_128x32_relu_medium_nn_v1 [11/26/2020-16:41:51] [V] [TRT] Tactic: 1062367460111450758 time 0.144384 [11/26/2020-16:41:51] [V] [TRT] Conv_14 + Relu_15 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148n_nt_v1 [11/26/2020-16:41:51] [V] [TRT] Tactic: 3827454225649558724 time 0.048128 [11/26/2020-16:41:51] [V] [TRT] Conv_14 + Relu_15 (scudnn) Set Tactic Name: maxwell_scudnn_128x64_relu_large_nn_v1 [11/26/2020-16:41:51] [V] [TRT] Tactic: 4337000649858996379 time 0.126976 [11/26/2020-16:41:51] [V] [TRT] Conv_14 + Relu_15 (scudnn) Set Tactic Name: maxwell_scudnn_128x128_relu_medium_nn_v1 [11/26/2020-16:41:51] [V] [TRT] Tactic: 4501471010995462441 time 0.116736 [11/26/2020-16:41:51] [V] [TRT] Conv_14 + Relu_15 (scudnn) Set Tactic Name: maxwell_scudnn_128x64_relu_small_nn_v1 [11/26/2020-16:41:51] [V] [TRT] Tactic: 5137655947464784826 time 0.114688 [11/26/2020-16:41:51] [V] [TRT] Conv_14 + Relu_15 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148m_nt_v1 [11/26/2020-16:41:51] [V] [TRT] Tactic: 5921334924264294896 time 0.052224 [11/26/2020-16:41:51] [V] [TRT] Conv_14 + Relu_15 (scudnn) Set Tactic Name: maxwell_scudnn_128x64_relu_medium_nn_v1 [11/26/2020-16:41:51] [V] [TRT] Tactic: 6645123197870846056 time 0.121856 [11/26/2020-16:41:51] [V] [TRT] Conv_14 + Relu_15 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148t_nt_v1 [11/26/2020-16:41:51] [V] [TRT] Tactic: 7852627285308570038 time 0.0512 [11/26/2020-16:41:51] [V] [TRT] Conv_14 + Relu_15 (scudnn) Set Tactic Name: maxwell_scudnn_128x128_relu_large_nn_v1 [11/26/2020-16:41:51] [V] [TRT] Tactic: -9137461792520977713 time 0.119808 [11/26/2020-16:41:51] [V] [TRT] Conv_14 + Relu_15 (scudnn) Set Tactic Name: maxwell_scudnn_128x32_relu_large_nn_v1 [11/26/2020-16:41:51] [V] [TRT] Tactic: -6092040395344634144 time 0.1536 [11/26/2020-16:41:51] [V] [TRT] Conv_14 + Relu_15 (scudnn) Set Tactic Name: maxwell_scudnn_128x32_relu_small_nn_v1 [11/26/2020-16:41:51] [V] [TRT] Tactic: -3456450830548107839 time 0.125952 [11/26/2020-16:41:51] [V] [TRT] Conv_14 + Relu_15 (scudnn) Set Tactic Name: maxwell_scudnn_128x128_relu_small_nn_v1 [11/26/2020-16:41:51] [V] [TRT] Tactic: -410470605513481746 time 0.11264 [11/26/2020-16:41:51] [V] [TRT] Fastest Tactic: 3827454225649558724 Time: 0.048128 [11/26/2020-16:41:51] [V] [TRT] --------------- Timing Runner: Conv_14 + Relu_15 (CudaConvolution) [11/26/2020-16:41:51] [V] [TRT] Tactic: 0 time 0.198656 [11/26/2020-16:41:51] [V] [TRT] Tactic: 1 time 0.136192 [11/26/2020-16:41:51] [V] [TRT] Tactic: 2 time 0.210944 [11/26/2020-16:41:51] [V] [TRT] Tactic: 5 time 1.79098 [11/26/2020-16:41:51] [V] [TRT] Tactic: 6 time 0.073728 [11/26/2020-16:41:51] [V] [TRT] Fastest Tactic: 6 Time: 0.073728 [11/26/2020-16:41:51] [V] [TRT] --------------- Timing Runner: Conv_14 + Relu_15 (CudaDepthwiseConvolution) [11/26/2020-16:41:51] [V] [TRT] CudaDepthwiseConvolution has no valid tactics for this config, skipping [11/26/2020-16:41:51] [V] [TRT] >>>>>>>>>>>>>>> Chose Runner Type: CaskConvolution Tactic: 3827454225649558724 [11/26/2020-16:41:51] [V] [TRT] Conv_14 + Relu_15 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148n_nt_v1 [11/26/2020-16:41:51] [V] [TRT] [11/26/2020-16:41:51] [V] [TRT] *************** Autotuning format combination: Half(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 1),(MUL_ADD 16 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 2048 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2048)) -> Half(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 1),(MUL_ADD 16 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 4096)) *************** [11/26/2020-16:41:51] [V] [TRT] --------------- Timing Runner: Conv_14 + Relu_15 (FusedConvActConvolution) [11/26/2020-16:41:51] [V] [TRT] FusedConvActConvolution has no valid tactics for this config, skipping [11/26/2020-16:41:51] [V] [TRT] --------------- Timing Runner: Conv_14 + Relu_15 (CaskConvolution) [11/26/2020-16:41:51] [V] [TRT] CaskConvolution has no valid tactics for this config, skipping [11/26/2020-16:41:51] [V] [TRT] --------------- Timing Runner: Conv_14 + Relu_15 (CudaConvolution) [11/26/2020-16:41:51] [V] [TRT] Tactic: 0 time 0.228352 [11/26/2020-16:41:51] [V] [TRT] Tactic: 1 time 0.149504 [11/26/2020-16:41:51] [V] [TRT] Tactic: 2 time 0.243712 [11/26/2020-16:41:51] [V] [TRT] Tactic: 5 time 1.97626 [11/26/2020-16:41:51] [V] [TRT] Tactic: 6 time 0.070656 [11/26/2020-16:41:51] [V] [TRT] Fastest Tactic: 6 Time: 0.070656 [11/26/2020-16:41:51] [V] [TRT] --------------- Timing Runner: Conv_14 + Relu_15 (CudaDepthwiseConvolution) [11/26/2020-16:41:51] [V] [TRT] CudaDepthwiseConvolution has no valid tactics for this config, skipping [11/26/2020-16:41:51] [V] [TRT] >>>>>>>>>>>>>>> Chose Runner Type: CudaConvolution Tactic: 6 [11/26/2020-16:41:51] [V] [TRT] [11/26/2020-16:41:51] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:41:51] [V] [TRT] Tactic: 0 time 0.015264 [11/26/2020-16:41:51] [V] [TRT] Fastest Tactic: 0 Time: 0.015264 [11/26/2020-16:41:51] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:41:51] [V] [TRT] Tactic: 0 time 0.011264 [11/26/2020-16:41:51] [V] [TRT] Fastest Tactic: 0 Time: 0.011264 [11/26/2020-16:41:51] [V] [TRT] *************** Autotuning format combination: Float(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 1),(MUL_ADD 16 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 4096)) -> Float(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 1),(MUL_ADD 16 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 4096)) *************** [11/26/2020-16:41:51] [V] [TRT] --------------- Timing Runner: Conv_16 (LegacySASSConvolution) [11/26/2020-16:41:51] [V] [TRT] Tactic: 0 time 0.280576 [11/26/2020-16:41:51] [V] [TRT] Tactic: 1 time 0.109472 [11/26/2020-16:41:51] [V] [TRT] Fastest Tactic: 1 Time: 0.109472 [11/26/2020-16:41:51] [V] [TRT] --------------- Timing Runner: Conv_16 (FusedConvActConvolution) [11/26/2020-16:41:51] [V] [TRT] Tactic: 7 time 0.260992 [11/26/2020-16:41:51] [V] [TRT] Tactic: 10 time 0.199488 [11/26/2020-16:41:51] [V] [TRT] Tactic: 14 time 0.166848 [11/26/2020-16:41:51] [V] [TRT] Tactic: 15 time 0.191424 [11/26/2020-16:41:51] [V] [TRT] Tactic: 25 time 0.17408 [11/26/2020-16:41:51] [V] [TRT] Tactic: 26 time 0.263168 [11/26/2020-16:41:51] [V] [TRT] Tactic: 29 time 0.24576 [11/26/2020-16:41:51] [V] [TRT] Tactic: 30 time 0.254976 [11/26/2020-16:41:51] [V] [TRT] Tactic: 33 time 0.285696 [11/26/2020-16:41:51] [V] [TRT] Tactic: 36 time 0.292864 [11/26/2020-16:41:51] [V] [TRT] Tactic: 39 time 0.188992 [11/26/2020-16:41:51] [V] [TRT] Tactic: 41 time 0.23552 [11/26/2020-16:41:51] [V] [TRT] Tactic: 42 time 0.3584 [11/26/2020-16:41:51] [V] [TRT] Tactic: 43 time 0.313312 [11/26/2020-16:41:51] [V] [TRT] Tactic: 45 time 0.166432 [11/26/2020-16:41:51] [V] [TRT] Tactic: 47 time 0.2048 [11/26/2020-16:41:51] [V] [TRT] Tactic: 52 time 0.20992 [11/26/2020-16:41:51] [V] [TRT] Tactic: 54 time 0.212992 [11/26/2020-16:41:51] [V] [TRT] Tactic: 56 time 0.224096 [11/26/2020-16:41:51] [V] [TRT] Tactic: 66 time 0.278528 [11/26/2020-16:41:51] [V] [TRT] Tactic: 76 time 0.187392 [11/26/2020-16:41:51] [V] [TRT] Tactic: 90 time 0.222784 [11/26/2020-16:41:51] [V] [TRT] Tactic: 93 time 0.20992 [11/26/2020-16:41:51] [V] [TRT] Tactic: 98 time 0.263712 [11/26/2020-16:41:51] [V] [TRT] Tactic: 104 time 0.221184 [11/26/2020-16:41:51] [V] [TRT] Tactic: 110 time 0.226304 [11/26/2020-16:41:51] [V] [TRT] Tactic: 119 time 0.234496 [11/26/2020-16:41:51] [V] [TRT] Tactic: 121 time 0.178176 [11/26/2020-16:41:51] [V] [TRT] Tactic: 130 time 0.246336 [11/26/2020-16:41:51] [V] [TRT] Tactic: 134 time 0.24064 [11/26/2020-16:41:51] [V] [TRT] Tactic: 136 time 0.278528 [11/26/2020-16:41:51] [V] [TRT] Tactic: 137 time 0.229376 [11/26/2020-16:41:51] [V] [TRT] Tactic: 139 time 0.215616 [11/26/2020-16:41:51] [V] [TRT] Tactic: 144 time 0.278528 [11/26/2020-16:41:51] [V] [TRT] Tactic: 149 time 0.203328 [11/26/2020-16:41:51] [V] [TRT] Tactic: 151 time 0.228896 [11/26/2020-16:41:51] [V] [TRT] Tactic: 152 time 0.228352 [11/26/2020-16:41:51] [V] [TRT] Tactic: 153 time 0.277472 [11/26/2020-16:41:51] [V] [TRT] Tactic: 156 time 0.247808 [11/26/2020-16:41:51] [V] [TRT] Tactic: 159 time 0.208896 [11/26/2020-16:41:51] [V] [TRT] Tactic: 162 time 0.223232 [11/26/2020-16:41:51] [V] [TRT] Tactic: 164 time 0.165888 [11/26/2020-16:41:51] [V] [TRT] Fastest Tactic: 164 Time: 0.165888 [11/26/2020-16:41:51] [V] [TRT] --------------- Timing Runner: Conv_16 (CaskConvolution) [11/26/2020-16:41:51] [V] [TRT] Conv_16 (scudnn) Set Tactic Name: maxwell_scudnn_128x32_relu_medium_nn_v1 [11/26/2020-16:41:51] [V] [TRT] Tactic: 1062367460111450758 time 0.282624 [11/26/2020-16:41:51] [V] [TRT] Conv_16 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148n_nt_v1 [11/26/2020-16:41:51] [V] [TRT] Tactic: 3827454225649558724 time 0.085344 [11/26/2020-16:41:51] [V] [TRT] Conv_16 (scudnn) Set Tactic Name: maxwell_scudnn_128x64_relu_large_nn_v1 [11/26/2020-16:41:51] [V] [TRT] Tactic: 4337000649858996379 time 0.256 [11/26/2020-16:41:51] [V] [TRT] Conv_16 (scudnn) Set Tactic Name: maxwell_scudnn_128x128_relu_medium_nn_v1 [11/26/2020-16:41:51] [V] [TRT] Tactic: 4501471010995462441 time 0.23552 [11/26/2020-16:41:51] [V] [TRT] Conv_16 (scudnn) Set Tactic Name: maxwell_scudnn_128x64_relu_small_nn_v1 [11/26/2020-16:41:51] [V] [TRT] Tactic: 5137655947464784826 time 0.2304 [11/26/2020-16:41:51] [V] [TRT] Conv_16 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148m_nt_v1 [11/26/2020-16:41:51] [V] [TRT] Tactic: 5921334924264294896 time 0.091136 [11/26/2020-16:41:51] [V] [TRT] Conv_16 (scudnn) Set Tactic Name: maxwell_scudnn_128x64_relu_medium_nn_v1 [11/26/2020-16:41:51] [V] [TRT] Tactic: 6645123197870846056 time 0.243712 [11/26/2020-16:41:51] [V] [TRT] Conv_16 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148t_nt_v1 [11/26/2020-16:41:51] [V] [TRT] Tactic: 7852627285308570038 time 0.092736 [11/26/2020-16:41:51] [V] [TRT] Conv_16 (scudnn) Set Tactic Name: maxwell_scudnn_128x128_relu_large_nn_v1 [11/26/2020-16:41:51] [V] [TRT] Tactic: -9137461792520977713 time 0.243712 [11/26/2020-16:41:51] [V] [TRT] Conv_16 (scudnn) Set Tactic Name: maxwell_scudnn_128x32_relu_large_nn_v1 [11/26/2020-16:41:51] [V] [TRT] Tactic: -6092040395344634144 time 0.297984 [11/26/2020-16:41:51] [V] [TRT] Conv_16 (scudnn) Set Tactic Name: maxwell_scudnn_128x32_relu_small_nn_v1 [11/26/2020-16:41:51] [V] [TRT] Tactic: -3456450830548107839 time 0.24576 [11/26/2020-16:41:51] [V] [TRT] Conv_16 (scudnn) Set Tactic Name: maxwell_scudnn_128x128_relu_small_nn_v1 [11/26/2020-16:41:51] [V] [TRT] Tactic: -410470605513481746 time 0.228288 [11/26/2020-16:41:51] [V] [TRT] Fastest Tactic: 3827454225649558724 Time: 0.085344 [11/26/2020-16:41:51] [V] [TRT] --------------- Timing Runner: Conv_16 (CudaConvolution) [11/26/2020-16:41:51] [V] [TRT] Tactic: 0 time 0.37376 [11/26/2020-16:41:51] [V] [TRT] Tactic: 1 time 0.272384 [11/26/2020-16:41:51] [V] [TRT] Tactic: 2 time 0.344064 [11/26/2020-16:41:51] [V] [TRT] Tactic: 5 time 3.48051 [11/26/2020-16:41:51] [V] [TRT] Tactic: 6 time 0.121856 [11/26/2020-16:41:51] [V] [TRT] Fastest Tactic: 6 Time: 0.121856 [11/26/2020-16:41:51] [V] [TRT] --------------- Timing Runner: Conv_16 (CudaDepthwiseConvolution) [11/26/2020-16:41:51] [V] [TRT] CudaDepthwiseConvolution has no valid tactics for this config, skipping [11/26/2020-16:41:51] [V] [TRT] >>>>>>>>>>>>>>> Chose Runner Type: CaskConvolution Tactic: 3827454225649558724 [11/26/2020-16:41:51] [V] [TRT] Conv_16 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148n_nt_v1 [11/26/2020-16:41:51] [V] [TRT] [11/26/2020-16:41:51] [V] [TRT] *************** Autotuning format combination: Half(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 1),(MUL_ADD 16 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 4096)) -> Half(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 1),(MUL_ADD 16 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 4096)) *************** [11/26/2020-16:41:51] [V] [TRT] --------------- Timing Runner: Conv_16 (FusedConvActConvolution) [11/26/2020-16:41:51] [V] [TRT] FusedConvActConvolution has no valid tactics for this config, skipping [11/26/2020-16:41:51] [V] [TRT] --------------- Timing Runner: Conv_16 (CaskConvolution) [11/26/2020-16:41:51] [V] [TRT] CaskConvolution has no valid tactics for this config, skipping [11/26/2020-16:41:51] [V] [TRT] --------------- Timing Runner: Conv_16 (CudaConvolution) [11/26/2020-16:41:51] [V] [TRT] Tactic: 0 time 0.408576 [11/26/2020-16:41:51] [V] [TRT] Tactic: 1 time 0.275456 [11/26/2020-16:41:51] [V] [TRT] Tactic: 2 time 0.375808 [11/26/2020-16:41:52] [V] [TRT] Tactic: 5 time 3.848 [11/26/2020-16:41:52] [V] [TRT] Tactic: 6 time 0.111616 [11/26/2020-16:41:52] [V] [TRT] Fastest Tactic: 6 Time: 0.111616 [11/26/2020-16:41:52] [V] [TRT] --------------- Timing Runner: Conv_16 (CudaDepthwiseConvolution) [11/26/2020-16:41:52] [V] [TRT] CudaDepthwiseConvolution has no valid tactics for this config, skipping [11/26/2020-16:41:52] [V] [TRT] >>>>>>>>>>>>>>> Chose Runner Type: CudaConvolution Tactic: 6 [11/26/2020-16:41:52] [V] [TRT] [11/26/2020-16:41:52] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:41:52] [V] [TRT] Tactic: 0 time 0.01024 [11/26/2020-16:41:52] [V] [TRT] Fastest Tactic: 0 Time: 0.01024 [11/26/2020-16:41:52] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:41:52] [V] [TRT] Tactic: 0 time 0.007168 [11/26/2020-16:41:52] [V] [TRT] Fastest Tactic: 0 Time: 0.007168 [11/26/2020-16:41:52] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:41:52] [V] [TRT] Tactic: 0 time 0.014144 [11/26/2020-16:41:52] [V] [TRT] Fastest Tactic: 0 Time: 0.014144 [11/26/2020-16:41:52] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:41:52] [V] [TRT] Tactic: 0 time 0.011264 [11/26/2020-16:41:52] [V] [TRT] Fastest Tactic: 0 Time: 0.011264 [11/26/2020-16:41:52] [V] [TRT] *************** Autotuning format combination: Float(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 1),(MUL_ADD 16 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 2048 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2048)), Float(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 1),(MUL_ADD 16 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 4096)) -> Float(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 1),(MUL_ADD 16 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 4096)) *************** [11/26/2020-16:41:52] [V] [TRT] --------------- Timing Runner: Conv_17 + Add_18 + Relu_19 (LegacySASSConvolution) [11/26/2020-16:41:52] [V] [TRT] Tactic: 0 time 0.025664 [11/26/2020-16:41:52] [V] [TRT] Fastest Tactic: 0 Time: 0.025664 [11/26/2020-16:41:52] [V] [TRT] --------------- Timing Runner: Conv_17 + Add_18 + Relu_19 (CaskConvolution) [11/26/2020-16:41:52] [V] [TRT] Conv_17 + Add_18 + Relu_19 (scudnn) Set Tactic Name: maxwell_scudnn_128x32_relu_medium_nn_v1 [11/26/2020-16:41:52] [V] [TRT] Tactic: 1062367460111450758 time 0.0256 [11/26/2020-16:41:52] [V] [TRT] Conv_17 + Add_18 + Relu_19 (scudnn) Set Tactic Name: maxwell_scudnn_128x128_relu_medium_nn_v1 [11/26/2020-16:41:52] [V] [TRT] Tactic: 4501471010995462441 time 0.023232 [11/26/2020-16:41:52] [V] [TRT] Conv_17 + Add_18 + Relu_19 (scudnn) Set Tactic Name: maxwell_scudnn_128x64_relu_small_nn_v1 [11/26/2020-16:41:52] [V] [TRT] Tactic: 5137655947464784826 time 0.022528 [11/26/2020-16:41:52] [V] [TRT] Conv_17 + Add_18 + Relu_19 (scudnn) Set Tactic Name: maxwell_scudnn_128x128_relu_interior_nn_v1 [11/26/2020-16:41:52] [V] [TRT] Tactic: 5326823351883942011 time 0.022528 [11/26/2020-16:41:52] [V] [TRT] Conv_17 + Add_18 + Relu_19 (scudnn) Set Tactic Name: maxwell_scudnn_128x64_relu_medium_nn_v1 [11/26/2020-16:41:52] [V] [TRT] Tactic: 6645123197870846056 time 0.02336 [11/26/2020-16:41:52] [V] [TRT] Conv_17 + Add_18 + Relu_19 (scudnn) Set Tactic Name: maxwell_scudnn_128x32_relu_interior_nn_v1 [11/26/2020-16:41:52] [V] [TRT] Tactic: -6576203419454146580 time 0.024352 [11/26/2020-16:41:52] [V] [TRT] Conv_17 + Add_18 + Relu_19 (scudnn) Set Tactic Name: maxwell_scudnn_128x32_relu_small_nn_v1 [11/26/2020-16:41:52] [V] [TRT] Tactic: -3456450830548107839 time 0.024576 [11/26/2020-16:41:52] [V] [TRT] Conv_17 + Add_18 + Relu_19 (scudnn) Set Tactic Name: maxwell_scudnn_128x128_relu_small_nn_v1 [11/26/2020-16:41:52] [V] [TRT] Tactic: -410470605513481746 time 0.022528 [11/26/2020-16:41:52] [V] [TRT] Conv_17 + Add_18 + Relu_19 (scudnn) Set Tactic Name: maxwell_scudnn_128x64_relu_interior_nn_v1 [11/26/2020-16:41:52] [V] [TRT] Tactic: -37215280111360163 time 0.022528 [11/26/2020-16:41:52] [V] [TRT] Fastest Tactic: 5137655947464784826 Time: 0.022528 [11/26/2020-16:41:52] [V] [TRT] --------------- Timing Runner: Conv_17 + Add_18 + Relu_19 (CudaConvolution) [11/26/2020-16:41:52] [V] [TRT] Tactic: 0 time 0.048128 [11/26/2020-16:41:52] [V] [TRT] Tactic: 1 time 0.045312 [11/26/2020-16:41:52] [V] [TRT] Tactic: 2 time 0.09728 [11/26/2020-16:41:52] [V] [TRT] Tactic: 5 time 0.177152 [11/26/2020-16:41:52] [V] [TRT] Fastest Tactic: 1 Time: 0.045312 [11/26/2020-16:41:52] [V] [TRT] >>>>>>>>>>>>>>> Chose Runner Type: CaskConvolution Tactic: 5137655947464784826 [11/26/2020-16:41:52] [V] [TRT] Conv_17 + Add_18 + Relu_19 (scudnn) Set Tactic Name: maxwell_scudnn_128x64_relu_small_nn_v1 [11/26/2020-16:41:52] [V] [TRT] [11/26/2020-16:41:52] [V] [TRT] *************** Autotuning format combination: Half(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 1),(MUL_ADD 16 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 2048 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2048)), Half(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 1),(MUL_ADD 16 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 4096)) -> Half(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 1),(MUL_ADD 16 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 4096)) *************** [11/26/2020-16:41:52] [V] [TRT] --------------- Timing Runner: Conv_17 + Add_18 + Relu_19 (CaskConvolution) [11/26/2020-16:41:52] [V] [TRT] CaskConvolution has no valid tactics for this config, skipping [11/26/2020-16:41:52] [V] [TRT] --------------- Timing Runner: Conv_17 + Add_18 + Relu_19 (CudaConvolution) [11/26/2020-16:41:52] [V] [TRT] Tactic: 0 time 0.075776 [11/26/2020-16:41:52] [V] [TRT] Tactic: 1 time 0.043008 [11/26/2020-16:41:52] [V] [TRT] Tactic: 2 time 0.121184 [11/26/2020-16:41:52] [V] [TRT] Tactic: 5 time 0.171008 [11/26/2020-16:41:52] [V] [TRT] Fastest Tactic: 1 Time: 0.043008 [11/26/2020-16:41:52] [V] [TRT] >>>>>>>>>>>>>>> Chose Runner Type: CudaConvolution Tactic: 1 [11/26/2020-16:41:52] [V] [TRT] [11/26/2020-16:41:52] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:41:52] [V] [TRT] Tactic: 0 time 0.014304 [11/26/2020-16:41:52] [V] [TRT] Fastest Tactic: 0 Time: 0.014304 [11/26/2020-16:41:52] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:41:52] [V] [TRT] Tactic: 0 time 0.011072 [11/26/2020-16:41:52] [V] [TRT] Fastest Tactic: 0 Time: 0.011072 [11/26/2020-16:41:52] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:41:52] [V] [TRT] Tactic: 0 time 0.014336 [11/26/2020-16:41:52] [V] [TRT] Fastest Tactic: 0 Time: 0.014336 [11/26/2020-16:41:52] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:41:52] [V] [TRT] Tactic: 0 time 0.011264 [11/26/2020-16:41:52] [V] [TRT] Fastest Tactic: 0 Time: 0.011264 [11/26/2020-16:41:52] [V] [TRT] *************** Autotuning format combination: Float(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 1),(MUL_ADD 16 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 4096)) -> Float(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 1),(MUL_ADD 16 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 4096)) *************** [11/26/2020-16:41:52] [V] [TRT] --------------- Timing Runner: Conv_20 + Relu_21 (LegacySASSConvolution) [11/26/2020-16:41:52] [V] [TRT] Tactic: 0 time 0.288768 [11/26/2020-16:41:52] [V] [TRT] Tactic: 1 time 0.108544 [11/26/2020-16:41:52] [V] [TRT] Fastest Tactic: 1 Time: 0.108544 [11/26/2020-16:41:52] [V] [TRT] --------------- Timing Runner: Conv_20 + Relu_21 (FusedConvActConvolution) [11/26/2020-16:41:52] [V] [TRT] Tactic: 7 time 0.268352 [11/26/2020-16:41:52] [V] [TRT] Tactic: 10 time 0.204608 [11/26/2020-16:41:52] [V] [TRT] Tactic: 14 time 0.165504 [11/26/2020-16:41:52] [V] [TRT] Tactic: 15 time 0.191392 [11/26/2020-16:41:52] [V] [TRT] Tactic: 25 time 0.181024 [11/26/2020-16:41:52] [V] [TRT] Tactic: 26 time 0.271296 [11/26/2020-16:41:52] [V] [TRT] Tactic: 29 time 0.248768 [11/26/2020-16:41:52] [V] [TRT] Tactic: 30 time 0.254976 [11/26/2020-16:41:52] [V] [TRT] Tactic: 33 time 0.289792 [11/26/2020-16:41:52] [V] [TRT] Tactic: 36 time 0.303552 [11/26/2020-16:41:52] [V] [TRT] Tactic: 39 time 0.192512 [11/26/2020-16:41:52] [V] [TRT] Tactic: 41 time 0.236096 [11/26/2020-16:41:52] [V] [TRT] Tactic: 42 time 0.3584 [11/26/2020-16:41:52] [V] [TRT] Tactic: 43 time 0.314368 [11/26/2020-16:41:52] [V] [TRT] Tactic: 45 time 0.165888 [11/26/2020-16:41:52] [V] [TRT] Tactic: 47 time 0.205824 [11/26/2020-16:41:52] [V] [TRT] Tactic: 52 time 0.211968 [11/26/2020-16:41:52] [V] [TRT] Tactic: 54 time 0.215616 [11/26/2020-16:41:52] [V] [TRT] Tactic: 56 time 0.226304 [11/26/2020-16:41:52] [V] [TRT] Tactic: 66 time 0.277504 [11/26/2020-16:41:52] [V] [TRT] Tactic: 76 time 0.18432 [11/26/2020-16:41:52] [V] [TRT] Tactic: 90 time 0.221728 [11/26/2020-16:41:52] [V] [TRT] Tactic: 93 time 0.216064 [11/26/2020-16:41:52] [V] [TRT] Tactic: 98 time 0.266496 [11/26/2020-16:41:52] [V] [TRT] Tactic: 104 time 0.221184 [11/26/2020-16:41:52] [V] [TRT] Tactic: 110 time 0.22016 [11/26/2020-16:41:52] [V] [TRT] Tactic: 119 time 0.234336 [11/26/2020-16:41:52] [V] [TRT] Tactic: 121 time 0.171008 [11/26/2020-16:41:52] [V] [TRT] Tactic: 130 time 0.24672 [11/26/2020-16:41:52] [V] [TRT] Tactic: 134 time 0.241664 [11/26/2020-16:41:52] [V] [TRT] Tactic: 136 time 0.278528 [11/26/2020-16:41:52] [V] [TRT] Tactic: 137 time 0.227328 [11/26/2020-16:41:52] [V] [TRT] Tactic: 139 time 0.212512 [11/26/2020-16:41:52] [V] [TRT] Tactic: 144 time 0.278528 [11/26/2020-16:41:52] [V] [TRT] Tactic: 149 time 0.198656 [11/26/2020-16:41:52] [V] [TRT] Tactic: 151 time 0.22528 [11/26/2020-16:41:52] [V] [TRT] Tactic: 152 time 0.22768 [11/26/2020-16:41:52] [V] [TRT] Tactic: 153 time 0.277504 [11/26/2020-16:41:52] [V] [TRT] Tactic: 156 time 0.247328 [11/26/2020-16:41:52] [V] [TRT] Tactic: 159 time 0.207872 [11/26/2020-16:41:53] [V] [TRT] Tactic: 162 time 0.223232 [11/26/2020-16:41:53] [V] [TRT] Tactic: 164 time 0.166912 [11/26/2020-16:41:53] [V] [TRT] Fastest Tactic: 14 Time: 0.165504 [11/26/2020-16:41:53] [V] [TRT] --------------- Timing Runner: Conv_20 + Relu_21 (CaskConvolution) [11/26/2020-16:41:53] [V] [TRT] Conv_20 + Relu_21 (scudnn) Set Tactic Name: maxwell_scudnn_128x32_relu_medium_nn_v1 [11/26/2020-16:41:53] [V] [TRT] Tactic: 1062367460111450758 time 0.282624 [11/26/2020-16:41:53] [V] [TRT] Conv_20 + Relu_21 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148n_nt_v1 [11/26/2020-16:41:53] [V] [TRT] Tactic: 3827454225649558724 time 0.090112 [11/26/2020-16:41:53] [V] [TRT] Conv_20 + Relu_21 (scudnn) Set Tactic Name: maxwell_scudnn_128x64_relu_large_nn_v1 [11/26/2020-16:41:53] [V] [TRT] Tactic: 4337000649858996379 time 0.256 [11/26/2020-16:41:53] [V] [TRT] Conv_20 + Relu_21 (scudnn) Set Tactic Name: maxwell_scudnn_128x128_relu_medium_nn_v1 [11/26/2020-16:41:53] [V] [TRT] Tactic: 4501471010995462441 time 0.235296 [11/26/2020-16:41:53] [V] [TRT] Conv_20 + Relu_21 (scudnn) Set Tactic Name: maxwell_scudnn_128x64_relu_small_nn_v1 [11/26/2020-16:41:53] [V] [TRT] Tactic: 5137655947464784826 time 0.229376 [11/26/2020-16:41:53] [V] [TRT] Conv_20 + Relu_21 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148m_nt_v1 [11/26/2020-16:41:53] [V] [TRT] Tactic: 5921334924264294896 time 0.091136 [11/26/2020-16:41:53] [V] [TRT] Conv_20 + Relu_21 (scudnn) Set Tactic Name: maxwell_scudnn_128x64_relu_medium_nn_v1 [11/26/2020-16:41:53] [V] [TRT] Tactic: 6645123197870846056 time 0.246784 [11/26/2020-16:41:53] [V] [TRT] Conv_20 + Relu_21 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148t_nt_v1 [11/26/2020-16:41:53] [V] [TRT] Tactic: 7852627285308570038 time 0.092928 [11/26/2020-16:41:53] [V] [TRT] Conv_20 + Relu_21 (scudnn) Set Tactic Name: maxwell_scudnn_128x128_relu_large_nn_v1 [11/26/2020-16:41:53] [V] [TRT] Tactic: -9137461792520977713 time 0.242688 [11/26/2020-16:41:53] [V] [TRT] Conv_20 + Relu_21 (scudnn) Set Tactic Name: maxwell_scudnn_128x32_relu_large_nn_v1 [11/26/2020-16:41:53] [V] [TRT] Tactic: -6092040395344634144 time 0.297984 [11/26/2020-16:41:53] [V] [TRT] Conv_20 + Relu_21 (scudnn) Set Tactic Name: maxwell_scudnn_128x32_relu_small_nn_v1 [11/26/2020-16:41:53] [V] [TRT] Tactic: -3456450830548107839 time 0.246784 [11/26/2020-16:41:53] [V] [TRT] Conv_20 + Relu_21 (scudnn) Set Tactic Name: maxwell_scudnn_128x128_relu_small_nn_v1 [11/26/2020-16:41:53] [V] [TRT] Tactic: -410470605513481746 time 0.227328 [11/26/2020-16:41:53] [V] [TRT] Fastest Tactic: 3827454225649558724 Time: 0.090112 [11/26/2020-16:41:53] [V] [TRT] --------------- Timing Runner: Conv_20 + Relu_21 (CudaConvolution) [11/26/2020-16:41:53] [V] [TRT] Tactic: 0 time 0.376832 [11/26/2020-16:41:53] [V] [TRT] Tactic: 1 time 0.273408 [11/26/2020-16:41:53] [V] [TRT] Tactic: 2 time 0.342016 [11/26/2020-16:41:53] [V] [TRT] Tactic: 5 time 3.7472 [11/26/2020-16:41:53] [V] [TRT] Tactic: 6 time 0.134144 [11/26/2020-16:41:53] [V] [TRT] Fastest Tactic: 6 Time: 0.134144 [11/26/2020-16:41:53] [V] [TRT] --------------- Timing Runner: Conv_20 + Relu_21 (CudaDepthwiseConvolution) [11/26/2020-16:41:53] [V] [TRT] CudaDepthwiseConvolution has no valid tactics for this config, skipping [11/26/2020-16:41:53] [V] [TRT] >>>>>>>>>>>>>>> Chose Runner Type: CaskConvolution Tactic: 3827454225649558724 [11/26/2020-16:41:53] [V] [TRT] Conv_20 + Relu_21 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148n_nt_v1 [11/26/2020-16:41:53] [V] [TRT] [11/26/2020-16:41:53] [V] [TRT] *************** Autotuning format combination: Half(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 1),(MUL_ADD 16 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 4096)) -> Half(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 1),(MUL_ADD 16 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 4096)) *************** [11/26/2020-16:41:53] [V] [TRT] --------------- Timing Runner: Conv_20 + Relu_21 (FusedConvActConvolution) [11/26/2020-16:41:53] [V] [TRT] FusedConvActConvolution has no valid tactics for this config, skipping [11/26/2020-16:41:53] [V] [TRT] --------------- Timing Runner: Conv_20 + Relu_21 (CaskConvolution) [11/26/2020-16:41:53] [V] [TRT] CaskConvolution has no valid tactics for this config, skipping [11/26/2020-16:41:53] [V] [TRT] --------------- Timing Runner: Conv_20 + Relu_21 (CudaConvolution) [11/26/2020-16:41:53] [V] [TRT] Tactic: 0 time 0.417792 [11/26/2020-16:41:53] [V] [TRT] Tactic: 1 time 0.279552 [11/26/2020-16:41:53] [V] [TRT] Tactic: 2 time 0.377856 [11/26/2020-16:41:53] [V] [TRT] Tactic: 5 time 3.87866 [11/26/2020-16:41:53] [V] [TRT] Tactic: 6 time 0.116736 [11/26/2020-16:41:53] [V] [TRT] Fastest Tactic: 6 Time: 0.116736 [11/26/2020-16:41:53] [V] [TRT] --------------- Timing Runner: Conv_20 + Relu_21 (CudaDepthwiseConvolution) [11/26/2020-16:41:53] [V] [TRT] CudaDepthwiseConvolution has no valid tactics for this config, skipping [11/26/2020-16:41:53] [V] [TRT] >>>>>>>>>>>>>>> Chose Runner Type: CudaConvolution Tactic: 6 [11/26/2020-16:41:53] [V] [TRT] [11/26/2020-16:41:53] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:41:53] [V] [TRT] Tactic: 0 time 0.014976 [11/26/2020-16:41:53] [V] [TRT] Fastest Tactic: 0 Time: 0.014976 [11/26/2020-16:41:53] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:41:53] [V] [TRT] Tactic: 0 time 0.011168 [11/26/2020-16:41:53] [V] [TRT] Fastest Tactic: 0 Time: 0.011168 [11/26/2020-16:41:53] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:41:53] [V] [TRT] Tactic: 0 time 0.014336 [11/26/2020-16:41:53] [V] [TRT] Fastest Tactic: 0 Time: 0.014336 [11/26/2020-16:41:53] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:41:53] [V] [TRT] Tactic: 0 time 0.011264 [11/26/2020-16:41:53] [V] [TRT] Fastest Tactic: 0 Time: 0.011264 [11/26/2020-16:41:53] [V] [TRT] *************** Autotuning format combination: Float(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 1),(MUL_ADD 16 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 4096)), Float(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 1),(MUL_ADD 16 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 4096)) -> Float(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 1),(MUL_ADD 16 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 4096)) *************** [11/26/2020-16:41:53] [V] [TRT] --------------- Timing Runner: Conv_22 + Add_23 + Relu_24 (LegacySASSConvolution) [11/26/2020-16:41:53] [V] [TRT] Tactic: 0 time 0.282624 [11/26/2020-16:41:53] [V] [TRT] Tactic: 1 time 0.111616 [11/26/2020-16:41:53] [V] [TRT] Fastest Tactic: 1 Time: 0.111616 [11/26/2020-16:41:53] [V] [TRT] --------------- Timing Runner: Conv_22 + Add_23 + Relu_24 (CaskConvolution) [11/26/2020-16:41:53] [V] [TRT] Conv_22 + Add_23 + Relu_24 (scudnn) Set Tactic Name: maxwell_scudnn_128x32_relu_medium_nn_v1 [11/26/2020-16:41:53] [V] [TRT] Tactic: 1062367460111450758 time 0.284672 [11/26/2020-16:41:53] [V] [TRT] Conv_22 + Add_23 + Relu_24 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148n_nt_v1 [11/26/2020-16:41:53] [V] [TRT] Tactic: 3827454225649558724 time 0.08704 [11/26/2020-16:41:53] [V] [TRT] Conv_22 + Add_23 + Relu_24 (scudnn) Set Tactic Name: maxwell_scudnn_128x64_relu_large_nn_v1 [11/26/2020-16:41:53] [V] [TRT] Tactic: 4337000649858996379 time 0.258048 [11/26/2020-16:41:53] [V] [TRT] Conv_22 + Add_23 + Relu_24 (scudnn) Set Tactic Name: maxwell_scudnn_128x128_relu_medium_nn_v1 [11/26/2020-16:41:53] [V] [TRT] Tactic: 4501471010995462441 time 0.242688 [11/26/2020-16:41:53] [V] [TRT] Conv_22 + Add_23 + Relu_24 (scudnn) Set Tactic Name: maxwell_scudnn_128x64_relu_small_nn_v1 [11/26/2020-16:41:53] [V] [TRT] Tactic: 5137655947464784826 time 0.23552 [11/26/2020-16:41:53] [V] [TRT] Conv_22 + Add_23 + Relu_24 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148m_nt_v1 [11/26/2020-16:41:53] [V] [TRT] Tactic: 5921334924264294896 time 0.09216 [11/26/2020-16:41:53] [V] [TRT] Conv_22 + Add_23 + Relu_24 (scudnn) Set Tactic Name: maxwell_scudnn_128x64_relu_medium_nn_v1 [11/26/2020-16:41:53] [V] [TRT] Tactic: 6645123197870846056 time 0.253952 [11/26/2020-16:41:53] [V] [TRT] Conv_22 + Add_23 + Relu_24 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148t_nt_v1 [11/26/2020-16:41:53] [V] [TRT] Tactic: 7852627285308570038 time 0.093184 [11/26/2020-16:41:53] [V] [TRT] Conv_22 + Add_23 + Relu_24 (scudnn) Set Tactic Name: maxwell_scudnn_128x128_relu_large_nn_v1 [11/26/2020-16:41:53] [V] [TRT] Tactic: -9137461792520977713 time 0.243648 [11/26/2020-16:41:53] [V] [TRT] Conv_22 + Add_23 + Relu_24 (scudnn) Set Tactic Name: maxwell_scudnn_128x32_relu_large_nn_v1 [11/26/2020-16:41:53] [V] [TRT] Tactic: -6092040395344634144 time 0.299008 [11/26/2020-16:41:53] [V] [TRT] Conv_22 + Add_23 + Relu_24 (scudnn) Set Tactic Name: maxwell_scudnn_128x32_relu_small_nn_v1 [11/26/2020-16:41:53] [V] [TRT] Tactic: -3456450830548107839 time 0.251008 [11/26/2020-16:41:53] [V] [TRT] Conv_22 + Add_23 + Relu_24 (scudnn) Set Tactic Name: maxwell_scudnn_128x128_relu_small_nn_v1 [11/26/2020-16:41:53] [V] [TRT] Tactic: -410470605513481746 time 0.22528 [11/26/2020-16:41:53] [V] [TRT] Fastest Tactic: 3827454225649558724 Time: 0.08704 [11/26/2020-16:41:53] [V] [TRT] --------------- Timing Runner: Conv_22 + Add_23 + Relu_24 (CudaConvolution) [11/26/2020-16:41:53] [V] [TRT] Tactic: 0 time 0.382976 [11/26/2020-16:41:53] [V] [TRT] Tactic: 1 time 0.279552 [11/26/2020-16:41:53] [V] [TRT] Tactic: 2 time 0.347712 [11/26/2020-16:41:53] [V] [TRT] Tactic: 5 time 3.51158 [11/26/2020-16:41:53] [V] [TRT] Tactic: 6 time 0.132832 [11/26/2020-16:41:53] [V] [TRT] Fastest Tactic: 6 Time: 0.132832 [11/26/2020-16:41:53] [V] [TRT] >>>>>>>>>>>>>>> Chose Runner Type: CaskConvolution Tactic: 3827454225649558724 [11/26/2020-16:41:53] [V] [TRT] Conv_22 + Add_23 + Relu_24 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148n_nt_v1 [11/26/2020-16:41:53] [V] [TRT] [11/26/2020-16:41:53] [V] [TRT] *************** Autotuning format combination: Half(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 1),(MUL_ADD 16 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 4096)), Half(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 1),(MUL_ADD 16 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 4096)) -> Half(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 1),(MUL_ADD 16 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 4096)) *************** [11/26/2020-16:41:53] [V] [TRT] --------------- Timing Runner: Conv_22 + Add_23 + Relu_24 (CaskConvolution) [11/26/2020-16:41:53] [V] [TRT] CaskConvolution has no valid tactics for this config, skipping [11/26/2020-16:41:53] [V] [TRT] --------------- Timing Runner: Conv_22 + Add_23 + Relu_24 (CudaConvolution) [11/26/2020-16:41:53] [V] [TRT] Tactic: 0 time 0.41984 [11/26/2020-16:41:53] [V] [TRT] Tactic: 1 time 0.29184 [11/26/2020-16:41:53] [V] [TRT] Tactic: 2 time 0.385024 [11/26/2020-16:41:53] [V] [TRT] Tactic: 5 time 3.81747 [11/26/2020-16:41:53] [V] [TRT] Tactic: 6 time 0.120832 [11/26/2020-16:41:53] [V] [TRT] Fastest Tactic: 6 Time: 0.120832 [11/26/2020-16:41:53] [V] [TRT] >>>>>>>>>>>>>>> Chose Runner Type: CudaConvolution Tactic: 6 [11/26/2020-16:41:53] [V] [TRT] [11/26/2020-16:41:53] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:41:53] [V] [TRT] Tactic: 0 time 0.014336 [11/26/2020-16:41:53] [V] [TRT] Fastest Tactic: 0 Time: 0.014336 [11/26/2020-16:41:53] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:41:53] [V] [TRT] Tactic: 0 time 0.011264 [11/26/2020-16:41:53] [V] [TRT] Fastest Tactic: 0 Time: 0.011264 [11/26/2020-16:41:53] [V] [TRT] *************** Autotuning format combination: Float(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 1),(MUL_ADD 16 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 4096)) -> Float(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 1),(MUL_ADD 16 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 4096)) *************** [11/26/2020-16:41:53] [V] [TRT] --------------- Timing Runner: Conv_25 + Relu_26 (LegacySASSConvolution) [11/26/2020-16:41:53] [V] [TRT] Tactic: 0 time 0.280416 [11/26/2020-16:41:53] [V] [TRT] Tactic: 1 time 0.110592 [11/26/2020-16:41:53] [V] [TRT] Fastest Tactic: 1 Time: 0.110592 [11/26/2020-16:41:53] [V] [TRT] --------------- Timing Runner: Conv_25 + Relu_26 (FusedConvActConvolution) [11/26/2020-16:41:53] [V] [TRT] Tactic: 7 time 0.259072 [11/26/2020-16:41:53] [V] [TRT] Tactic: 10 time 0.2048 [11/26/2020-16:41:53] [V] [TRT] Tactic: 14 time 0.166912 [11/26/2020-16:41:53] [V] [TRT] Tactic: 15 time 0.184896 [11/26/2020-16:41:53] [V] [TRT] Tactic: 25 time 0.180224 [11/26/2020-16:41:53] [V] [TRT] Tactic: 26 time 0.262688 [11/26/2020-16:41:53] [V] [TRT] Tactic: 29 time 0.246784 [11/26/2020-16:41:53] [V] [TRT] Tactic: 30 time 0.252928 [11/26/2020-16:41:53] [V] [TRT] Tactic: 33 time 0.287744 [11/26/2020-16:41:53] [V] [TRT] Tactic: 36 time 0.29344 [11/26/2020-16:41:53] [V] [TRT] Tactic: 39 time 0.189344 [11/26/2020-16:41:53] [V] [TRT] Tactic: 41 time 0.23504 [11/26/2020-16:41:53] [V] [TRT] Tactic: 42 time 0.358848 [11/26/2020-16:41:53] [V] [TRT] Tactic: 43 time 0.314912 [11/26/2020-16:41:53] [V] [TRT] Tactic: 45 time 0.166912 [11/26/2020-16:41:53] [V] [TRT] Tactic: 47 time 0.205824 [11/26/2020-16:41:53] [V] [TRT] Tactic: 52 time 0.208448 [11/26/2020-16:41:53] [V] [TRT] Tactic: 54 time 0.216064 [11/26/2020-16:41:53] [V] [TRT] Tactic: 56 time 0.22176 [11/26/2020-16:41:53] [V] [TRT] Tactic: 66 time 0.275456 [11/26/2020-16:41:53] [V] [TRT] Tactic: 76 time 0.186368 [11/26/2020-16:41:53] [V] [TRT] Tactic: 90 time 0.223232 [11/26/2020-16:41:53] [V] [TRT] Tactic: 93 time 0.211968 [11/26/2020-16:41:54] [V] [TRT] Tactic: 98 time 0.260096 [11/26/2020-16:41:54] [V] [TRT] Tactic: 104 time 0.222208 [11/26/2020-16:41:54] [V] [TRT] Tactic: 110 time 0.223232 [11/26/2020-16:41:54] [V] [TRT] Tactic: 119 time 0.234496 [11/26/2020-16:41:54] [V] [TRT] Tactic: 121 time 0.173952 [11/26/2020-16:41:54] [V] [TRT] Tactic: 130 time 0.24576 [11/26/2020-16:41:54] [V] [TRT] Tactic: 134 time 0.24064 [11/26/2020-16:41:54] [V] [TRT] Tactic: 136 time 0.278528 [11/26/2020-16:41:54] [V] [TRT] Tactic: 137 time 0.228352 [11/26/2020-16:41:54] [V] [TRT] Tactic: 139 time 0.216064 [11/26/2020-16:41:54] [V] [TRT] Tactic: 144 time 0.278528 [11/26/2020-16:41:54] [V] [TRT] Tactic: 149 time 0.2048 [11/26/2020-16:41:54] [V] [TRT] Tactic: 151 time 0.227328 [11/26/2020-16:41:54] [V] [TRT] Tactic: 152 time 0.227328 [11/26/2020-16:41:54] [V] [TRT] Tactic: 153 time 0.278336 [11/26/2020-16:41:54] [V] [TRT] Tactic: 156 time 0.245664 [11/26/2020-16:41:54] [V] [TRT] Tactic: 159 time 0.207648 [11/26/2020-16:41:54] [V] [TRT] Tactic: 162 time 0.220064 [11/26/2020-16:41:54] [V] [TRT] Tactic: 164 time 0.168832 [11/26/2020-16:41:54] [V] [TRT] Fastest Tactic: 14 Time: 0.166912 [11/26/2020-16:41:54] [V] [TRT] --------------- Timing Runner: Conv_25 + Relu_26 (CaskConvolution) [11/26/2020-16:41:54] [V] [TRT] Conv_25 + Relu_26 (scudnn) Set Tactic Name: maxwell_scudnn_128x32_relu_medium_nn_v1 [11/26/2020-16:41:54] [V] [TRT] Tactic: 1062367460111450758 time 0.2816 [11/26/2020-16:41:54] [V] [TRT] Conv_25 + Relu_26 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148n_nt_v1 [11/26/2020-16:41:54] [V] [TRT] Tactic: 3827454225649558724 time 0.086016 [11/26/2020-16:41:54] [V] [TRT] Conv_25 + Relu_26 (scudnn) Set Tactic Name: maxwell_scudnn_128x64_relu_large_nn_v1 [11/26/2020-16:41:54] [V] [TRT] Tactic: 4337000649858996379 time 0.256 [11/26/2020-16:41:54] [V] [TRT] Conv_25 + Relu_26 (scudnn) Set Tactic Name: maxwell_scudnn_128x128_relu_medium_nn_v1 [11/26/2020-16:41:54] [V] [TRT] Tactic: 4501471010995462441 time 0.235616 [11/26/2020-16:41:54] [V] [TRT] Conv_25 + Relu_26 (scudnn) Set Tactic Name: maxwell_scudnn_128x64_relu_small_nn_v1 [11/26/2020-16:41:54] [V] [TRT] Tactic: 5137655947464784826 time 0.228352 [11/26/2020-16:41:54] [V] [TRT] Conv_25 + Relu_26 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148m_nt_v1 [11/26/2020-16:41:54] [V] [TRT] Tactic: 5921334924264294896 time 0.091136 [11/26/2020-16:41:54] [V] [TRT] Conv_25 + Relu_26 (scudnn) Set Tactic Name: maxwell_scudnn_128x64_relu_medium_nn_v1 [11/26/2020-16:41:54] [V] [TRT] Tactic: 6645123197870846056 time 0.243712 [11/26/2020-16:41:54] [V] [TRT] Conv_25 + Relu_26 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148t_nt_v1 [11/26/2020-16:41:54] [V] [TRT] Tactic: 7852627285308570038 time 0.091136 [11/26/2020-16:41:54] [V] [TRT] Conv_25 + Relu_26 (scudnn) Set Tactic Name: maxwell_scudnn_128x128_relu_large_nn_v1 [11/26/2020-16:41:54] [V] [TRT] Tactic: -9137461792520977713 time 0.243808 [11/26/2020-16:41:54] [V] [TRT] Conv_25 + Relu_26 (scudnn) Set Tactic Name: maxwell_scudnn_128x32_relu_large_nn_v1 [11/26/2020-16:41:54] [V] [TRT] Tactic: -6092040395344634144 time 0.295936 [11/26/2020-16:41:54] [V] [TRT] Conv_25 + Relu_26 (scudnn) Set Tactic Name: maxwell_scudnn_128x32_relu_small_nn_v1 [11/26/2020-16:41:54] [V] [TRT] Tactic: -3456450830548107839 time 0.247808 [11/26/2020-16:41:54] [V] [TRT] Conv_25 + Relu_26 (scudnn) Set Tactic Name: maxwell_scudnn_128x128_relu_small_nn_v1 [11/26/2020-16:41:54] [V] [TRT] Tactic: -410470605513481746 time 0.2304 [11/26/2020-16:41:54] [V] [TRT] Fastest Tactic: 3827454225649558724 Time: 0.086016 [11/26/2020-16:41:54] [V] [TRT] --------------- Timing Runner: Conv_25 + Relu_26 (CudaConvolution) [11/26/2020-16:41:54] [V] [TRT] Tactic: 0 time 0.377856 [11/26/2020-16:41:54] [V] [TRT] Tactic: 1 time 0.273408 [11/26/2020-16:41:54] [V] [TRT] Tactic: 2 time 0.34304 [11/26/2020-16:41:54] [V] [TRT] Tactic: 5 time 3.5 [11/26/2020-16:41:54] [V] [TRT] Tactic: 6 time 0.124928 [11/26/2020-16:41:54] [V] [TRT] Fastest Tactic: 6 Time: 0.124928 [11/26/2020-16:41:54] [V] [TRT] --------------- Timing Runner: Conv_25 + Relu_26 (CudaDepthwiseConvolution) [11/26/2020-16:41:54] [V] [TRT] CudaDepthwiseConvolution has no valid tactics for this config, skipping [11/26/2020-16:41:54] [V] [TRT] >>>>>>>>>>>>>>> Chose Runner Type: CaskConvolution Tactic: 3827454225649558724 [11/26/2020-16:41:54] [V] [TRT] Conv_25 + Relu_26 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148n_nt_v1 [11/26/2020-16:41:54] [V] [TRT] [11/26/2020-16:41:54] [V] [TRT] *************** Autotuning format combination: Half(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 1),(MUL_ADD 16 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 4096)) -> Half(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 1),(MUL_ADD 16 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 4096)) *************** [11/26/2020-16:41:54] [V] [TRT] --------------- Timing Runner: Conv_25 + Relu_26 (FusedConvActConvolution) [11/26/2020-16:41:54] [V] [TRT] FusedConvActConvolution has no valid tactics for this config, skipping [11/26/2020-16:41:54] [V] [TRT] --------------- Timing Runner: Conv_25 + Relu_26 (CaskConvolution) [11/26/2020-16:41:54] [V] [TRT] CaskConvolution has no valid tactics for this config, skipping [11/26/2020-16:41:54] [V] [TRT] --------------- Timing Runner: Conv_25 + Relu_26 (CudaConvolution) [11/26/2020-16:41:54] [V] [TRT] Tactic: 0 time 0.410624 [11/26/2020-16:41:54] [V] [TRT] Tactic: 1 time 0.27648 [11/26/2020-16:41:54] [V] [TRT] Tactic: 2 time 0.377856 [11/26/2020-16:41:54] [V] [TRT] Tactic: 5 time 3.75194 [11/26/2020-16:41:54] [V] [TRT] Tactic: 6 time 0.11776 [11/26/2020-16:41:54] [V] [TRT] Fastest Tactic: 6 Time: 0.11776 [11/26/2020-16:41:54] [V] [TRT] --------------- Timing Runner: Conv_25 + Relu_26 (CudaDepthwiseConvolution) [11/26/2020-16:41:54] [V] [TRT] CudaDepthwiseConvolution has no valid tactics for this config, skipping [11/26/2020-16:41:54] [V] [TRT] >>>>>>>>>>>>>>> Chose Runner Type: CudaConvolution Tactic: 6 [11/26/2020-16:41:54] [V] [TRT] [11/26/2020-16:41:54] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:41:54] [V] [TRT] Tactic: 0 time 0.016384 [11/26/2020-16:41:54] [V] [TRT] Fastest Tactic: 0 Time: 0.016384 [11/26/2020-16:41:54] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:41:54] [V] [TRT] Tactic: 0 time 0.011264 [11/26/2020-16:41:54] [V] [TRT] Fastest Tactic: 0 Time: 0.011264 [11/26/2020-16:41:54] [V] [TRT] *************** Autotuning format combination: Float(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 1),(MUL_ADD 16 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 4096)) -> Float(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2),(MUL_ADD 8 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 2048 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 4096)) *************** [11/26/2020-16:41:54] [V] [TRT] --------------- Timing Runner: MaxPool_27 (Pooling) [11/26/2020-16:41:54] [V] [TRT] Tactic: -1 time 0.011264 [11/26/2020-16:41:54] [V] [TRT] Fastest Tactic: -1 Time: 0.011264 [11/26/2020-16:41:54] [V] [TRT] --------------- Timing Runner: MaxPool_27 (TiledPooling) [11/26/2020-16:41:54] [V] [TRT] Tactic: 7340289 time 0.237312 [11/26/2020-16:41:54] [V] [TRT] Tactic: 7340290 time 0.121696 [11/26/2020-16:41:54] [V] [TRT] Tactic: 7340292 time 0.064192 [11/26/2020-16:41:54] [V] [TRT] Tactic: 7340293 time 0.052224 [11/26/2020-16:41:54] [V] [TRT] Tactic: 7340545 time 0.120672 [11/26/2020-16:41:54] [V] [TRT] Tactic: 7340546 time 0.062304 [11/26/2020-16:41:54] [V] [TRT] Tactic: 7340548 time 0.033792 [11/26/2020-16:41:54] [V] [TRT] Tactic: 7340549 time 0.02848 [11/26/2020-16:41:54] [V] [TRT] Tactic: 7340801 time 0.090944 [11/26/2020-16:41:54] [V] [TRT] Tactic: 7340802 time 0.047104 [11/26/2020-16:41:54] [V] [TRT] Tactic: 7340804 time 0.0264 [11/26/2020-16:41:54] [V] [TRT] Tactic: 7340805 time 0.03072 [11/26/2020-16:41:54] [V] [TRT] Tactic: 7341057 time 0.06144 [11/26/2020-16:41:54] [V] [TRT] Tactic: 7341058 time 0.032768 [11/26/2020-16:41:54] [V] [TRT] Tactic: 7341060 time 0.0256 [11/26/2020-16:41:54] [V] [TRT] Tactic: 7341061 time 0.021504 [11/26/2020-16:41:54] [V] [TRT] Tactic: 7341313 time 0.06144 [11/26/2020-16:41:54] [V] [TRT] Tactic: 7341314 time 0.033568 [11/26/2020-16:41:54] [V] [TRT] Tactic: 7341316 time 0.0256 [11/26/2020-16:41:54] [V] [TRT] Tactic: 7341317 time 0.021504 [11/26/2020-16:41:54] [V] [TRT] Tactic: 7341569 time 0.062464 [11/26/2020-16:41:54] [V] [TRT] Tactic: 7341570 time 0.046624 [11/26/2020-16:41:54] [V] [TRT] Tactic: 7341572 time 0.025568 [11/26/2020-16:41:54] [V] [TRT] Tactic: 7341573 time 0.023552 [11/26/2020-16:41:54] [V] [TRT] Tactic: 7341825 time 0.062464 [11/26/2020-16:41:54] [V] [TRT] Tactic: 7341826 time 0.047104 [11/26/2020-16:41:54] [V] [TRT] Tactic: 7341828 time 0.028672 [11/26/2020-16:41:54] [V] [TRT] Tactic: 7341829 time 0.026624 [11/26/2020-16:41:54] [V] [TRT] Tactic: 7342081 time 0.032768 [11/26/2020-16:41:54] [V] [TRT] Tactic: 7342082 time 0.026624 [11/26/2020-16:41:54] [V] [TRT] Tactic: 7342084 time 0.01536 [11/26/2020-16:41:54] [V] [TRT] Tactic: 7342085 time 0.01536 [11/26/2020-16:41:54] [V] [TRT] Fastest Tactic: 7342084 Time: 0.01536 [11/26/2020-16:41:54] [V] [TRT] >>>>>>>>>>>>>>> Chose Runner Type: Pooling Tactic: -1 [11/26/2020-16:41:54] [V] [TRT] [11/26/2020-16:41:54] [V] [TRT] *************** Autotuning format combination: Half(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 1),(MUL_ADD 16 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 4096)) -> Half(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2),(MUL_ADD 8 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 2048 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 4096)) *************** [11/26/2020-16:41:54] [V] [TRT] --------------- Timing Runner: MaxPool_27 (Pooling) [11/26/2020-16:41:54] [V] [TRT] Tactic: -1 time 0.01104 [11/26/2020-16:41:54] [V] [TRT] Fastest Tactic: -1 Time: 0.01104 [11/26/2020-16:41:54] [V] [TRT] --------------- Timing Runner: MaxPool_27 (TiledPooling) [11/26/2020-16:41:54] [V] [TRT] TiledPooling has no valid tactics for this config, skipping [11/26/2020-16:41:54] [V] [TRT] >>>>>>>>>>>>>>> Chose Runner Type: Pooling Tactic: -1 [11/26/2020-16:41:54] [V] [TRT] [11/26/2020-16:41:54] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:41:54] [V] [TRT] Tactic: 0 time 0.009216 [11/26/2020-16:41:54] [V] [TRT] Fastest Tactic: 0 Time: 0.009216 [11/26/2020-16:41:54] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:41:54] [V] [TRT] Tactic: 0 time 0.007168 [11/26/2020-16:41:54] [V] [TRT] Fastest Tactic: 0 Time: 0.007168 [11/26/2020-16:41:54] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:41:54] [V] [TRT] Tactic: 0 time 0.008864 [11/26/2020-16:41:54] [V] [TRT] Fastest Tactic: 0 Time: 0.008864 [11/26/2020-16:41:54] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:41:54] [V] [TRT] Tactic: 0 time 0.00688 [11/26/2020-16:41:54] [V] [TRT] Fastest Tactic: 0 Time: 0.00688 [11/26/2020-16:41:54] [V] [TRT] *************** Autotuning format combination: Float(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2),(MUL_ADD 8 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 2048 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 4096)) -> Float(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2),(MUL_ADD 8 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 8192)) *************** [11/26/2020-16:41:54] [V] [TRT] --------------- Timing Runner: Conv_28 + Relu_29 (LegacySASSConvolution) [11/26/2020-16:41:54] [V] [TRT] Tactic: 0 time 0.280576 [11/26/2020-16:41:54] [V] [TRT] Tactic: 1 time 0.123424 [11/26/2020-16:41:54] [V] [TRT] Fastest Tactic: 1 Time: 0.123424 [11/26/2020-16:41:54] [V] [TRT] --------------- Timing Runner: Conv_28 + Relu_29 (FusedConvActConvolution) [11/26/2020-16:41:54] [V] [TRT] Tactic: 7 time 0.308 [11/26/2020-16:41:54] [V] [TRT] Tactic: 10 time 0.3072 [11/26/2020-16:41:54] [V] [TRT] Tactic: 14 time 0.16384 [11/26/2020-16:41:54] [V] [TRT] Tactic: 15 time 0.195584 [11/26/2020-16:41:54] [V] [TRT] Tactic: 25 time 0.190464 [11/26/2020-16:41:54] [V] [TRT] Tactic: 26 time 0.269312 [11/26/2020-16:41:54] [V] [TRT] Tactic: 29 time 0.248832 [11/26/2020-16:41:54] [V] [TRT] Tactic: 30 time 0.251584 [11/26/2020-16:41:54] [V] [TRT] Tactic: 33 time 0.318464 [11/26/2020-16:41:54] [V] [TRT] Tactic: 36 time 0.443392 [11/26/2020-16:41:55] [V] [TRT] Tactic: 39 time 0.195584 [11/26/2020-16:41:55] [V] [TRT] Tactic: 41 time 0.244736 [11/26/2020-16:41:55] [V] [TRT] Tactic: 42 time 0.3328 [11/26/2020-16:41:55] [V] [TRT] Tactic: 43 time 0.314368 [11/26/2020-16:41:55] [V] [TRT] Tactic: 45 time 0.167936 [11/26/2020-16:41:55] [V] [TRT] Tactic: 47 time 0.205824 [11/26/2020-16:41:55] [V] [TRT] Tactic: 52 time 0.20992 [11/26/2020-16:41:55] [V] [TRT] Tactic: 54 time 0.312864 [11/26/2020-16:41:55] [V] [TRT] Tactic: 56 time 0.229376 [11/26/2020-16:41:55] [V] [TRT] Tactic: 66 time 0.35872 [11/26/2020-16:41:55] [V] [TRT] Tactic: 76 time 0.190464 [11/26/2020-16:41:55] [V] [TRT] Tactic: 90 time 0.22416 [11/26/2020-16:41:55] [V] [TRT] Tactic: 93 time 0.210944 [11/26/2020-16:41:55] [V] [TRT] Tactic: 98 time 0.32768 [11/26/2020-16:41:55] [V] [TRT] Tactic: 104 time 0.278528 [11/26/2020-16:41:55] [V] [TRT] Tactic: 110 time 0.238592 [11/26/2020-16:41:55] [V] [TRT] Tactic: 119 time 0.326496 [11/26/2020-16:41:55] [V] [TRT] Tactic: 121 time 0.22016 [11/26/2020-16:41:55] [V] [TRT] Tactic: 130 time 0.299552 [11/26/2020-16:41:55] [V] [TRT] Tactic: 134 time 0.311296 [11/26/2020-16:41:55] [V] [TRT] Tactic: 136 time 0.279552 [11/26/2020-16:41:55] [V] [TRT] Tactic: 137 time 0.29984 [11/26/2020-16:41:55] [V] [TRT] Tactic: 139 time 0.304128 [11/26/2020-16:41:55] [V] [TRT] Tactic: 144 time 0.276032 [11/26/2020-16:41:55] [V] [TRT] Tactic: 149 time 0.197856 [11/26/2020-16:41:55] [V] [TRT] Tactic: 151 time 0.324608 [11/26/2020-16:41:55] [V] [TRT] Tactic: 152 time 0.233472 [11/26/2020-16:41:55] [V] [TRT] Tactic: 153 time 0.355136 [11/26/2020-16:41:55] [V] [TRT] Tactic: 156 time 0.249856 [11/26/2020-16:41:55] [V] [TRT] Tactic: 159 time 0.218112 [11/26/2020-16:41:55] [V] [TRT] Tactic: 162 time 0.254528 [11/26/2020-16:41:55] [V] [TRT] Tactic: 164 time 0.196608 [11/26/2020-16:41:55] [V] [TRT] Fastest Tactic: 14 Time: 0.16384 [11/26/2020-16:41:55] [V] [TRT] --------------- Timing Runner: Conv_28 + Relu_29 (CaskConvolution) [11/26/2020-16:41:55] [V] [TRT] Conv_28 + Relu_29 (scudnn) Set Tactic Name: maxwell_scudnn_128x32_relu_medium_nn_v1 [11/26/2020-16:41:55] [V] [TRT] Tactic: 1062367460111450758 time 0.284672 [11/26/2020-16:41:55] [V] [TRT] Conv_28 + Relu_29 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148n_nt_v1 [11/26/2020-16:41:55] [V] [TRT] Tactic: 3827454225649558724 time 0.111488 [11/26/2020-16:41:55] [V] [TRT] Conv_28 + Relu_29 (scudnn) Set Tactic Name: maxwell_scudnn_128x64_relu_large_nn_v1 [11/26/2020-16:41:55] [V] [TRT] Tactic: 4337000649858996379 time 0.262144 [11/26/2020-16:41:55] [V] [TRT] Conv_28 + Relu_29 (scudnn) Set Tactic Name: maxwell_scudnn_128x128_relu_medium_nn_v1 [11/26/2020-16:41:55] [V] [TRT] Tactic: 4501471010995462441 time 0.249856 [11/26/2020-16:41:55] [V] [TRT] Conv_28 + Relu_29 (scudnn) Set Tactic Name: maxwell_scudnn_128x64_relu_small_nn_v1 [11/26/2020-16:41:55] [V] [TRT] Tactic: 5137655947464784826 time 0.233024 [11/26/2020-16:41:55] [V] [TRT] Conv_28 + Relu_29 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148m_nt_v1 [11/26/2020-16:41:55] [V] [TRT] Tactic: 5921334924264294896 time 0.109568 [11/26/2020-16:41:55] [V] [TRT] Conv_28 + Relu_29 (scudnn) Set Tactic Name: maxwell_scudnn_128x64_relu_medium_nn_v1 [11/26/2020-16:41:55] [V] [TRT] Tactic: 6645123197870846056 time 0.251072 [11/26/2020-16:41:55] [V] [TRT] Conv_28 + Relu_29 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148t_nt_v1 [11/26/2020-16:41:55] [V] [TRT] Tactic: 7852627285308570038 time 0.113664 [11/26/2020-16:41:55] [V] [TRT] Conv_28 + Relu_29 (scudnn) Set Tactic Name: maxwell_scudnn_128x128_relu_large_nn_v1 [11/26/2020-16:41:55] [V] [TRT] Tactic: -9137461792520977713 time 0.253952 [11/26/2020-16:41:55] [V] [TRT] Conv_28 + Relu_29 (scudnn) Set Tactic Name: maxwell_scudnn_128x32_relu_large_nn_v1 [11/26/2020-16:41:55] [V] [TRT] Tactic: -6092040395344634144 time 0.3072 [11/26/2020-16:41:55] [V] [TRT] Conv_28 + Relu_29 (scudnn) Set Tactic Name: maxwell_scudnn_128x32_relu_small_nn_v1 [11/26/2020-16:41:55] [V] [TRT] Tactic: -3456450830548107839 time 0.250368 [11/26/2020-16:41:55] [V] [TRT] Conv_28 + Relu_29 (scudnn) Set Tactic Name: maxwell_scudnn_128x128_relu_small_nn_v1 [11/26/2020-16:41:55] [V] [TRT] Tactic: -410470605513481746 time 0.241664 [11/26/2020-16:41:55] [V] [TRT] Fastest Tactic: 5921334924264294896 Time: 0.109568 [11/26/2020-16:41:55] [V] [TRT] --------------- Timing Runner: Conv_28 + Relu_29 (CudaConvolution) [11/26/2020-16:41:55] [V] [TRT] Tactic: 0 time 0.371712 [11/26/2020-16:41:55] [V] [TRT] Tactic: 1 time 0.277504 [11/26/2020-16:41:55] [V] [TRT] Tactic: 2 time 0.35328 [11/26/2020-16:41:55] [V] [TRT] Tactic: 5 time 6.92317 [11/26/2020-16:41:55] [V] [TRT] Tactic: 6 time 0.160768 [11/26/2020-16:41:55] [V] [TRT] Fastest Tactic: 6 Time: 0.160768 [11/26/2020-16:41:55] [V] [TRT] --------------- Timing Runner: Conv_28 + Relu_29 (CudaDepthwiseConvolution) [11/26/2020-16:41:55] [V] [TRT] CudaDepthwiseConvolution has no valid tactics for this config, skipping [11/26/2020-16:41:55] [V] [TRT] >>>>>>>>>>>>>>> Chose Runner Type: CaskConvolution Tactic: 5921334924264294896 [11/26/2020-16:41:55] [V] [TRT] Conv_28 + Relu_29 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148m_nt_v1 [11/26/2020-16:41:55] [V] [TRT] [11/26/2020-16:41:55] [V] [TRT] *************** Autotuning format combination: Half(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2),(MUL_ADD 8 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 2048 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 4096)) -> Half(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2),(MUL_ADD 8 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 8192)) *************** [11/26/2020-16:41:55] [V] [TRT] --------------- Timing Runner: Conv_28 + Relu_29 (FusedConvActConvolution) [11/26/2020-16:41:55] [V] [TRT] FusedConvActConvolution has no valid tactics for this config, skipping [11/26/2020-16:41:55] [V] [TRT] --------------- Timing Runner: Conv_28 + Relu_29 (CaskConvolution) [11/26/2020-16:41:55] [V] [TRT] CaskConvolution has no valid tactics for this config, skipping [11/26/2020-16:41:55] [V] [TRT] --------------- Timing Runner: Conv_28 + Relu_29 (CudaConvolution) [11/26/2020-16:41:55] [V] [TRT] Tactic: 0 time 0.4096 [11/26/2020-16:41:55] [V] [TRT] Tactic: 1 time 0.294912 [11/26/2020-16:41:55] [V] [TRT] Tactic: 2 time 0.364544 [11/26/2020-16:41:55] [V] [TRT] Tactic: 5 time 6.87002 [11/26/2020-16:41:55] [V] [TRT] Tactic: 6 time 0.155648 [11/26/2020-16:41:55] [V] [TRT] Fastest Tactic: 6 Time: 0.155648 [11/26/2020-16:41:55] [V] [TRT] --------------- Timing Runner: Conv_28 + Relu_29 (CudaDepthwiseConvolution) [11/26/2020-16:41:55] [V] [TRT] CudaDepthwiseConvolution has no valid tactics for this config, skipping [11/26/2020-16:41:55] [V] [TRT] >>>>>>>>>>>>>>> Chose Runner Type: CudaConvolution Tactic: 6 [11/26/2020-16:41:55] [V] [TRT] [11/26/2020-16:41:55] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:41:55] [V] [TRT] Tactic: 0 time 0.015008 [11/26/2020-16:41:55] [V] [TRT] Fastest Tactic: 0 Time: 0.015008 [11/26/2020-16:41:55] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:41:55] [V] [TRT] Tactic: 0 time 0.011072 [11/26/2020-16:41:55] [V] [TRT] Fastest Tactic: 0 Time: 0.011072 [11/26/2020-16:41:55] [V] [TRT] *************** Autotuning format combination: Float(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2),(MUL_ADD 8 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 8192)) -> Float(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2),(MUL_ADD 8 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 8192)) *************** [11/26/2020-16:41:55] [V] [TRT] --------------- Timing Runner: Conv_30 (LegacySASSConvolution) [11/26/2020-16:41:55] [V] [TRT] Tactic: 0 time 0.551936 [11/26/2020-16:41:55] [V] [TRT] Tactic: 1 time 0.227328 [11/26/2020-16:41:55] [V] [TRT] Fastest Tactic: 1 Time: 0.227328 [11/26/2020-16:41:55] [V] [TRT] --------------- Timing Runner: Conv_30 (FusedConvActConvolution) [11/26/2020-16:41:55] [V] [TRT] Tactic: 7 time 0.566272 [11/26/2020-16:41:55] [V] [TRT] Tactic: 10 time 0.526336 [11/26/2020-16:41:55] [V] [TRT] Tactic: 14 time 0.316416 [11/26/2020-16:41:55] [V] [TRT] Tactic: 15 time 0.34304 [11/26/2020-16:41:55] [V] [TRT] Tactic: 25 time 0.342016 [11/26/2020-16:41:55] [V] [TRT] Tactic: 26 time 0.503808 [11/26/2020-16:41:55] [V] [TRT] Tactic: 29 time 0.468992 [11/26/2020-16:41:55] [V] [TRT] Tactic: 30 time 0.48896 [11/26/2020-16:41:56] [V] [TRT] Tactic: 33 time 0.615424 [11/26/2020-16:41:56] [V] [TRT] Tactic: 36 time 0.864256 [11/26/2020-16:41:56] [V] [TRT] Tactic: 39 time 0.374432 [11/26/2020-16:41:56] [V] [TRT] Tactic: 41 time 0.448512 [11/26/2020-16:41:56] [V] [TRT] Tactic: 42 time 0.620544 [11/26/2020-16:41:56] [V] [TRT] Tactic: 43 time 0.611328 [11/26/2020-16:41:56] [V] [TRT] Tactic: 45 time 0.318464 [11/26/2020-16:41:56] [V] [TRT] Tactic: 47 time 0.400384 [11/26/2020-16:41:56] [V] [TRT] Tactic: 52 time 0.39424 [11/26/2020-16:41:56] [V] [TRT] Tactic: 54 time 0.608256 [11/26/2020-16:41:56] [V] [TRT] Tactic: 56 time 0.436224 [11/26/2020-16:41:56] [V] [TRT] Tactic: 66 time 0.6912 [11/26/2020-16:41:56] [V] [TRT] Tactic: 76 time 0.384832 [11/26/2020-16:41:56] [V] [TRT] Tactic: 90 time 0.421792 [11/26/2020-16:41:56] [V] [TRT] Tactic: 93 time 0.385024 [11/26/2020-16:41:56] [V] [TRT] Tactic: 98 time 0.570368 [11/26/2020-16:41:56] [V] [TRT] Tactic: 104 time 0.534816 [11/26/2020-16:41:56] [V] [TRT] Tactic: 110 time 0.453632 [11/26/2020-16:41:56] [V] [TRT] Tactic: 119 time 0.62064 [11/26/2020-16:41:56] [V] [TRT] Tactic: 121 time 0.4096 [11/26/2020-16:41:56] [V] [TRT] Tactic: 130 time 0.570368 [11/26/2020-16:41:56] [V] [TRT] Tactic: 134 time 0.590848 [11/26/2020-16:41:56] [V] [TRT] Tactic: 136 time 0.540672 [11/26/2020-16:41:56] [V] [TRT] Tactic: 137 time 0.571232 [11/26/2020-16:41:56] [V] [TRT] Tactic: 139 time 0.607232 [11/26/2020-16:41:56] [V] [TRT] Tactic: 144 time 0.533504 [11/26/2020-16:41:56] [V] [TRT] Tactic: 149 time 0.388032 [11/26/2020-16:41:56] [V] [TRT] Tactic: 151 time 0.62768 [11/26/2020-16:41:56] [V] [TRT] Tactic: 152 time 0.47616 [11/26/2020-16:41:56] [V] [TRT] Tactic: 153 time 0.69216 [11/26/2020-16:41:56] [V] [TRT] Tactic: 156 time 0.468992 [11/26/2020-16:41:56] [V] [TRT] Tactic: 159 time 0.433152 [11/26/2020-16:41:56] [V] [TRT] Tactic: 162 time 0.506752 [11/26/2020-16:41:56] [V] [TRT] Tactic: 164 time 0.365568 [11/26/2020-16:41:56] [V] [TRT] Fastest Tactic: 14 Time: 0.316416 [11/26/2020-16:41:56] [V] [TRT] --------------- Timing Runner: Conv_30 (CaskConvolution) [11/26/2020-16:41:56] [V] [TRT] Conv_30 (scudnn) Set Tactic Name: maxwell_scudnn_128x32_relu_medium_nn_v1 [11/26/2020-16:41:56] [V] [TRT] Tactic: 1062367460111450758 time 0.557056 [11/26/2020-16:41:56] [V] [TRT] Conv_30 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148n_nt_v1 [11/26/2020-16:41:56] [V] [TRT] Tactic: 3827454225649558724 time 0.195584 [11/26/2020-16:41:56] [V] [TRT] Conv_30 (scudnn) Set Tactic Name: maxwell_scudnn_128x64_relu_large_nn_v1 [11/26/2020-16:41:56] [V] [TRT] Tactic: 4337000649858996379 time 0.51712 [11/26/2020-16:41:56] [V] [TRT] Conv_30 (scudnn) Set Tactic Name: maxwell_scudnn_128x128_relu_medium_nn_v1 [11/26/2020-16:41:56] [V] [TRT] Tactic: 4501471010995462441 time 0.48208 [11/26/2020-16:41:56] [V] [TRT] Conv_30 (scudnn) Set Tactic Name: maxwell_scudnn_128x64_relu_small_nn_v1 [11/26/2020-16:41:56] [V] [TRT] Tactic: 5137655947464784826 time 0.461824 [11/26/2020-16:41:56] [V] [TRT] Conv_30 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148m_nt_v1 [11/26/2020-16:41:56] [V] [TRT] Tactic: 5921334924264294896 time 0.208896 [11/26/2020-16:41:56] [V] [TRT] Conv_30 (scudnn) Set Tactic Name: maxwell_scudnn_128x64_relu_medium_nn_v1 [11/26/2020-16:41:56] [V] [TRT] Tactic: 6645123197870846056 time 0.49664 [11/26/2020-16:41:56] [V] [TRT] Conv_30 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148t_nt_v1 [11/26/2020-16:41:56] [V] [TRT] Tactic: 7852627285308570038 time 0.212224 [11/26/2020-16:41:56] [V] [TRT] Conv_30 (scudnn) Set Tactic Name: maxwell_scudnn_128x128_relu_large_nn_v1 [11/26/2020-16:41:56] [V] [TRT] Tactic: -9137461792520977713 time 0.507904 [11/26/2020-16:41:56] [V] [TRT] Conv_30 (scudnn) Set Tactic Name: maxwell_scudnn_128x32_relu_large_nn_v1 [11/26/2020-16:41:56] [V] [TRT] Tactic: -6092040395344634144 time 0.603136 [11/26/2020-16:41:56] [V] [TRT] Conv_30 (scudnn) Set Tactic Name: maxwell_scudnn_128x32_relu_small_nn_v1 [11/26/2020-16:41:56] [V] [TRT] Tactic: -3456450830548107839 time 0.485376 [11/26/2020-16:41:56] [V] [TRT] Conv_30 (scudnn) Set Tactic Name: maxwell_scudnn_128x128_relu_small_nn_v1 [11/26/2020-16:41:56] [V] [TRT] Tactic: -410470605513481746 time 0.478208 [11/26/2020-16:41:56] [V] [TRT] Fastest Tactic: 3827454225649558724 Time: 0.195584 [11/26/2020-16:41:56] [V] [TRT] --------------- Timing Runner: Conv_30 (CudaConvolution) [11/26/2020-16:41:56] [V] [TRT] Tactic: 0 time 0.715168 [11/26/2020-16:41:56] [V] [TRT] Tactic: 1 time 0.536576 [11/26/2020-16:41:56] [V] [TRT] Tactic: 2 time 0.641024 [11/26/2020-16:41:56] [V] [TRT] Tactic: 5 time 15.444 [11/26/2020-16:41:56] [V] [TRT] Tactic: 6 time 0.29184 [11/26/2020-16:41:56] [V] [TRT] Fastest Tactic: 6 Time: 0.29184 [11/26/2020-16:41:56] [V] [TRT] --------------- Timing Runner: Conv_30 (CudaDepthwiseConvolution) [11/26/2020-16:41:56] [V] [TRT] CudaDepthwiseConvolution has no valid tactics for this config, skipping [11/26/2020-16:41:56] [V] [TRT] >>>>>>>>>>>>>>> Chose Runner Type: CaskConvolution Tactic: 3827454225649558724 [11/26/2020-16:41:56] [V] [TRT] Conv_30 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148n_nt_v1 [11/26/2020-16:41:56] [V] [TRT] [11/26/2020-16:41:56] [V] [TRT] *************** Autotuning format combination: Half(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2),(MUL_ADD 8 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 8192)) -> Half(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2),(MUL_ADD 8 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 8192)) *************** [11/26/2020-16:41:56] [V] [TRT] --------------- Timing Runner: Conv_30 (FusedConvActConvolution) [11/26/2020-16:41:56] [V] [TRT] FusedConvActConvolution has no valid tactics for this config, skipping [11/26/2020-16:41:56] [V] [TRT] --------------- Timing Runner: Conv_30 (CaskConvolution) [11/26/2020-16:41:56] [V] [TRT] CaskConvolution has no valid tactics for this config, skipping [11/26/2020-16:41:56] [V] [TRT] --------------- Timing Runner: Conv_30 (CudaConvolution) [11/26/2020-16:41:56] [V] [TRT] Tactic: 0 time 0.77216 [11/26/2020-16:41:57] [V] [TRT] Tactic: 1 time 0.552224 [11/26/2020-16:41:57] [V] [TRT] Tactic: 2 time 0.64 [11/26/2020-16:41:57] [V] [TRT] Tactic: 5 time 15.0712 [11/26/2020-16:41:57] [V] [TRT] Tactic: 6 time 0.273408 [11/26/2020-16:41:57] [V] [TRT] Fastest Tactic: 6 Time: 0.273408 [11/26/2020-16:41:57] [V] [TRT] --------------- Timing Runner: Conv_30 (CudaDepthwiseConvolution) [11/26/2020-16:41:57] [V] [TRT] CudaDepthwiseConvolution has no valid tactics for this config, skipping [11/26/2020-16:41:57] [V] [TRT] >>>>>>>>>>>>>>> Chose Runner Type: CudaConvolution Tactic: 6 [11/26/2020-16:41:57] [V] [TRT] [11/26/2020-16:41:57] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:41:57] [V] [TRT] Tactic: 0 time 0.009216 [11/26/2020-16:41:57] [V] [TRT] Fastest Tactic: 0 Time: 0.009216 [11/26/2020-16:41:57] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:41:57] [V] [TRT] Tactic: 0 time 0.007712 [11/26/2020-16:41:57] [V] [TRT] Fastest Tactic: 0 Time: 0.007712 [11/26/2020-16:41:57] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:41:57] [V] [TRT] Tactic: 0 time 0.014624 [11/26/2020-16:41:57] [V] [TRT] Fastest Tactic: 0 Time: 0.014624 [11/26/2020-16:41:57] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:41:57] [V] [TRT] Tactic: 0 time 0.011104 [11/26/2020-16:41:57] [V] [TRT] Fastest Tactic: 0 Time: 0.011104 [11/26/2020-16:41:57] [V] [TRT] *************** Autotuning format combination: Float(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2),(MUL_ADD 8 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 2048 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 4096)), Float(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2),(MUL_ADD 8 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 8192)) -> Float(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2),(MUL_ADD 8 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 8192)) *************** [11/26/2020-16:41:57] [V] [TRT] --------------- Timing Runner: Conv_31 + Add_32 + Relu_33 (LegacySASSConvolution) [11/26/2020-16:41:57] [V] [TRT] Tactic: 0 time 0.051072 [11/26/2020-16:41:57] [V] [TRT] Fastest Tactic: 0 Time: 0.051072 [11/26/2020-16:41:57] [V] [TRT] --------------- Timing Runner: Conv_31 + Add_32 + Relu_33 (CaskConvolution) [11/26/2020-16:41:57] [V] [TRT] Conv_31 + Add_32 + Relu_33 (scudnn) Set Tactic Name: maxwell_scudnn_128x32_relu_medium_nn_v1 [11/26/2020-16:41:57] [V] [TRT] Tactic: 1062367460111450758 time 0.042208 [11/26/2020-16:41:57] [V] [TRT] Conv_31 + Add_32 + Relu_33 (scudnn) Set Tactic Name: maxwell_scudnn_128x128_relu_medium_nn_v1 [11/26/2020-16:41:57] [V] [TRT] Tactic: 4501471010995462441 time 0.0376 [11/26/2020-16:41:57] [V] [TRT] Conv_31 + Add_32 + Relu_33 (scudnn) Set Tactic Name: maxwell_scudnn_128x64_relu_small_nn_v1 [11/26/2020-16:41:57] [V] [TRT] Tactic: 5137655947464784826 time 0.035744 [11/26/2020-16:41:57] [V] [TRT] Conv_31 + Add_32 + Relu_33 (scudnn) Set Tactic Name: maxwell_scudnn_128x128_relu_interior_nn_v1 [11/26/2020-16:41:57] [V] [TRT] Tactic: 5326823351883942011 time 0.035552 [11/26/2020-16:41:57] [V] [TRT] Conv_31 + Add_32 + Relu_33 (scudnn) Set Tactic Name: maxwell_scudnn_128x64_relu_medium_nn_v1 [11/26/2020-16:41:57] [V] [TRT] Tactic: 6645123197870846056 time 0.039936 [11/26/2020-16:41:57] [V] [TRT] Conv_31 + Add_32 + Relu_33 (scudnn) Set Tactic Name: maxwell_scudnn_128x32_relu_interior_nn_v1 [11/26/2020-16:41:57] [V] [TRT] Tactic: -6576203419454146580 time 0.036864 [11/26/2020-16:41:57] [V] [TRT] Conv_31 + Add_32 + Relu_33 (scudnn) Set Tactic Name: maxwell_scudnn_128x32_relu_small_nn_v1 [11/26/2020-16:41:57] [V] [TRT] Tactic: -3456450830548107839 time 0.038912 [11/26/2020-16:41:57] [V] [TRT] Conv_31 + Add_32 + Relu_33 (scudnn) Set Tactic Name: maxwell_scudnn_128x128_relu_small_nn_v1 [11/26/2020-16:41:57] [V] [TRT] Tactic: -410470605513481746 time 0.036864 [11/26/2020-16:41:57] [V] [TRT] Conv_31 + Add_32 + Relu_33 (scudnn) Set Tactic Name: maxwell_scudnn_128x64_relu_interior_nn_v1 [11/26/2020-16:41:57] [V] [TRT] Tactic: -37215280111360163 time 0.034816 [11/26/2020-16:41:57] [V] [TRT] Fastest Tactic: -37215280111360163 Time: 0.034816 [11/26/2020-16:41:57] [V] [TRT] --------------- Timing Runner: Conv_31 + Add_32 + Relu_33 (CudaConvolution) [11/26/2020-16:41:57] [V] [TRT] Tactic: 0 time 0.070656 [11/26/2020-16:41:57] [V] [TRT] Tactic: 1 time 0.06144 [11/26/2020-16:41:57] [V] [TRT] Tactic: 2 time 0.16384 [11/26/2020-16:41:57] [V] [TRT] Tactic: 5 time 0.407552 [11/26/2020-16:41:57] [V] [TRT] Fastest Tactic: 1 Time: 0.06144 [11/26/2020-16:41:57] [V] [TRT] >>>>>>>>>>>>>>> Chose Runner Type: CaskConvolution Tactic: -37215280111360163 [11/26/2020-16:41:57] [V] [TRT] Conv_31 + Add_32 + Relu_33 (scudnn) Set Tactic Name: maxwell_scudnn_128x64_relu_interior_nn_v1 [11/26/2020-16:41:57] [V] [TRT] [11/26/2020-16:41:57] [V] [TRT] *************** Autotuning format combination: Half(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2),(MUL_ADD 8 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 2048 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 4096)), Half(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2),(MUL_ADD 8 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 8192)) -> Half(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2),(MUL_ADD 8 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 8192)) *************** [11/26/2020-16:41:57] [V] [TRT] --------------- Timing Runner: Conv_31 + Add_32 + Relu_33 (CaskConvolution) [11/26/2020-16:41:57] [V] [TRT] CaskConvolution has no valid tactics for this config, skipping [11/26/2020-16:41:57] [V] [TRT] --------------- Timing Runner: Conv_31 + Add_32 + Relu_33 (CudaConvolution) [11/26/2020-16:41:57] [V] [TRT] Tactic: 0 time 0.094208 [11/26/2020-16:41:57] [V] [TRT] Tactic: 1 time 0.05632 [11/26/2020-16:41:57] [V] [TRT] Tactic: 2 time 0.18432 [11/26/2020-16:41:57] [V] [TRT] Tactic: 5 time 0.396288 [11/26/2020-16:41:57] [V] [TRT] Fastest Tactic: 1 Time: 0.05632 [11/26/2020-16:41:57] [V] [TRT] >>>>>>>>>>>>>>> Chose Runner Type: CudaConvolution Tactic: 1 [11/26/2020-16:41:57] [V] [TRT] [11/26/2020-16:41:57] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:41:57] [V] [TRT] Tactic: 0 time 0.0152 [11/26/2020-16:41:57] [V] [TRT] Fastest Tactic: 0 Time: 0.0152 [11/26/2020-16:41:57] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:41:57] [V] [TRT] Tactic: 0 time 0.011104 [11/26/2020-16:41:57] [V] [TRT] Fastest Tactic: 0 Time: 0.011104 [11/26/2020-16:41:57] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:41:57] [V] [TRT] Tactic: 0 time 0.015232 [11/26/2020-16:41:57] [V] [TRT] Fastest Tactic: 0 Time: 0.015232 [11/26/2020-16:41:57] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:41:57] [V] [TRT] Tactic: 0 time 0.0112 [11/26/2020-16:41:57] [V] [TRT] Fastest Tactic: 0 Time: 0.0112 [11/26/2020-16:41:57] [V] [TRT] *************** Autotuning format combination: Float(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2),(MUL_ADD 8 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 8192)) -> Float(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2),(MUL_ADD 8 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 8192)) *************** [11/26/2020-16:41:57] [V] [TRT] --------------- Timing Runner: Conv_34 + Relu_35 (LegacySASSConvolution) [11/26/2020-16:41:57] [V] [TRT] Tactic: 0 time 0.561152 [11/26/2020-16:41:57] [V] [TRT] Tactic: 1 time 0.229216 [11/26/2020-16:41:57] [V] [TRT] Fastest Tactic: 1 Time: 0.229216 [11/26/2020-16:41:57] [V] [TRT] --------------- Timing Runner: Conv_34 + Relu_35 (FusedConvActConvolution) [11/26/2020-16:41:57] [V] [TRT] Tactic: 7 time 0.566112 [11/26/2020-16:41:57] [V] [TRT] Tactic: 10 time 0.531328 [11/26/2020-16:41:57] [V] [TRT] Tactic: 14 time 0.351008 [11/26/2020-16:41:57] [V] [TRT] Tactic: 15 time 0.347872 [11/26/2020-16:41:57] [V] [TRT] Tactic: 25 time 0.351232 [11/26/2020-16:41:57] [V] [TRT] Tactic: 26 time 0.504672 [11/26/2020-16:41:57] [V] [TRT] Tactic: 29 time 0.4696 [11/26/2020-16:41:57] [V] [TRT] Tactic: 30 time 0.490496 [11/26/2020-16:41:57] [V] [TRT] Tactic: 33 time 0.615168 [11/26/2020-16:41:57] [V] [TRT] Tactic: 36 time 0.864256 [11/26/2020-16:41:57] [V] [TRT] Tactic: 39 time 0.37888 [11/26/2020-16:41:57] [V] [TRT] Tactic: 41 time 0.451552 [11/26/2020-16:41:57] [V] [TRT] Tactic: 42 time 0.620544 [11/26/2020-16:41:57] [V] [TRT] Tactic: 43 time 0.60928 [11/26/2020-16:41:58] [V] [TRT] Tactic: 45 time 0.318464 [11/26/2020-16:41:58] [V] [TRT] Tactic: 47 time 0.397312 [11/26/2020-16:41:58] [V] [TRT] Tactic: 52 time 0.38912 [11/26/2020-16:41:58] [V] [TRT] Tactic: 54 time 0.606208 [11/26/2020-16:41:58] [V] [TRT] Tactic: 56 time 0.439296 [11/26/2020-16:41:58] [V] [TRT] Tactic: 66 time 0.694272 [11/26/2020-16:41:58] [V] [TRT] Tactic: 76 time 0.381408 [11/26/2020-16:41:58] [V] [TRT] Tactic: 90 time 0.420864 [11/26/2020-16:41:58] [V] [TRT] Tactic: 93 time 0.387072 [11/26/2020-16:41:58] [V] [TRT] Tactic: 98 time 0.575488 [11/26/2020-16:41:58] [V] [TRT] Tactic: 104 time 0.534528 [11/26/2020-16:41:58] [V] [TRT] Tactic: 110 time 0.451584 [11/26/2020-16:41:58] [V] [TRT] Tactic: 119 time 0.623616 [11/26/2020-16:41:58] [V] [TRT] Tactic: 121 time 0.41728 [11/26/2020-16:41:58] [V] [TRT] Tactic: 130 time 0.56832 [11/26/2020-16:41:58] [V] [TRT] Tactic: 134 time 0.592896 [11/26/2020-16:41:58] [V] [TRT] Tactic: 136 time 0.542336 [11/26/2020-16:41:58] [V] [TRT] Tactic: 137 time 0.570592 [11/26/2020-16:41:58] [V] [TRT] Tactic: 139 time 0.607232 [11/26/2020-16:41:58] [V] [TRT] Tactic: 144 time 0.525952 [11/26/2020-16:41:58] [V] [TRT] Tactic: 149 time 0.382912 [11/26/2020-16:41:58] [V] [TRT] Tactic: 151 time 0.62976 [11/26/2020-16:41:58] [V] [TRT] Tactic: 152 time 0.48128 [11/26/2020-16:41:58] [V] [TRT] Tactic: 153 time 0.697344 [11/26/2020-16:41:58] [V] [TRT] Tactic: 156 time 0.470016 [11/26/2020-16:41:58] [V] [TRT] Tactic: 159 time 0.425984 [11/26/2020-16:41:58] [V] [TRT] Tactic: 162 time 0.49488 [11/26/2020-16:41:58] [V] [TRT] Tactic: 164 time 0.370688 [11/26/2020-16:41:58] [V] [TRT] Fastest Tactic: 45 Time: 0.318464 [11/26/2020-16:41:58] [V] [TRT] --------------- Timing Runner: Conv_34 + Relu_35 (CaskConvolution) [11/26/2020-16:41:58] [V] [TRT] Conv_34 + Relu_35 (scudnn) Set Tactic Name: maxwell_scudnn_128x32_relu_medium_nn_v1 [11/26/2020-16:41:58] [V] [TRT] Tactic: 1062367460111450758 time 0.566272 [11/26/2020-16:41:58] [V] [TRT] Conv_34 + Relu_35 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148n_nt_v1 [11/26/2020-16:41:58] [V] [TRT] Tactic: 3827454225649558724 time 0.195584 [11/26/2020-16:41:58] [V] [TRT] Conv_34 + Relu_35 (scudnn) Set Tactic Name: maxwell_scudnn_128x64_relu_large_nn_v1 [11/26/2020-16:41:58] [V] [TRT] Tactic: 4337000649858996379 time 0.516096 [11/26/2020-16:41:58] [V] [TRT] Conv_34 + Relu_35 (scudnn) Set Tactic Name: maxwell_scudnn_128x128_relu_medium_nn_v1 [11/26/2020-16:41:58] [V] [TRT] Tactic: 4501471010995462441 time 0.494592 [11/26/2020-16:41:58] [V] [TRT] Conv_34 + Relu_35 (scudnn) Set Tactic Name: maxwell_scudnn_128x64_relu_small_nn_v1 [11/26/2020-16:41:58] [V] [TRT] Tactic: 5137655947464784826 time 0.4608 [11/26/2020-16:41:58] [V] [TRT] Conv_34 + Relu_35 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148m_nt_v1 [11/26/2020-16:41:58] [V] [TRT] Tactic: 5921334924264294896 time 0.206848 [11/26/2020-16:41:58] [V] [TRT] Conv_34 + Relu_35 (scudnn) Set Tactic Name: maxwell_scudnn_128x64_relu_medium_nn_v1 [11/26/2020-16:41:58] [V] [TRT] Tactic: 6645123197870846056 time 0.503808 [11/26/2020-16:41:58] [V] [TRT] Conv_34 + Relu_35 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148t_nt_v1 [11/26/2020-16:41:58] [V] [TRT] Tactic: 7852627285308570038 time 0.216064 [11/26/2020-16:41:58] [V] [TRT] Conv_34 + Relu_35 (scudnn) Set Tactic Name: maxwell_scudnn_128x128_relu_large_nn_v1 [11/26/2020-16:41:58] [V] [TRT] Tactic: -9137461792520977713 time 0.50688 [11/26/2020-16:41:58] [V] [TRT] Conv_34 + Relu_35 (scudnn) Set Tactic Name: maxwell_scudnn_128x32_relu_large_nn_v1 [11/26/2020-16:41:58] [V] [TRT] Tactic: -6092040395344634144 time 0.610304 [11/26/2020-16:41:58] [V] [TRT] Conv_34 + Relu_35 (scudnn) Set Tactic Name: maxwell_scudnn_128x32_relu_small_nn_v1 [11/26/2020-16:41:58] [V] [TRT] Tactic: -3456450830548107839 time 0.489472 [11/26/2020-16:41:58] [V] [TRT] Conv_34 + Relu_35 (scudnn) Set Tactic Name: maxwell_scudnn_128x128_relu_small_nn_v1 [11/26/2020-16:41:58] [V] [TRT] Tactic: -410470605513481746 time 0.47712 [11/26/2020-16:41:58] [V] [TRT] Fastest Tactic: 3827454225649558724 Time: 0.195584 [11/26/2020-16:41:58] [V] [TRT] --------------- Timing Runner: Conv_34 + Relu_35 (CudaConvolution) [11/26/2020-16:41:58] [V] [TRT] Tactic: 0 time 0.71984 [11/26/2020-16:41:58] [V] [TRT] Tactic: 1 time 0.539648 [11/26/2020-16:41:58] [V] [TRT] Tactic: 2 time 0.64 [11/26/2020-16:41:58] [V] [TRT] Tactic: 5 time 15.8413 [11/26/2020-16:41:58] [V] [TRT] Tactic: 6 time 0.29568 [11/26/2020-16:41:58] [V] [TRT] Fastest Tactic: 6 Time: 0.29568 [11/26/2020-16:41:58] [V] [TRT] --------------- Timing Runner: Conv_34 + Relu_35 (CudaDepthwiseConvolution) [11/26/2020-16:41:58] [V] [TRT] CudaDepthwiseConvolution has no valid tactics for this config, skipping [11/26/2020-16:41:58] [V] [TRT] >>>>>>>>>>>>>>> Chose Runner Type: CaskConvolution Tactic: 3827454225649558724 [11/26/2020-16:41:58] [V] [TRT] Conv_34 + Relu_35 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148n_nt_v1 [11/26/2020-16:41:58] [V] [TRT] [11/26/2020-16:41:58] [V] [TRT] *************** Autotuning format combination: Half(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2),(MUL_ADD 8 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 8192)) -> Half(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2),(MUL_ADD 8 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 8192)) *************** [11/26/2020-16:41:58] [V] [TRT] --------------- Timing Runner: Conv_34 + Relu_35 (FusedConvActConvolution) [11/26/2020-16:41:58] [V] [TRT] FusedConvActConvolution has no valid tactics for this config, skipping [11/26/2020-16:41:58] [V] [TRT] --------------- Timing Runner: Conv_34 + Relu_35 (CaskConvolution) [11/26/2020-16:41:58] [V] [TRT] CaskConvolution has no valid tactics for this config, skipping [11/26/2020-16:41:58] [V] [TRT] --------------- Timing Runner: Conv_34 + Relu_35 (CudaConvolution) [11/26/2020-16:41:58] [V] [TRT] Tactic: 0 time 0.774144 [11/26/2020-16:41:58] [V] [TRT] Tactic: 1 time 0.556032 [11/26/2020-16:41:58] [V] [TRT] Tactic: 2 time 0.641024 [11/26/2020-16:41:58] [V] [TRT] Tactic: 5 time 15.4245 [11/26/2020-16:41:58] [V] [TRT] Tactic: 6 time 0.275456 [11/26/2020-16:41:58] [V] [TRT] Fastest Tactic: 6 Time: 0.275456 [11/26/2020-16:41:58] [V] [TRT] --------------- Timing Runner: Conv_34 + Relu_35 (CudaDepthwiseConvolution) [11/26/2020-16:41:58] [V] [TRT] CudaDepthwiseConvolution has no valid tactics for this config, skipping [11/26/2020-16:41:58] [V] [TRT] >>>>>>>>>>>>>>> Chose Runner Type: CudaConvolution Tactic: 6 [11/26/2020-16:41:58] [V] [TRT] [11/26/2020-16:41:58] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:41:58] [V] [TRT] Tactic: 0 time 0.014336 [11/26/2020-16:41:58] [V] [TRT] Fastest Tactic: 0 Time: 0.014336 [11/26/2020-16:41:58] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:41:58] [V] [TRT] Tactic: 0 time 0.012224 [11/26/2020-16:41:58] [V] [TRT] Fastest Tactic: 0 Time: 0.012224 [11/26/2020-16:41:59] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:41:59] [V] [TRT] Tactic: 0 time 0.014336 [11/26/2020-16:41:59] [V] [TRT] Fastest Tactic: 0 Time: 0.014336 [11/26/2020-16:41:59] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:41:59] [V] [TRT] Tactic: 0 time 0.011968 [11/26/2020-16:41:59] [V] [TRT] Fastest Tactic: 0 Time: 0.011968 [11/26/2020-16:41:59] [V] [TRT] *************** Autotuning format combination: Float(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2),(MUL_ADD 8 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 8192)), Float(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2),(MUL_ADD 8 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 8192)) -> Float(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2),(MUL_ADD 8 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 8192)) *************** [11/26/2020-16:41:59] [V] [TRT] --------------- Timing Runner: Conv_36 + Add_37 + Relu_38 (LegacySASSConvolution) [11/26/2020-16:41:59] [V] [TRT] Tactic: 0 time 0.556032 [11/26/2020-16:41:59] [V] [TRT] Tactic: 1 time 0.234432 [11/26/2020-16:41:59] [V] [TRT] Fastest Tactic: 1 Time: 0.234432 [11/26/2020-16:41:59] [V] [TRT] --------------- Timing Runner: Conv_36 + Add_37 + Relu_38 (CaskConvolution) [11/26/2020-16:41:59] [V] [TRT] Conv_36 + Add_37 + Relu_38 (scudnn) Set Tactic Name: maxwell_scudnn_128x32_relu_medium_nn_v1 [11/26/2020-16:41:59] [V] [TRT] Tactic: 1062367460111450758 time 0.57856 [11/26/2020-16:41:59] [V] [TRT] Conv_36 + Add_37 + Relu_38 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148n_nt_v1 [11/26/2020-16:41:59] [V] [TRT] Tactic: 3827454225649558724 time 0.209856 [11/26/2020-16:41:59] [V] [TRT] Conv_36 + Add_37 + Relu_38 (scudnn) Set Tactic Name: maxwell_scudnn_128x64_relu_large_nn_v1 [11/26/2020-16:41:59] [V] [TRT] Tactic: 4337000649858996379 time 0.516096 [11/26/2020-16:41:59] [V] [TRT] Conv_36 + Add_37 + Relu_38 (scudnn) Set Tactic Name: maxwell_scudnn_128x128_relu_medium_nn_v1 [11/26/2020-16:41:59] [V] [TRT] Tactic: 4501471010995462441 time 0.489472 [11/26/2020-16:41:59] [V] [TRT] Conv_36 + Add_37 + Relu_38 (scudnn) Set Tactic Name: maxwell_scudnn_128x64_relu_small_nn_v1 [11/26/2020-16:41:59] [V] [TRT] Tactic: 5137655947464784826 time 0.46128 [11/26/2020-16:41:59] [V] [TRT] Conv_36 + Add_37 + Relu_38 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148m_nt_v1 [11/26/2020-16:41:59] [V] [TRT] Tactic: 5921334924264294896 time 0.208544 [11/26/2020-16:41:59] [V] [TRT] Conv_36 + Add_37 + Relu_38 (scudnn) Set Tactic Name: maxwell_scudnn_128x64_relu_medium_nn_v1 [11/26/2020-16:41:59] [V] [TRT] Tactic: 6645123197870846056 time 0.49664 [11/26/2020-16:41:59] [V] [TRT] Conv_36 + Add_37 + Relu_38 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148t_nt_v1 [11/26/2020-16:41:59] [V] [TRT] Tactic: 7852627285308570038 time 0.221184 [11/26/2020-16:41:59] [V] [TRT] Conv_36 + Add_37 + Relu_38 (scudnn) Set Tactic Name: maxwell_scudnn_128x128_relu_large_nn_v1 [11/26/2020-16:41:59] [V] [TRT] Tactic: -9137461792520977713 time 0.505856 [11/26/2020-16:41:59] [V] [TRT] Conv_36 + Add_37 + Relu_38 (scudnn) Set Tactic Name: maxwell_scudnn_128x32_relu_large_nn_v1 [11/26/2020-16:41:59] [V] [TRT] Tactic: -6092040395344634144 time 0.60416 [11/26/2020-16:41:59] [V] [TRT] Conv_36 + Add_37 + Relu_38 (scudnn) Set Tactic Name: maxwell_scudnn_128x32_relu_small_nn_v1 [11/26/2020-16:41:59] [V] [TRT] Tactic: -3456450830548107839 time 0.493568 [11/26/2020-16:41:59] [V] [TRT] Conv_36 + Add_37 + Relu_38 (scudnn) Set Tactic Name: maxwell_scudnn_128x128_relu_small_nn_v1 [11/26/2020-16:41:59] [V] [TRT] Tactic: -410470605513481746 time 0.479232 [11/26/2020-16:41:59] [V] [TRT] Fastest Tactic: 5921334924264294896 Time: 0.208544 [11/26/2020-16:41:59] [V] [TRT] --------------- Timing Runner: Conv_36 + Add_37 + Relu_38 (CudaConvolution) [11/26/2020-16:41:59] [V] [TRT] Tactic: 0 time 0.728064 [11/26/2020-16:41:59] [V] [TRT] Tactic: 1 time 0.548864 [11/26/2020-16:41:59] [V] [TRT] Tactic: 2 time 0.648 [11/26/2020-16:41:59] [V] [TRT] Tactic: 5 time 13.8844 [11/26/2020-16:41:59] [V] [TRT] Tactic: 6 time 0.300032 [11/26/2020-16:41:59] [V] [TRT] Fastest Tactic: 6 Time: 0.300032 [11/26/2020-16:41:59] [V] [TRT] >>>>>>>>>>>>>>> Chose Runner Type: CaskConvolution Tactic: 5921334924264294896 [11/26/2020-16:41:59] [V] [TRT] Conv_36 + Add_37 + Relu_38 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148m_nt_v1 [11/26/2020-16:41:59] [V] [TRT] [11/26/2020-16:41:59] [V] [TRT] *************** Autotuning format combination: Half(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2),(MUL_ADD 8 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 8192)), Half(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2),(MUL_ADD 8 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 8192)) -> Half(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2),(MUL_ADD 8 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 8192)) *************** [11/26/2020-16:41:59] [V] [TRT] --------------- Timing Runner: Conv_36 + Add_37 + Relu_38 (CaskConvolution) [11/26/2020-16:41:59] [V] [TRT] CaskConvolution has no valid tactics for this config, skipping [11/26/2020-16:41:59] [V] [TRT] --------------- Timing Runner: Conv_36 + Add_37 + Relu_38 (CudaConvolution) [11/26/2020-16:41:59] [V] [TRT] Tactic: 0 time 0.784384 [11/26/2020-16:41:59] [V] [TRT] Tactic: 1 time 0.562176 [11/26/2020-16:41:59] [V] [TRT] Tactic: 2 time 0.64512 [11/26/2020-16:41:59] [V] [TRT] Tactic: 5 time 15.0313 [11/26/2020-16:41:59] [V] [TRT] Tactic: 6 time 0.28128 [11/26/2020-16:41:59] [V] [TRT] Fastest Tactic: 6 Time: 0.28128 [11/26/2020-16:41:59] [V] [TRT] >>>>>>>>>>>>>>> Chose Runner Type: CudaConvolution Tactic: 6 [11/26/2020-16:41:59] [V] [TRT] [11/26/2020-16:41:59] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:41:59] [V] [TRT] Tactic: 0 time 0.014336 [11/26/2020-16:41:59] [V] [TRT] Fastest Tactic: 0 Time: 0.014336 [11/26/2020-16:41:59] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:41:59] [V] [TRT] Tactic: 0 time 0.011136 [11/26/2020-16:41:59] [V] [TRT] Fastest Tactic: 0 Time: 0.011136 [11/26/2020-16:41:59] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:41:59] [V] [TRT] Tactic: 0 time 0.015072 [11/26/2020-16:41:59] [V] [TRT] Fastest Tactic: 0 Time: 0.015072 [11/26/2020-16:41:59] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:41:59] [V] [TRT] Tactic: 0 time 0.011264 [11/26/2020-16:41:59] [V] [TRT] Fastest Tactic: 0 Time: 0.011264 [11/26/2020-16:41:59] [V] [TRT] *************** Autotuning format combination: Float(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2),(MUL_ADD 8 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 8192)) -> Float(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2),(MUL_ADD 8 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 8192)) *************** [11/26/2020-16:41:59] [V] [TRT] --------------- Timing Runner: Conv_39 + Relu_40 (LegacySASSConvolution) [11/26/2020-16:41:59] [V] [TRT] Tactic: 0 time 0.560128 [11/26/2020-16:41:59] [V] [TRT] Tactic: 1 time 0.230304 [11/26/2020-16:41:59] [V] [TRT] Fastest Tactic: 1 Time: 0.230304 [11/26/2020-16:41:59] [V] [TRT] --------------- Timing Runner: Conv_39 + Relu_40 (FusedConvActConvolution) [11/26/2020-16:41:59] [V] [TRT] Tactic: 7 time 0.56624 [11/26/2020-16:41:59] [V] [TRT] Tactic: 10 time 0.525056 [11/26/2020-16:41:59] [V] [TRT] Tactic: 14 time 0.310112 [11/26/2020-16:41:59] [V] [TRT] Tactic: 15 time 0.344064 [11/26/2020-16:41:59] [V] [TRT] Tactic: 25 time 0.35328 [11/26/2020-16:42:00] [V] [TRT] Tactic: 26 time 0.503808 [11/26/2020-16:42:00] [V] [TRT] Tactic: 29 time 0.469568 [11/26/2020-16:42:00] [V] [TRT] Tactic: 30 time 0.490496 [11/26/2020-16:42:00] [V] [TRT] Tactic: 33 time 0.610048 [11/26/2020-16:42:00] [V] [TRT] Tactic: 36 time 0.863232 [11/26/2020-16:42:00] [V] [TRT] Tactic: 39 time 0.376384 [11/26/2020-16:42:00] [V] [TRT] Tactic: 41 time 0.451584 [11/26/2020-16:42:00] [V] [TRT] Tactic: 42 time 0.622432 [11/26/2020-16:42:00] [V] [TRT] Tactic: 43 time 0.609856 [11/26/2020-16:42:00] [V] [TRT] Tactic: 45 time 0.318016 [11/26/2020-16:42:00] [V] [TRT] Tactic: 47 time 0.396288 [11/26/2020-16:42:00] [V] [TRT] Tactic: 52 time 0.401408 [11/26/2020-16:42:00] [V] [TRT] Tactic: 54 time 0.608256 [11/26/2020-16:42:00] [V] [TRT] Tactic: 56 time 0.443392 [11/26/2020-16:42:00] [V] [TRT] Tactic: 66 time 0.6912 [11/26/2020-16:42:00] [V] [TRT] Tactic: 76 time 0.379904 [11/26/2020-16:42:00] [V] [TRT] Tactic: 90 time 0.42144 [11/26/2020-16:42:00] [V] [TRT] Tactic: 93 time 0.385024 [11/26/2020-16:42:00] [V] [TRT] Tactic: 98 time 0.567296 [11/26/2020-16:42:00] [V] [TRT] Tactic: 104 time 0.534528 [11/26/2020-16:42:00] [V] [TRT] Tactic: 110 time 0.451584 [11/26/2020-16:42:00] [V] [TRT] Tactic: 119 time 0.618496 [11/26/2020-16:42:00] [V] [TRT] Tactic: 121 time 0.42496 [11/26/2020-16:42:00] [V] [TRT] Tactic: 130 time 0.56992 [11/26/2020-16:42:00] [V] [TRT] Tactic: 134 time 0.590848 [11/26/2020-16:42:00] [V] [TRT] Tactic: 136 time 0.54272 [11/26/2020-16:42:00] [V] [TRT] Tactic: 137 time 0.571392 [11/26/2020-16:42:00] [V] [TRT] Tactic: 139 time 0.608256 [11/26/2020-16:42:00] [V] [TRT] Tactic: 144 time 0.52736 [11/26/2020-16:42:00] [V] [TRT] Tactic: 149 time 0.388096 [11/26/2020-16:42:00] [V] [TRT] Tactic: 151 time 0.627712 [11/26/2020-16:42:00] [V] [TRT] Tactic: 152 time 0.477088 [11/26/2020-16:42:00] [V] [TRT] Tactic: 153 time 0.695296 [11/26/2020-16:42:00] [V] [TRT] Tactic: 156 time 0.467968 [11/26/2020-16:42:00] [V] [TRT] Tactic: 159 time 0.429056 [11/26/2020-16:42:00] [V] [TRT] Tactic: 162 time 0.49664 [11/26/2020-16:42:00] [V] [TRT] Tactic: 164 time 0.364544 [11/26/2020-16:42:00] [V] [TRT] Fastest Tactic: 14 Time: 0.310112 [11/26/2020-16:42:00] [V] [TRT] --------------- Timing Runner: Conv_39 + Relu_40 (CaskConvolution) [11/26/2020-16:42:00] [V] [TRT] Conv_39 + Relu_40 (scudnn) Set Tactic Name: maxwell_scudnn_128x32_relu_medium_nn_v1 [11/26/2020-16:42:00] [V] [TRT] Tactic: 1062367460111450758 time 0.558048 [11/26/2020-16:42:00] [V] [TRT] Conv_39 + Relu_40 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148n_nt_v1 [11/26/2020-16:42:00] [V] [TRT] Tactic: 3827454225649558724 time 0.19456 [11/26/2020-16:42:00] [V] [TRT] Conv_39 + Relu_40 (scudnn) Set Tactic Name: maxwell_scudnn_128x64_relu_large_nn_v1 [11/26/2020-16:42:00] [V] [TRT] Tactic: 4337000649858996379 time 0.523264 [11/26/2020-16:42:00] [V] [TRT] Conv_39 + Relu_40 (scudnn) Set Tactic Name: maxwell_scudnn_128x128_relu_medium_nn_v1 [11/26/2020-16:42:00] [V] [TRT] Tactic: 4501471010995462441 time 0.487328 [11/26/2020-16:42:00] [V] [TRT] Conv_39 + Relu_40 (scudnn) Set Tactic Name: maxwell_scudnn_128x64_relu_small_nn_v1 [11/26/2020-16:42:00] [V] [TRT] Tactic: 5137655947464784826 time 0.457728 [11/26/2020-16:42:00] [V] [TRT] Conv_39 + Relu_40 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148m_nt_v1 [11/26/2020-16:42:00] [V] [TRT] Tactic: 5921334924264294896 time 0.203776 [11/26/2020-16:42:00] [V] [TRT] Conv_39 + Relu_40 (scudnn) Set Tactic Name: maxwell_scudnn_128x64_relu_medium_nn_v1 [11/26/2020-16:42:00] [V] [TRT] Tactic: 6645123197870846056 time 0.495616 [11/26/2020-16:42:00] [V] [TRT] Conv_39 + Relu_40 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148t_nt_v1 [11/26/2020-16:42:00] [V] [TRT] Tactic: 7852627285308570038 time 0.213568 [11/26/2020-16:42:00] [V] [TRT] Conv_39 + Relu_40 (scudnn) Set Tactic Name: maxwell_scudnn_128x128_relu_large_nn_v1 [11/26/2020-16:42:00] [V] [TRT] Tactic: -9137461792520977713 time 0.514048 [11/26/2020-16:42:00] [V] [TRT] Conv_39 + Relu_40 (scudnn) Set Tactic Name: maxwell_scudnn_128x32_relu_large_nn_v1 [11/26/2020-16:42:00] [V] [TRT] Tactic: -6092040395344634144 time 0.600064 [11/26/2020-16:42:00] [V] [TRT] Conv_39 + Relu_40 (scudnn) Set Tactic Name: maxwell_scudnn_128x32_relu_small_nn_v1 [11/26/2020-16:42:00] [V] [TRT] Tactic: -3456450830548107839 time 0.486976 [11/26/2020-16:42:00] [V] [TRT] Conv_39 + Relu_40 (scudnn) Set Tactic Name: maxwell_scudnn_128x128_relu_small_nn_v1 [11/26/2020-16:42:00] [V] [TRT] Tactic: -410470605513481746 time 0.47616 [11/26/2020-16:42:00] [V] [TRT] Fastest Tactic: 3827454225649558724 Time: 0.19456 [11/26/2020-16:42:00] [V] [TRT] --------------- Timing Runner: Conv_39 + Relu_40 (CudaConvolution) [11/26/2020-16:42:00] [V] [TRT] Tactic: 0 time 0.717824 [11/26/2020-16:42:00] [V] [TRT] Tactic: 1 time 0.544768 [11/26/2020-16:42:00] [V] [TRT] Tactic: 2 time 0.64 [11/26/2020-16:42:00] [V] [TRT] Tactic: 5 time 15.3713 [11/26/2020-16:42:00] [V] [TRT] Tactic: 6 time 0.290816 [11/26/2020-16:42:00] [V] [TRT] Fastest Tactic: 6 Time: 0.290816 [11/26/2020-16:42:00] [V] [TRT] --------------- Timing Runner: Conv_39 + Relu_40 (CudaDepthwiseConvolution) [11/26/2020-16:42:00] [V] [TRT] CudaDepthwiseConvolution has no valid tactics for this config, skipping [11/26/2020-16:42:00] [V] [TRT] >>>>>>>>>>>>>>> Chose Runner Type: CaskConvolution Tactic: 3827454225649558724 [11/26/2020-16:42:00] [V] [TRT] Conv_39 + Relu_40 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148n_nt_v1 [11/26/2020-16:42:00] [V] [TRT] [11/26/2020-16:42:01] [V] [TRT] *************** Autotuning format combination: Half(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2),(MUL_ADD 8 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 8192)) -> Half(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2),(MUL_ADD 8 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 8192)) *************** [11/26/2020-16:42:01] [V] [TRT] --------------- Timing Runner: Conv_39 + Relu_40 (FusedConvActConvolution) [11/26/2020-16:42:01] [V] [TRT] FusedConvActConvolution has no valid tactics for this config, skipping [11/26/2020-16:42:01] [V] [TRT] --------------- Timing Runner: Conv_39 + Relu_40 (CaskConvolution) [11/26/2020-16:42:01] [V] [TRT] CaskConvolution has no valid tactics for this config, skipping [11/26/2020-16:42:01] [V] [TRT] --------------- Timing Runner: Conv_39 + Relu_40 (CudaConvolution) [11/26/2020-16:42:01] [V] [TRT] Tactic: 0 time 0.775168 [11/26/2020-16:42:01] [V] [TRT] Tactic: 1 time 0.559104 [11/26/2020-16:42:01] [V] [TRT] Tactic: 2 time 0.642048 [11/26/2020-16:42:01] [V] [TRT] Tactic: 5 time 15.231 [11/26/2020-16:42:01] [V] [TRT] Tactic: 6 time 0.277472 [11/26/2020-16:42:01] [V] [TRT] Fastest Tactic: 6 Time: 0.277472 [11/26/2020-16:42:01] [V] [TRT] --------------- Timing Runner: Conv_39 + Relu_40 (CudaDepthwiseConvolution) [11/26/2020-16:42:01] [V] [TRT] CudaDepthwiseConvolution has no valid tactics for this config, skipping [11/26/2020-16:42:01] [V] [TRT] >>>>>>>>>>>>>>> Chose Runner Type: CudaConvolution Tactic: 6 [11/26/2020-16:42:01] [V] [TRT] [11/26/2020-16:42:01] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:42:01] [V] [TRT] Tactic: 0 time 0.014336 [11/26/2020-16:42:01] [V] [TRT] Fastest Tactic: 0 Time: 0.014336 [11/26/2020-16:42:01] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:42:01] [V] [TRT] Tactic: 0 time 0.011072 [11/26/2020-16:42:01] [V] [TRT] Fastest Tactic: 0 Time: 0.011072 [11/26/2020-16:42:01] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:42:01] [V] [TRT] Tactic: 0 time 0.014336 [11/26/2020-16:42:01] [V] [TRT] Fastest Tactic: 0 Time: 0.014336 [11/26/2020-16:42:01] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:42:01] [V] [TRT] Tactic: 0 time 0.012032 [11/26/2020-16:42:01] [V] [TRT] Fastest Tactic: 0 Time: 0.012032 [11/26/2020-16:42:01] [V] [TRT] *************** Autotuning format combination: Float(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2),(MUL_ADD 8 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 8192)), Float(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2),(MUL_ADD 8 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 8192)) -> Float(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2),(MUL_ADD 8 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 8192)) *************** [11/26/2020-16:42:01] [V] [TRT] --------------- Timing Runner: Conv_41 + Add_42 + Relu_43 (LegacySASSConvolution) [11/26/2020-16:42:01] [V] [TRT] Tactic: 0 time 0.563584 [11/26/2020-16:42:01] [V] [TRT] Tactic: 1 time 0.233376 [11/26/2020-16:42:01] [V] [TRT] Fastest Tactic: 1 Time: 0.233376 [11/26/2020-16:42:01] [V] [TRT] --------------- Timing Runner: Conv_41 + Add_42 + Relu_43 (CaskConvolution) [11/26/2020-16:42:01] [V] [TRT] Conv_41 + Add_42 + Relu_43 (scudnn) Set Tactic Name: maxwell_scudnn_128x32_relu_medium_nn_v1 [11/26/2020-16:42:01] [V] [TRT] Tactic: 1062367460111450758 time 0.570336 [11/26/2020-16:42:01] [V] [TRT] Conv_41 + Add_42 + Relu_43 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148n_nt_v1 [11/26/2020-16:42:01] [V] [TRT] Tactic: 3827454225649558724 time 0.208896 [11/26/2020-16:42:01] [V] [TRT] Conv_41 + Add_42 + Relu_43 (scudnn) Set Tactic Name: maxwell_scudnn_128x64_relu_large_nn_v1 [11/26/2020-16:42:01] [V] [TRT] Tactic: 4337000649858996379 time 0.52224 [11/26/2020-16:42:01] [V] [TRT] Conv_41 + Add_42 + Relu_43 (scudnn) Set Tactic Name: maxwell_scudnn_128x128_relu_medium_nn_v1 [11/26/2020-16:42:01] [V] [TRT] Tactic: 4501471010995462441 time 0.499712 [11/26/2020-16:42:01] [V] [TRT] Conv_41 + Add_42 + Relu_43 (scudnn) Set Tactic Name: maxwell_scudnn_128x64_relu_small_nn_v1 [11/26/2020-16:42:01] [V] [TRT] Tactic: 5137655947464784826 time 0.462848 [11/26/2020-16:42:01] [V] [TRT] Conv_41 + Add_42 + Relu_43 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148m_nt_v1 [11/26/2020-16:42:01] [V] [TRT] Tactic: 5921334924264294896 time 0.208448 [11/26/2020-16:42:01] [V] [TRT] Conv_41 + Add_42 + Relu_43 (scudnn) Set Tactic Name: maxwell_scudnn_128x64_relu_medium_nn_v1 [11/26/2020-16:42:01] [V] [TRT] Tactic: 6645123197870846056 time 0.505856 [11/26/2020-16:42:01] [V] [TRT] Conv_41 + Add_42 + Relu_43 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148t_nt_v1 [11/26/2020-16:42:01] [V] [TRT] Tactic: 7852627285308570038 time 0.21504 [11/26/2020-16:42:01] [V] [TRT] Conv_41 + Add_42 + Relu_43 (scudnn) Set Tactic Name: maxwell_scudnn_128x128_relu_large_nn_v1 [11/26/2020-16:42:01] [V] [TRT] Tactic: -9137461792520977713 time 0.507904 [11/26/2020-16:42:01] [V] [TRT] Conv_41 + Add_42 + Relu_43 (scudnn) Set Tactic Name: maxwell_scudnn_128x32_relu_large_nn_v1 [11/26/2020-16:42:01] [V] [TRT] Tactic: -6092040395344634144 time 0.612352 [11/26/2020-16:42:01] [V] [TRT] Conv_41 + Add_42 + Relu_43 (scudnn) Set Tactic Name: maxwell_scudnn_128x32_relu_small_nn_v1 [11/26/2020-16:42:01] [V] [TRT] Tactic: -3456450830548107839 time 0.487424 [11/26/2020-16:42:01] [V] [TRT] Conv_41 + Add_42 + Relu_43 (scudnn) Set Tactic Name: maxwell_scudnn_128x128_relu_small_nn_v1 [11/26/2020-16:42:01] [V] [TRT] Tactic: -410470605513481746 time 0.476928 [11/26/2020-16:42:01] [V] [TRT] Fastest Tactic: 5921334924264294896 Time: 0.208448 [11/26/2020-16:42:01] [V] [TRT] --------------- Timing Runner: Conv_41 + Add_42 + Relu_43 (CudaConvolution) [11/26/2020-16:42:01] [V] [TRT] Tactic: 0 time 0.724992 [11/26/2020-16:42:01] [V] [TRT] Tactic: 1 time 0.55296 [11/26/2020-16:42:01] [V] [TRT] Tactic: 2 time 0.649216 [11/26/2020-16:42:01] [V] [TRT] Tactic: 5 time 14.763 [11/26/2020-16:42:01] [V] [TRT] Tactic: 6 time 0.303104 [11/26/2020-16:42:01] [V] [TRT] Fastest Tactic: 6 Time: 0.303104 [11/26/2020-16:42:01] [V] [TRT] >>>>>>>>>>>>>>> Chose Runner Type: CaskConvolution Tactic: 5921334924264294896 [11/26/2020-16:42:01] [V] [TRT] Conv_41 + Add_42 + Relu_43 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148m_nt_v1 [11/26/2020-16:42:01] [V] [TRT] [11/26/2020-16:42:01] [V] [TRT] *************** Autotuning format combination: Half(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2),(MUL_ADD 8 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 8192)), Half(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2),(MUL_ADD 8 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 8192)) -> Half(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2),(MUL_ADD 8 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 8192)) *************** [11/26/2020-16:42:01] [V] [TRT] --------------- Timing Runner: Conv_41 + Add_42 + Relu_43 (CaskConvolution) [11/26/2020-16:42:01] [V] [TRT] CaskConvolution has no valid tactics for this config, skipping [11/26/2020-16:42:01] [V] [TRT] --------------- Timing Runner: Conv_41 + Add_42 + Relu_43 (CudaConvolution) [11/26/2020-16:42:01] [V] [TRT] Tactic: 0 time 0.782336 [11/26/2020-16:42:01] [V] [TRT] Tactic: 1 time 0.566272 [11/26/2020-16:42:01] [V] [TRT] Tactic: 2 time 0.646272 [11/26/2020-16:42:01] [V] [TRT] Tactic: 5 time 14.5469 [11/26/2020-16:42:01] [V] [TRT] Tactic: 6 time 0.2816 [11/26/2020-16:42:01] [V] [TRT] Fastest Tactic: 6 Time: 0.2816 [11/26/2020-16:42:01] [V] [TRT] >>>>>>>>>>>>>>> Chose Runner Type: CudaConvolution Tactic: 6 [11/26/2020-16:42:01] [V] [TRT] [11/26/2020-16:42:01] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:42:01] [V] [TRT] Tactic: 0 time 0.014336 [11/26/2020-16:42:01] [V] [TRT] Fastest Tactic: 0 Time: 0.014336 [11/26/2020-16:42:01] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:42:01] [V] [TRT] Tactic: 0 time 0.011232 [11/26/2020-16:42:01] [V] [TRT] Fastest Tactic: 0 Time: 0.011232 [11/26/2020-16:42:01] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:42:01] [V] [TRT] Tactic: 0 time 0.014336 [11/26/2020-16:42:01] [V] [TRT] Fastest Tactic: 0 Time: 0.014336 [11/26/2020-16:42:01] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:42:01] [V] [TRT] Tactic: 0 time 0.011264 [11/26/2020-16:42:01] [V] [TRT] Fastest Tactic: 0 Time: 0.011264 [11/26/2020-16:42:01] [V] [TRT] *************** Autotuning format combination: Float(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2),(MUL_ADD 8 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 8192)) -> Float(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2),(MUL_ADD 8 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 8192)) *************** [11/26/2020-16:42:01] [V] [TRT] --------------- Timing Runner: Conv_44 + Relu_45 (LegacySASSConvolution) [11/26/2020-16:42:02] [V] [TRT] Tactic: 0 time 0.552576 [11/26/2020-16:42:02] [V] [TRT] Tactic: 1 time 0.231776 [11/26/2020-16:42:02] [V] [TRT] Fastest Tactic: 1 Time: 0.231776 [11/26/2020-16:42:02] [V] [TRT] --------------- Timing Runner: Conv_44 + Relu_45 (FusedConvActConvolution) [11/26/2020-16:42:02] [V] [TRT] Tactic: 7 time 0.567296 [11/26/2020-16:42:02] [V] [TRT] Tactic: 10 time 0.534464 [11/26/2020-16:42:02] [V] [TRT] Tactic: 14 time 0.309248 [11/26/2020-16:42:02] [V] [TRT] Tactic: 15 time 0.344064 [11/26/2020-16:42:02] [V] [TRT] Tactic: 25 time 0.34304 [11/26/2020-16:42:02] [V] [TRT] Tactic: 26 time 0.502784 [11/26/2020-16:42:02] [V] [TRT] Tactic: 29 time 0.467968 [11/26/2020-16:42:02] [V] [TRT] Tactic: 30 time 0.492064 [11/26/2020-16:42:02] [V] [TRT] Tactic: 33 time 0.625664 [11/26/2020-16:42:02] [V] [TRT] Tactic: 36 time 0.873472 [11/26/2020-16:42:02] [V] [TRT] Tactic: 39 time 0.371456 [11/26/2020-16:42:02] [V] [TRT] Tactic: 41 time 0.45056 [11/26/2020-16:42:02] [V] [TRT] Tactic: 42 time 0.61952 [11/26/2020-16:42:02] [V] [TRT] Tactic: 43 time 0.608256 [11/26/2020-16:42:02] [V] [TRT] Tactic: 45 time 0.320256 [11/26/2020-16:42:02] [V] [TRT] Tactic: 47 time 0.39808 [11/26/2020-16:42:02] [V] [TRT] Tactic: 52 time 0.393216 [11/26/2020-16:42:02] [V] [TRT] Tactic: 54 time 0.607232 [11/26/2020-16:42:02] [V] [TRT] Tactic: 56 time 0.443392 [11/26/2020-16:42:02] [V] [TRT] Tactic: 66 time 0.692224 [11/26/2020-16:42:02] [V] [TRT] Tactic: 76 time 0.370688 [11/26/2020-16:42:02] [V] [TRT] Tactic: 90 time 0.422912 [11/26/2020-16:42:02] [V] [TRT] Tactic: 93 time 0.384 [11/26/2020-16:42:02] [V] [TRT] Tactic: 98 time 0.56832 [11/26/2020-16:42:02] [V] [TRT] Tactic: 104 time 0.539648 [11/26/2020-16:42:02] [V] [TRT] Tactic: 110 time 0.461824 [11/26/2020-16:42:02] [V] [TRT] Tactic: 119 time 0.620544 [11/26/2020-16:42:02] [V] [TRT] Tactic: 121 time 0.410624 [11/26/2020-16:42:02] [V] [TRT] Tactic: 130 time 0.572416 [11/26/2020-16:42:02] [V] [TRT] Tactic: 134 time 0.5888 [11/26/2020-16:42:02] [V] [TRT] Tactic: 136 time 0.540672 [11/26/2020-16:42:02] [V] [TRT] Tactic: 137 time 0.570368 [11/26/2020-16:42:02] [V] [TRT] Tactic: 139 time 0.607168 [11/26/2020-16:42:02] [V] [TRT] Tactic: 144 time 0.53248 [11/26/2020-16:42:02] [V] [TRT] Tactic: 149 time 0.38352 [11/26/2020-16:42:02] [V] [TRT] Tactic: 151 time 0.627616 [11/26/2020-16:42:02] [V] [TRT] Tactic: 152 time 0.475136 [11/26/2020-16:42:02] [V] [TRT] Tactic: 153 time 0.692224 [11/26/2020-16:42:02] [V] [TRT] Tactic: 156 time 0.466304 [11/26/2020-16:42:02] [V] [TRT] Tactic: 159 time 0.427008 [11/26/2020-16:42:03] [V] [TRT] Tactic: 162 time 0.51712 [11/26/2020-16:42:03] [V] [TRT] Tactic: 164 time 0.370688 [11/26/2020-16:42:03] [V] [TRT] Fastest Tactic: 14 Time: 0.309248 [11/26/2020-16:42:03] [V] [TRT] --------------- Timing Runner: Conv_44 + Relu_45 (CaskConvolution) [11/26/2020-16:42:03] [V] [TRT] Conv_44 + Relu_45 (scudnn) Set Tactic Name: maxwell_scudnn_128x32_relu_medium_nn_v1 [11/26/2020-16:42:03] [V] [TRT] Tactic: 1062367460111450758 time 0.567296 [11/26/2020-16:42:03] [V] [TRT] Conv_44 + Relu_45 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148n_nt_v1 [11/26/2020-16:42:03] [V] [TRT] Tactic: 3827454225649558724 time 0.196608 [11/26/2020-16:42:03] [V] [TRT] Conv_44 + Relu_45 (scudnn) Set Tactic Name: maxwell_scudnn_128x64_relu_large_nn_v1 [11/26/2020-16:42:03] [V] [TRT] Tactic: 4337000649858996379 time 0.514048 [11/26/2020-16:42:03] [V] [TRT] Conv_44 + Relu_45 (scudnn) Set Tactic Name: maxwell_scudnn_128x128_relu_medium_nn_v1 [11/26/2020-16:42:03] [V] [TRT] Tactic: 4501471010995462441 time 0.484352 [11/26/2020-16:42:03] [V] [TRT] Conv_44 + Relu_45 (scudnn) Set Tactic Name: maxwell_scudnn_128x64_relu_small_nn_v1 [11/26/2020-16:42:03] [V] [TRT] Tactic: 5137655947464784826 time 0.458752 [11/26/2020-16:42:03] [V] [TRT] Conv_44 + Relu_45 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148m_nt_v1 [11/26/2020-16:42:03] [V] [TRT] Tactic: 5921334924264294896 time 0.207424 [11/26/2020-16:42:03] [V] [TRT] Conv_44 + Relu_45 (scudnn) Set Tactic Name: maxwell_scudnn_128x64_relu_medium_nn_v1 [11/26/2020-16:42:03] [V] [TRT] Tactic: 6645123197870846056 time 0.502784 [11/26/2020-16:42:03] [V] [TRT] Conv_44 + Relu_45 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148t_nt_v1 [11/26/2020-16:42:03] [V] [TRT] Tactic: 7852627285308570038 time 0.214016 [11/26/2020-16:42:03] [V] [TRT] Conv_44 + Relu_45 (scudnn) Set Tactic Name: maxwell_scudnn_128x128_relu_large_nn_v1 [11/26/2020-16:42:03] [V] [TRT] Tactic: -9137461792520977713 time 0.504832 [11/26/2020-16:42:03] [V] [TRT] Conv_44 + Relu_45 (scudnn) Set Tactic Name: maxwell_scudnn_128x32_relu_large_nn_v1 [11/26/2020-16:42:03] [V] [TRT] Tactic: -6092040395344634144 time 0.60416 [11/26/2020-16:42:03] [V] [TRT] Conv_44 + Relu_45 (scudnn) Set Tactic Name: maxwell_scudnn_128x32_relu_small_nn_v1 [11/26/2020-16:42:03] [V] [TRT] Tactic: -3456450830548107839 time 0.497088 [11/26/2020-16:42:03] [V] [TRT] Conv_44 + Relu_45 (scudnn) Set Tactic Name: maxwell_scudnn_128x128_relu_small_nn_v1 [11/26/2020-16:42:03] [V] [TRT] Tactic: -410470605513481746 time 0.477184 [11/26/2020-16:42:03] [V] [TRT] Fastest Tactic: 3827454225649558724 Time: 0.196608 [11/26/2020-16:42:03] [V] [TRT] --------------- Timing Runner: Conv_44 + Relu_45 (CudaConvolution) [11/26/2020-16:42:03] [V] [TRT] Tactic: 0 time 0.722944 [11/26/2020-16:42:03] [V] [TRT] Tactic: 1 time 0.541696 [11/26/2020-16:42:03] [V] [TRT] Tactic: 2 time 0.641024 [11/26/2020-16:42:03] [V] [TRT] Tactic: 5 time 15.019 [11/26/2020-16:42:03] [V] [TRT] Tactic: 6 time 0.294656 [11/26/2020-16:42:03] [V] [TRT] Fastest Tactic: 6 Time: 0.294656 [11/26/2020-16:42:03] [V] [TRT] --------------- Timing Runner: Conv_44 + Relu_45 (CudaDepthwiseConvolution) [11/26/2020-16:42:03] [V] [TRT] CudaDepthwiseConvolution has no valid tactics for this config, skipping [11/26/2020-16:42:03] [V] [TRT] >>>>>>>>>>>>>>> Chose Runner Type: CaskConvolution Tactic: 3827454225649558724 [11/26/2020-16:42:03] [V] [TRT] Conv_44 + Relu_45 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148n_nt_v1 [11/26/2020-16:42:03] [V] [TRT] [11/26/2020-16:42:03] [V] [TRT] *************** Autotuning format combination: Half(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2),(MUL_ADD 8 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 8192)) -> Half(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2),(MUL_ADD 8 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 8192)) *************** [11/26/2020-16:42:03] [V] [TRT] --------------- Timing Runner: Conv_44 + Relu_45 (FusedConvActConvolution) [11/26/2020-16:42:03] [V] [TRT] FusedConvActConvolution has no valid tactics for this config, skipping [11/26/2020-16:42:03] [V] [TRT] --------------- Timing Runner: Conv_44 + Relu_45 (CaskConvolution) [11/26/2020-16:42:03] [V] [TRT] CaskConvolution has no valid tactics for this config, skipping [11/26/2020-16:42:03] [V] [TRT] --------------- Timing Runner: Conv_44 + Relu_45 (CudaConvolution) [11/26/2020-16:42:03] [V] [TRT] Tactic: 0 time 0.78336 [11/26/2020-16:42:03] [V] [TRT] Tactic: 1 time 0.553984 [11/26/2020-16:42:03] [V] [TRT] Tactic: 2 time 0.641024 [11/26/2020-16:42:03] [V] [TRT] Tactic: 5 time 14.6063 [11/26/2020-16:42:03] [V] [TRT] Tactic: 6 time 0.274432 [11/26/2020-16:42:03] [V] [TRT] Fastest Tactic: 6 Time: 0.274432 [11/26/2020-16:42:03] [V] [TRT] --------------- Timing Runner: Conv_44 + Relu_45 (CudaDepthwiseConvolution) [11/26/2020-16:42:03] [V] [TRT] CudaDepthwiseConvolution has no valid tactics for this config, skipping [11/26/2020-16:42:03] [V] [TRT] >>>>>>>>>>>>>>> Chose Runner Type: CudaConvolution Tactic: 6 [11/26/2020-16:42:03] [V] [TRT] [11/26/2020-16:42:03] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:42:03] [V] [TRT] Tactic: 0 time 0.016384 [11/26/2020-16:42:03] [V] [TRT] Fastest Tactic: 0 Time: 0.016384 [11/26/2020-16:42:03] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:42:03] [V] [TRT] Tactic: 0 time 0.011264 [11/26/2020-16:42:03] [V] [TRT] Fastest Tactic: 0 Time: 0.011264 [11/26/2020-16:42:03] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:42:03] [V] [TRT] Tactic: 0 time 0.014336 [11/26/2020-16:42:03] [V] [TRT] Fastest Tactic: 0 Time: 0.014336 [11/26/2020-16:42:03] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:42:03] [V] [TRT] Tactic: 0 time 0.011168 [11/26/2020-16:42:03] [V] [TRT] Fastest Tactic: 0 Time: 0.011168 [11/26/2020-16:42:03] [V] [TRT] *************** Autotuning format combination: Float(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2),(MUL_ADD 8 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 8192)), Float(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2),(MUL_ADD 8 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 8192)) -> Float(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2),(MUL_ADD 8 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 8192)) *************** [11/26/2020-16:42:03] [V] [TRT] --------------- Timing Runner: Conv_46 + Add_47 + Relu_48 (LegacySASSConvolution) [11/26/2020-16:42:03] [V] [TRT] Tactic: 0 time 0.555008 [11/26/2020-16:42:03] [V] [TRT] Tactic: 1 time 0.23248 [11/26/2020-16:42:03] [V] [TRT] Fastest Tactic: 1 Time: 0.23248 [11/26/2020-16:42:03] [V] [TRT] --------------- Timing Runner: Conv_46 + Add_47 + Relu_48 (CaskConvolution) [11/26/2020-16:42:03] [V] [TRT] Conv_46 + Add_47 + Relu_48 (scudnn) Set Tactic Name: maxwell_scudnn_128x32_relu_medium_nn_v1 [11/26/2020-16:42:03] [V] [TRT] Tactic: 1062367460111450758 time 0.568096 [11/26/2020-16:42:03] [V] [TRT] Conv_46 + Add_47 + Relu_48 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148n_nt_v1 [11/26/2020-16:42:03] [V] [TRT] Tactic: 3827454225649558724 time 0.205824 [11/26/2020-16:42:03] [V] [TRT] Conv_46 + Add_47 + Relu_48 (scudnn) Set Tactic Name: maxwell_scudnn_128x64_relu_large_nn_v1 [11/26/2020-16:42:03] [V] [TRT] Tactic: 4337000649858996379 time 0.516096 [11/26/2020-16:42:03] [V] [TRT] Conv_46 + Add_47 + Relu_48 (scudnn) Set Tactic Name: maxwell_scudnn_128x128_relu_medium_nn_v1 [11/26/2020-16:42:03] [V] [TRT] Tactic: 4501471010995462441 time 0.500736 [11/26/2020-16:42:03] [V] [TRT] Conv_46 + Add_47 + Relu_48 (scudnn) Set Tactic Name: maxwell_scudnn_128x64_relu_small_nn_v1 [11/26/2020-16:42:03] [V] [TRT] Tactic: 5137655947464784826 time 0.457728 [11/26/2020-16:42:03] [V] [TRT] Conv_46 + Add_47 + Relu_48 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148m_nt_v1 [11/26/2020-16:42:03] [V] [TRT] Tactic: 5921334924264294896 time 0.208448 [11/26/2020-16:42:03] [V] [TRT] Conv_46 + Add_47 + Relu_48 (scudnn) Set Tactic Name: maxwell_scudnn_128x64_relu_medium_nn_v1 [11/26/2020-16:42:03] [V] [TRT] Tactic: 6645123197870846056 time 0.504832 [11/26/2020-16:42:03] [V] [TRT] Conv_46 + Add_47 + Relu_48 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148t_nt_v1 [11/26/2020-16:42:03] [V] [TRT] Tactic: 7852627285308570038 time 0.214016 [11/26/2020-16:42:03] [V] [TRT] Conv_46 + Add_47 + Relu_48 (scudnn) Set Tactic Name: maxwell_scudnn_128x128_relu_large_nn_v1 [11/26/2020-16:42:03] [V] [TRT] Tactic: -9137461792520977713 time 0.512 [11/26/2020-16:42:03] [V] [TRT] Conv_46 + Add_47 + Relu_48 (scudnn) Set Tactic Name: maxwell_scudnn_128x32_relu_large_nn_v1 [11/26/2020-16:42:03] [V] [TRT] Tactic: -6092040395344634144 time 0.611328 [11/26/2020-16:42:03] [V] [TRT] Conv_46 + Add_47 + Relu_48 (scudnn) Set Tactic Name: maxwell_scudnn_128x32_relu_small_nn_v1 [11/26/2020-16:42:03] [V] [TRT] Tactic: -3456450830548107839 time 0.490496 [11/26/2020-16:42:03] [V] [TRT] Conv_46 + Add_47 + Relu_48 (scudnn) Set Tactic Name: maxwell_scudnn_128x128_relu_small_nn_v1 [11/26/2020-16:42:03] [V] [TRT] Tactic: -410470605513481746 time 0.475936 [11/26/2020-16:42:03] [V] [TRT] Fastest Tactic: 3827454225649558724 Time: 0.205824 [11/26/2020-16:42:03] [V] [TRT] --------------- Timing Runner: Conv_46 + Add_47 + Relu_48 (CudaConvolution) [11/26/2020-16:42:03] [V] [TRT] Tactic: 0 time 0.722944 [11/26/2020-16:42:03] [V] [TRT] Tactic: 1 time 0.564224 [11/26/2020-16:42:03] [V] [TRT] Tactic: 2 time 0.65024 [11/26/2020-16:42:03] [V] [TRT] Tactic: 5 time 15.787 [11/26/2020-16:42:03] [V] [TRT] Tactic: 6 time 0.301056 [11/26/2020-16:42:03] [V] [TRT] Fastest Tactic: 6 Time: 0.301056 [11/26/2020-16:42:03] [V] [TRT] >>>>>>>>>>>>>>> Chose Runner Type: CaskConvolution Tactic: 3827454225649558724 [11/26/2020-16:42:03] [V] [TRT] Conv_46 + Add_47 + Relu_48 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148n_nt_v1 [11/26/2020-16:42:03] [V] [TRT] [11/26/2020-16:42:03] [V] [TRT] *************** Autotuning format combination: Half(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2),(MUL_ADD 8 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 8192)), Half(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2),(MUL_ADD 8 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 8192)) -> Half(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2),(MUL_ADD 8 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 8192)) *************** [11/26/2020-16:42:03] [V] [TRT] --------------- Timing Runner: Conv_46 + Add_47 + Relu_48 (CaskConvolution) [11/26/2020-16:42:03] [V] [TRT] CaskConvolution has no valid tactics for this config, skipping [11/26/2020-16:42:03] [V] [TRT] --------------- Timing Runner: Conv_46 + Add_47 + Relu_48 (CudaConvolution) [11/26/2020-16:42:03] [V] [TRT] Tactic: 0 time 0.77824 [11/26/2020-16:42:03] [V] [TRT] Tactic: 1 time 0.5632 [11/26/2020-16:42:03] [V] [TRT] Tactic: 2 time 0.643072 [11/26/2020-16:42:04] [V] [TRT] Tactic: 5 time 14.4527 [11/26/2020-16:42:04] [V] [TRT] Tactic: 6 time 0.282624 [11/26/2020-16:42:04] [V] [TRT] Fastest Tactic: 6 Time: 0.282624 [11/26/2020-16:42:04] [V] [TRT] >>>>>>>>>>>>>>> Chose Runner Type: CudaConvolution Tactic: 6 [11/26/2020-16:42:04] [V] [TRT] [11/26/2020-16:42:04] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:42:04] [V] [TRT] Tactic: 0 time 0.014336 [11/26/2020-16:42:04] [V] [TRT] Fastest Tactic: 0 Time: 0.014336 [11/26/2020-16:42:04] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:42:04] [V] [TRT] Tactic: 0 time 0.011072 [11/26/2020-16:42:04] [V] [TRT] Fastest Tactic: 0 Time: 0.011072 [11/26/2020-16:42:04] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:42:04] [V] [TRT] Tactic: 0 time 0.015136 [11/26/2020-16:42:04] [V] [TRT] Fastest Tactic: 0 Time: 0.015136 [11/26/2020-16:42:04] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:42:04] [V] [TRT] Tactic: 0 time 0.011168 [11/26/2020-16:42:04] [V] [TRT] Fastest Tactic: 0 Time: 0.011168 [11/26/2020-16:42:04] [V] [TRT] *************** Autotuning format combination: Float(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2),(MUL_ADD 8 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 8192)) -> Float(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2),(MUL_ADD 8 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 8192)) *************** [11/26/2020-16:42:04] [V] [TRT] --------------- Timing Runner: Conv_49 + Relu_50 (LegacySASSConvolution) [11/26/2020-16:42:04] [V] [TRT] Tactic: 0 time 0.551936 [11/26/2020-16:42:04] [V] [TRT] Tactic: 1 time 0.230496 [11/26/2020-16:42:04] [V] [TRT] Fastest Tactic: 1 Time: 0.230496 [11/26/2020-16:42:04] [V] [TRT] --------------- Timing Runner: Conv_49 + Relu_50 (FusedConvActConvolution) [11/26/2020-16:42:04] [V] [TRT] Tactic: 7 time 0.566272 [11/26/2020-16:42:04] [V] [TRT] Tactic: 10 time 0.538624 [11/26/2020-16:42:04] [V] [TRT] Tactic: 14 time 0.316416 [11/26/2020-16:42:04] [V] [TRT] Tactic: 15 time 0.343776 [11/26/2020-16:42:04] [V] [TRT] Tactic: 25 time 0.342016 [11/26/2020-16:42:04] [V] [TRT] Tactic: 26 time 0.509952 [11/26/2020-16:42:04] [V] [TRT] Tactic: 29 time 0.473088 [11/26/2020-16:42:04] [V] [TRT] Tactic: 30 time 0.49152 [11/26/2020-16:42:04] [V] [TRT] Tactic: 33 time 0.615168 [11/26/2020-16:42:04] [V] [TRT] Tactic: 36 time 0.864928 [11/26/2020-16:42:04] [V] [TRT] Tactic: 39 time 0.384 [11/26/2020-16:42:04] [V] [TRT] Tactic: 41 time 0.451584 [11/26/2020-16:42:04] [V] [TRT] Tactic: 42 time 0.63488 [11/26/2020-16:42:04] [V] [TRT] Tactic: 43 time 0.614048 [11/26/2020-16:42:04] [V] [TRT] Tactic: 45 time 0.318464 [11/26/2020-16:42:04] [V] [TRT] Tactic: 47 time 0.405504 [11/26/2020-16:42:04] [V] [TRT] Tactic: 52 time 0.391168 [11/26/2020-16:42:04] [V] [TRT] Tactic: 54 time 0.607008 [11/26/2020-16:42:04] [V] [TRT] Tactic: 56 time 0.442368 [11/26/2020-16:42:04] [V] [TRT] Tactic: 66 time 0.692224 [11/26/2020-16:42:04] [V] [TRT] Tactic: 76 time 0.381952 [11/26/2020-16:42:04] [V] [TRT] Tactic: 90 time 0.423552 [11/26/2020-16:42:04] [V] [TRT] Tactic: 93 time 0.387072 [11/26/2020-16:42:04] [V] [TRT] Tactic: 98 time 0.581632 [11/26/2020-16:42:04] [V] [TRT] Tactic: 104 time 0.534528 [11/26/2020-16:42:04] [V] [TRT] Tactic: 110 time 0.456704 [11/26/2020-16:42:04] [V] [TRT] Tactic: 119 time 0.62464 [11/26/2020-16:42:04] [V] [TRT] Tactic: 121 time 0.412672 [11/26/2020-16:42:04] [V] [TRT] Tactic: 130 time 0.572416 [11/26/2020-16:42:05] [V] [TRT] Tactic: 134 time 0.590848 [11/26/2020-16:42:05] [V] [TRT] Tactic: 136 time 0.539648 [11/26/2020-16:42:05] [V] [TRT] Tactic: 137 time 0.571392 [11/26/2020-16:42:05] [V] [TRT] Tactic: 139 time 0.616384 [11/26/2020-16:42:05] [V] [TRT] Tactic: 144 time 0.52736 [11/26/2020-16:42:05] [V] [TRT] Tactic: 149 time 0.383936 [11/26/2020-16:42:05] [V] [TRT] Tactic: 151 time 0.627712 [11/26/2020-16:42:05] [V] [TRT] Tactic: 152 time 0.47616 [11/26/2020-16:42:05] [V] [TRT] Tactic: 153 time 0.691744 [11/26/2020-16:42:05] [V] [TRT] Tactic: 156 time 0.468992 [11/26/2020-16:42:05] [V] [TRT] Tactic: 159 time 0.4328 [11/26/2020-16:42:05] [V] [TRT] Tactic: 162 time 0.507904 [11/26/2020-16:42:05] [V] [TRT] Tactic: 164 time 0.368192 [11/26/2020-16:42:05] [V] [TRT] Fastest Tactic: 14 Time: 0.316416 [11/26/2020-16:42:05] [V] [TRT] --------------- Timing Runner: Conv_49 + Relu_50 (CaskConvolution) [11/26/2020-16:42:05] [V] [TRT] Conv_49 + Relu_50 (scudnn) Set Tactic Name: maxwell_scudnn_128x32_relu_medium_nn_v1 [11/26/2020-16:42:05] [V] [TRT] Tactic: 1062367460111450758 time 0.560128 [11/26/2020-16:42:05] [V] [TRT] Conv_49 + Relu_50 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148n_nt_v1 [11/26/2020-16:42:05] [V] [TRT] Tactic: 3827454225649558724 time 0.197632 [11/26/2020-16:42:05] [V] [TRT] Conv_49 + Relu_50 (scudnn) Set Tactic Name: maxwell_scudnn_128x64_relu_large_nn_v1 [11/26/2020-16:42:05] [V] [TRT] Tactic: 4337000649858996379 time 0.516096 [11/26/2020-16:42:05] [V] [TRT] Conv_49 + Relu_50 (scudnn) Set Tactic Name: maxwell_scudnn_128x128_relu_medium_nn_v1 [11/26/2020-16:42:05] [V] [TRT] Tactic: 4501471010995462441 time 0.493568 [11/26/2020-16:42:05] [V] [TRT] Conv_49 + Relu_50 (scudnn) Set Tactic Name: maxwell_scudnn_128x64_relu_small_nn_v1 [11/26/2020-16:42:05] [V] [TRT] Tactic: 5137655947464784826 time 0.4608 [11/26/2020-16:42:05] [V] [TRT] Conv_49 + Relu_50 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148m_nt_v1 [11/26/2020-16:42:05] [V] [TRT] Tactic: 5921334924264294896 time 0.2048 [11/26/2020-16:42:05] [V] [TRT] Conv_49 + Relu_50 (scudnn) Set Tactic Name: maxwell_scudnn_128x64_relu_medium_nn_v1 [11/26/2020-16:42:05] [V] [TRT] Tactic: 6645123197870846056 time 0.49664 [11/26/2020-16:42:05] [V] [TRT] Conv_49 + Relu_50 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148t_nt_v1 [11/26/2020-16:42:05] [V] [TRT] Tactic: 7852627285308570038 time 0.21504 [11/26/2020-16:42:05] [V] [TRT] Conv_49 + Relu_50 (scudnn) Set Tactic Name: maxwell_scudnn_128x128_relu_large_nn_v1 [11/26/2020-16:42:05] [V] [TRT] Tactic: -9137461792520977713 time 0.50688 [11/26/2020-16:42:05] [V] [TRT] Conv_49 + Relu_50 (scudnn) Set Tactic Name: maxwell_scudnn_128x32_relu_large_nn_v1 [11/26/2020-16:42:05] [V] [TRT] Tactic: -6092040395344634144 time 0.6016 [11/26/2020-16:42:05] [V] [TRT] Conv_49 + Relu_50 (scudnn) Set Tactic Name: maxwell_scudnn_128x32_relu_small_nn_v1 [11/26/2020-16:42:05] [V] [TRT] Tactic: -3456450830548107839 time 0.497216 [11/26/2020-16:42:05] [V] [TRT] Conv_49 + Relu_50 (scudnn) Set Tactic Name: maxwell_scudnn_128x128_relu_small_nn_v1 [11/26/2020-16:42:05] [V] [TRT] Tactic: -410470605513481746 time 0.477184 [11/26/2020-16:42:05] [V] [TRT] Fastest Tactic: 3827454225649558724 Time: 0.197632 [11/26/2020-16:42:05] [V] [TRT] --------------- Timing Runner: Conv_49 + Relu_50 (CudaConvolution) [11/26/2020-16:42:05] [V] [TRT] Tactic: 0 time 0.720896 [11/26/2020-16:42:05] [V] [TRT] Tactic: 1 time 0.536576 [11/26/2020-16:42:05] [V] [TRT] Tactic: 2 time 0.644096 [11/26/2020-16:42:05] [V] [TRT] Tactic: 5 time 15.4399 [11/26/2020-16:42:05] [V] [TRT] Tactic: 6 time 0.295936 [11/26/2020-16:42:05] [V] [TRT] Fastest Tactic: 6 Time: 0.295936 [11/26/2020-16:42:05] [V] [TRT] --------------- Timing Runner: Conv_49 + Relu_50 (CudaDepthwiseConvolution) [11/26/2020-16:42:05] [V] [TRT] CudaDepthwiseConvolution has no valid tactics for this config, skipping [11/26/2020-16:42:05] [V] [TRT] >>>>>>>>>>>>>>> Chose Runner Type: CaskConvolution Tactic: 3827454225649558724 [11/26/2020-16:42:05] [V] [TRT] Conv_49 + Relu_50 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148n_nt_v1 [11/26/2020-16:42:05] [V] [TRT] [11/26/2020-16:42:05] [V] [TRT] *************** Autotuning format combination: Half(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2),(MUL_ADD 8 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 8192)) -> Half(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2),(MUL_ADD 8 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 8192)) *************** [11/26/2020-16:42:05] [V] [TRT] --------------- Timing Runner: Conv_49 + Relu_50 (FusedConvActConvolution) [11/26/2020-16:42:05] [V] [TRT] FusedConvActConvolution has no valid tactics for this config, skipping [11/26/2020-16:42:05] [V] [TRT] --------------- Timing Runner: Conv_49 + Relu_50 (CaskConvolution) [11/26/2020-16:42:05] [V] [TRT] CaskConvolution has no valid tactics for this config, skipping [11/26/2020-16:42:05] [V] [TRT] --------------- Timing Runner: Conv_49 + Relu_50 (CudaConvolution) [11/26/2020-16:42:05] [V] [TRT] Tactic: 0 time 0.77824 [11/26/2020-16:42:05] [V] [TRT] Tactic: 1 time 0.557056 [11/26/2020-16:42:05] [V] [TRT] Tactic: 2 time 0.64 [11/26/2020-16:42:05] [V] [TRT] Tactic: 5 time 15.2228 [11/26/2020-16:42:05] [V] [TRT] Tactic: 6 time 0.275456 [11/26/2020-16:42:05] [V] [TRT] Fastest Tactic: 6 Time: 0.275456 [11/26/2020-16:42:05] [V] [TRT] --------------- Timing Runner: Conv_49 + Relu_50 (CudaDepthwiseConvolution) [11/26/2020-16:42:05] [V] [TRT] CudaDepthwiseConvolution has no valid tactics for this config, skipping [11/26/2020-16:42:05] [V] [TRT] >>>>>>>>>>>>>>> Chose Runner Type: CudaConvolution Tactic: 6 [11/26/2020-16:42:05] [V] [TRT] [11/26/2020-16:42:05] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:42:05] [V] [TRT] Tactic: 0 time 0.014336 [11/26/2020-16:42:05] [V] [TRT] Fastest Tactic: 0 Time: 0.014336 [11/26/2020-16:42:05] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:42:05] [V] [TRT] Tactic: 0 time 0.011552 [11/26/2020-16:42:05] [V] [TRT] Fastest Tactic: 0 Time: 0.011552 [11/26/2020-16:42:05] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:42:05] [V] [TRT] Tactic: 0 time 0.014336 [11/26/2020-16:42:05] [V] [TRT] Fastest Tactic: 0 Time: 0.014336 [11/26/2020-16:42:05] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:42:05] [V] [TRT] Tactic: 0 time 0.011264 [11/26/2020-16:42:05] [V] [TRT] Fastest Tactic: 0 Time: 0.011264 [11/26/2020-16:42:05] [V] [TRT] *************** Autotuning format combination: Float(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2),(MUL_ADD 8 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 8192)), Float(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2),(MUL_ADD 8 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 8192)) -> Float(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2),(MUL_ADD 8 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 8192)) *************** [11/26/2020-16:42:05] [V] [TRT] --------------- Timing Runner: Conv_51 + Add_52 + Relu_53 (LegacySASSConvolution) [11/26/2020-16:42:05] [V] [TRT] Tactic: 0 time 0.555008 [11/26/2020-16:42:05] [V] [TRT] Tactic: 1 time 0.23136 [11/26/2020-16:42:05] [V] [TRT] Fastest Tactic: 1 Time: 0.23136 [11/26/2020-16:42:05] [V] [TRT] --------------- Timing Runner: Conv_51 + Add_52 + Relu_53 (CaskConvolution) [11/26/2020-16:42:05] [V] [TRT] Conv_51 + Add_52 + Relu_53 (scudnn) Set Tactic Name: maxwell_scudnn_128x32_relu_medium_nn_v1 [11/26/2020-16:42:05] [V] [TRT] Tactic: 1062367460111450758 time 0.57856 [11/26/2020-16:42:05] [V] [TRT] Conv_51 + Add_52 + Relu_53 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148n_nt_v1 [11/26/2020-16:42:05] [V] [TRT] Tactic: 3827454225649558724 time 0.22624 [11/26/2020-16:42:05] [V] [TRT] Conv_51 + Add_52 + Relu_53 (scudnn) Set Tactic Name: maxwell_scudnn_128x64_relu_large_nn_v1 [11/26/2020-16:42:05] [V] [TRT] Tactic: 4337000649858996379 time 0.518144 [11/26/2020-16:42:05] [V] [TRT] Conv_51 + Add_52 + Relu_53 (scudnn) Set Tactic Name: maxwell_scudnn_128x128_relu_medium_nn_v1 [11/26/2020-16:42:05] [V] [TRT] Tactic: 4501471010995462441 time 0.49152 [11/26/2020-16:42:05] [V] [TRT] Conv_51 + Add_52 + Relu_53 (scudnn) Set Tactic Name: maxwell_scudnn_128x64_relu_small_nn_v1 [11/26/2020-16:42:05] [V] [TRT] Tactic: 5137655947464784826 time 0.462848 [11/26/2020-16:42:05] [V] [TRT] Conv_51 + Add_52 + Relu_53 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148m_nt_v1 [11/26/2020-16:42:05] [V] [TRT] Tactic: 5921334924264294896 time 0.20992 [11/26/2020-16:42:05] [V] [TRT] Conv_51 + Add_52 + Relu_53 (scudnn) Set Tactic Name: maxwell_scudnn_128x64_relu_medium_nn_v1 [11/26/2020-16:42:06] [V] [TRT] Tactic: 6645123197870846056 time 0.494752 [11/26/2020-16:42:06] [V] [TRT] Conv_51 + Add_52 + Relu_53 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148t_nt_v1 [11/26/2020-16:42:06] [V] [TRT] Tactic: 7852627285308570038 time 0.223232 [11/26/2020-16:42:06] [V] [TRT] Conv_51 + Add_52 + Relu_53 (scudnn) Set Tactic Name: maxwell_scudnn_128x128_relu_large_nn_v1 [11/26/2020-16:42:06] [V] [TRT] Tactic: -9137461792520977713 time 0.510976 [11/26/2020-16:42:06] [V] [TRT] Conv_51 + Add_52 + Relu_53 (scudnn) Set Tactic Name: maxwell_scudnn_128x32_relu_large_nn_v1 [11/26/2020-16:42:06] [V] [TRT] Tactic: -6092040395344634144 time 0.60416 [11/26/2020-16:42:06] [V] [TRT] Conv_51 + Add_52 + Relu_53 (scudnn) Set Tactic Name: maxwell_scudnn_128x32_relu_small_nn_v1 [11/26/2020-16:42:06] [V] [TRT] Tactic: -3456450830548107839 time 0.490944 [11/26/2020-16:42:06] [V] [TRT] Conv_51 + Add_52 + Relu_53 (scudnn) Set Tactic Name: maxwell_scudnn_128x128_relu_small_nn_v1 [11/26/2020-16:42:06] [V] [TRT] Tactic: -410470605513481746 time 0.47616 [11/26/2020-16:42:06] [V] [TRT] Fastest Tactic: 5921334924264294896 Time: 0.20992 [11/26/2020-16:42:06] [V] [TRT] --------------- Timing Runner: Conv_51 + Add_52 + Relu_53 (CudaConvolution) [11/26/2020-16:42:06] [V] [TRT] Tactic: 0 time 0.72192 [11/26/2020-16:42:06] [V] [TRT] Tactic: 1 time 0.55808 [11/26/2020-16:42:06] [V] [TRT] Tactic: 2 time 0.648384 [11/26/2020-16:42:06] [V] [TRT] Tactic: 5 time 15.9775 [11/26/2020-16:42:06] [V] [TRT] Tactic: 6 time 0.30208 [11/26/2020-16:42:06] [V] [TRT] Fastest Tactic: 6 Time: 0.30208 [11/26/2020-16:42:06] [V] [TRT] >>>>>>>>>>>>>>> Chose Runner Type: CaskConvolution Tactic: 5921334924264294896 [11/26/2020-16:42:06] [V] [TRT] Conv_51 + Add_52 + Relu_53 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148m_nt_v1 [11/26/2020-16:42:06] [V] [TRT] [11/26/2020-16:42:06] [V] [TRT] *************** Autotuning format combination: Half(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2),(MUL_ADD 8 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 8192)), Half(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2),(MUL_ADD 8 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 8192)) -> Half(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2),(MUL_ADD 8 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 8192)) *************** [11/26/2020-16:42:06] [V] [TRT] --------------- Timing Runner: Conv_51 + Add_52 + Relu_53 (CaskConvolution) [11/26/2020-16:42:06] [V] [TRT] CaskConvolution has no valid tactics for this config, skipping [11/26/2020-16:42:06] [V] [TRT] --------------- Timing Runner: Conv_51 + Add_52 + Relu_53 (CudaConvolution) [11/26/2020-16:42:06] [V] [TRT] Tactic: 0 time 0.785408 [11/26/2020-16:42:06] [V] [TRT] Tactic: 1 time 0.569344 [11/26/2020-16:42:06] [V] [TRT] Tactic: 2 time 0.64512 [11/26/2020-16:42:06] [V] [TRT] Tactic: 5 time 14.7927 [11/26/2020-16:42:06] [V] [TRT] Tactic: 6 time 0.2816 [11/26/2020-16:42:06] [V] [TRT] Fastest Tactic: 6 Time: 0.2816 [11/26/2020-16:42:06] [V] [TRT] >>>>>>>>>>>>>>> Chose Runner Type: CudaConvolution Tactic: 6 [11/26/2020-16:42:06] [V] [TRT] [11/26/2020-16:42:06] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:42:06] [V] [TRT] Tactic: 0 time 0.014336 [11/26/2020-16:42:06] [V] [TRT] Fastest Tactic: 0 Time: 0.014336 [11/26/2020-16:42:06] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:42:06] [V] [TRT] Tactic: 0 time 0.012672 [11/26/2020-16:42:06] [V] [TRT] Fastest Tactic: 0 Time: 0.012672 [11/26/2020-16:42:06] [V] [TRT] *************** Autotuning format combination: Float(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2),(MUL_ADD 8 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 8192)) -> Float(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2),(MUL_ADD 8 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 8192)) *************** [11/26/2020-16:42:06] [V] [TRT] --------------- Timing Runner: Conv_54 + Relu_55 (LegacySASSConvolution) [11/26/2020-16:42:06] [V] [TRT] Tactic: 0 time 0.551936 [11/26/2020-16:42:06] [V] [TRT] Tactic: 1 time 0.229376 [11/26/2020-16:42:06] [V] [TRT] Fastest Tactic: 1 Time: 0.229376 [11/26/2020-16:42:06] [V] [TRT] --------------- Timing Runner: Conv_54 + Relu_55 (FusedConvActConvolution) [11/26/2020-16:42:06] [V] [TRT] Tactic: 7 time 0.565248 [11/26/2020-16:42:06] [V] [TRT] Tactic: 10 time 0.53248 [11/26/2020-16:42:06] [V] [TRT] Tactic: 14 time 0.31088 [11/26/2020-16:42:06] [V] [TRT] Tactic: 15 time 0.342016 [11/26/2020-16:42:06] [V] [TRT] Tactic: 25 time 0.340992 [11/26/2020-16:42:06] [V] [TRT] Tactic: 26 time 0.504384 [11/26/2020-16:42:06] [V] [TRT] Tactic: 29 time 0.468992 [11/26/2020-16:42:06] [V] [TRT] Tactic: 30 time 0.489472 [11/26/2020-16:42:06] [V] [TRT] Tactic: 33 time 0.613952 [11/26/2020-16:42:06] [V] [TRT] Tactic: 36 time 0.863232 [11/26/2020-16:42:06] [V] [TRT] Tactic: 39 time 0.373248 [11/26/2020-16:42:06] [V] [TRT] Tactic: 41 time 0.459776 [11/26/2020-16:42:06] [V] [TRT] Tactic: 42 time 0.637952 [11/26/2020-16:42:06] [V] [TRT] Tactic: 43 time 0.607232 [11/26/2020-16:42:06] [V] [TRT] Tactic: 45 time 0.31904 [11/26/2020-16:42:06] [V] [TRT] Tactic: 47 time 0.396864 [11/26/2020-16:42:06] [V] [TRT] Tactic: 52 time 0.388096 [11/26/2020-16:42:06] [V] [TRT] Tactic: 54 time 0.608256 [11/26/2020-16:42:06] [V] [TRT] Tactic: 56 time 0.438272 [11/26/2020-16:42:06] [V] [TRT] Tactic: 66 time 0.6912 [11/26/2020-16:42:06] [V] [TRT] Tactic: 76 time 0.36544 [11/26/2020-16:42:06] [V] [TRT] Tactic: 90 time 0.423808 [11/26/2020-16:42:06] [V] [TRT] Tactic: 93 time 0.383552 [11/26/2020-16:42:07] [V] [TRT] Tactic: 98 time 0.62464 [11/26/2020-16:42:07] [V] [TRT] Tactic: 104 time 0.540672 [11/26/2020-16:42:07] [V] [TRT] Tactic: 110 time 0.452192 [11/26/2020-16:42:07] [V] [TRT] Tactic: 119 time 0.620544 [11/26/2020-16:42:07] [V] [TRT] Tactic: 121 time 0.410688 [11/26/2020-16:42:07] [V] [TRT] Tactic: 130 time 0.574464 [11/26/2020-16:42:07] [V] [TRT] Tactic: 134 time 0.594592 [11/26/2020-16:42:07] [V] [TRT] Tactic: 136 time 0.539648 [11/26/2020-16:42:07] [V] [TRT] Tactic: 137 time 0.571392 [11/26/2020-16:42:07] [V] [TRT] Tactic: 139 time 0.608832 [11/26/2020-16:42:07] [V] [TRT] Tactic: 144 time 0.52736 [11/26/2020-16:42:07] [V] [TRT] Tactic: 149 time 0.383872 [11/26/2020-16:42:07] [V] [TRT] Tactic: 151 time 0.629472 [11/26/2020-16:42:07] [V] [TRT] Tactic: 152 time 0.47616 [11/26/2020-16:42:07] [V] [TRT] Tactic: 153 time 0.69424 [11/26/2020-16:42:07] [V] [TRT] Tactic: 156 time 0.470592 [11/26/2020-16:42:07] [V] [TRT] Tactic: 159 time 0.425536 [11/26/2020-16:42:07] [V] [TRT] Tactic: 162 time 0.498688 [11/26/2020-16:42:07] [V] [TRT] Tactic: 164 time 0.367616 [11/26/2020-16:42:07] [V] [TRT] Fastest Tactic: 14 Time: 0.31088 [11/26/2020-16:42:07] [V] [TRT] --------------- Timing Runner: Conv_54 + Relu_55 (CaskConvolution) [11/26/2020-16:42:07] [V] [TRT] Conv_54 + Relu_55 (scudnn) Set Tactic Name: maxwell_scudnn_128x32_relu_medium_nn_v1 [11/26/2020-16:42:07] [V] [TRT] Tactic: 1062367460111450758 time 0.556832 [11/26/2020-16:42:07] [V] [TRT] Conv_54 + Relu_55 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148n_nt_v1 [11/26/2020-16:42:07] [V] [TRT] Tactic: 3827454225649558724 time 0.195584 [11/26/2020-16:42:07] [V] [TRT] Conv_54 + Relu_55 (scudnn) Set Tactic Name: maxwell_scudnn_128x64_relu_large_nn_v1 [11/26/2020-16:42:07] [V] [TRT] Tactic: 4337000649858996379 time 0.51904 [11/26/2020-16:42:07] [V] [TRT] Conv_54 + Relu_55 (scudnn) Set Tactic Name: maxwell_scudnn_128x128_relu_medium_nn_v1 [11/26/2020-16:42:07] [V] [TRT] Tactic: 4501471010995462441 time 0.485376 [11/26/2020-16:42:07] [V] [TRT] Conv_54 + Relu_55 (scudnn) Set Tactic Name: maxwell_scudnn_128x64_relu_small_nn_v1 [11/26/2020-16:42:07] [V] [TRT] Tactic: 5137655947464784826 time 0.462848 [11/26/2020-16:42:07] [V] [TRT] Conv_54 + Relu_55 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148m_nt_v1 [11/26/2020-16:42:07] [V] [TRT] Tactic: 5921334924264294896 time 0.207872 [11/26/2020-16:42:07] [V] [TRT] Conv_54 + Relu_55 (scudnn) Set Tactic Name: maxwell_scudnn_128x64_relu_medium_nn_v1 [11/26/2020-16:42:07] [V] [TRT] Tactic: 6645123197870846056 time 0.497664 [11/26/2020-16:42:07] [V] [TRT] Conv_54 + Relu_55 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148t_nt_v1 [11/26/2020-16:42:07] [V] [TRT] Tactic: 7852627285308570038 time 0.212992 [11/26/2020-16:42:07] [V] [TRT] Conv_54 + Relu_55 (scudnn) Set Tactic Name: maxwell_scudnn_128x128_relu_large_nn_v1 [11/26/2020-16:42:07] [V] [TRT] Tactic: -9137461792520977713 time 0.50688 [11/26/2020-16:42:07] [V] [TRT] Conv_54 + Relu_55 (scudnn) Set Tactic Name: maxwell_scudnn_128x32_relu_large_nn_v1 [11/26/2020-16:42:07] [V] [TRT] Tactic: -6092040395344634144 time 0.610304 [11/26/2020-16:42:07] [V] [TRT] Conv_54 + Relu_55 (scudnn) Set Tactic Name: maxwell_scudnn_128x32_relu_small_nn_v1 [11/26/2020-16:42:07] [V] [TRT] Tactic: -3456450830548107839 time 0.488448 [11/26/2020-16:42:07] [V] [TRT] Conv_54 + Relu_55 (scudnn) Set Tactic Name: maxwell_scudnn_128x128_relu_small_nn_v1 [11/26/2020-16:42:07] [V] [TRT] Tactic: -410470605513481746 time 0.475136 [11/26/2020-16:42:07] [V] [TRT] Fastest Tactic: 3827454225649558724 Time: 0.195584 [11/26/2020-16:42:07] [V] [TRT] --------------- Timing Runner: Conv_54 + Relu_55 (CudaConvolution) [11/26/2020-16:42:07] [V] [TRT] Tactic: 0 time 0.72192 [11/26/2020-16:42:07] [V] [TRT] Tactic: 1 time 0.535552 [11/26/2020-16:42:07] [V] [TRT] Tactic: 2 time 0.642432 [11/26/2020-16:42:07] [V] [TRT] Tactic: 5 time 15.1173 [11/26/2020-16:42:07] [V] [TRT] Tactic: 6 time 0.293888 [11/26/2020-16:42:07] [V] [TRT] Fastest Tactic: 6 Time: 0.293888 [11/26/2020-16:42:07] [V] [TRT] --------------- Timing Runner: Conv_54 + Relu_55 (CudaDepthwiseConvolution) [11/26/2020-16:42:07] [V] [TRT] CudaDepthwiseConvolution has no valid tactics for this config, skipping [11/26/2020-16:42:07] [V] [TRT] >>>>>>>>>>>>>>> Chose Runner Type: CaskConvolution Tactic: 3827454225649558724 [11/26/2020-16:42:07] [V] [TRT] Conv_54 + Relu_55 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148n_nt_v1 [11/26/2020-16:42:07] [V] [TRT] [11/26/2020-16:42:07] [V] [TRT] *************** Autotuning format combination: Half(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2),(MUL_ADD 8 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 8192)) -> Half(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2),(MUL_ADD 8 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 8192)) *************** [11/26/2020-16:42:07] [V] [TRT] --------------- Timing Runner: Conv_54 + Relu_55 (FusedConvActConvolution) [11/26/2020-16:42:07] [V] [TRT] FusedConvActConvolution has no valid tactics for this config, skipping [11/26/2020-16:42:07] [V] [TRT] --------------- Timing Runner: Conv_54 + Relu_55 (CaskConvolution) [11/26/2020-16:42:07] [V] [TRT] CaskConvolution has no valid tactics for this config, skipping [11/26/2020-16:42:07] [V] [TRT] --------------- Timing Runner: Conv_54 + Relu_55 (CudaConvolution) [11/26/2020-16:42:07] [V] [TRT] Tactic: 0 time 0.77824 [11/26/2020-16:42:07] [V] [TRT] Tactic: 1 time 0.564224 [11/26/2020-16:42:07] [V] [TRT] Tactic: 2 time 0.638976 [11/26/2020-16:42:07] [V] [TRT] Tactic: 5 time 14.2848 [11/26/2020-16:42:07] [V] [TRT] Tactic: 6 time 0.278528 [11/26/2020-16:42:07] [V] [TRT] Fastest Tactic: 6 Time: 0.278528 [11/26/2020-16:42:07] [V] [TRT] --------------- Timing Runner: Conv_54 + Relu_55 (CudaDepthwiseConvolution) [11/26/2020-16:42:07] [V] [TRT] CudaDepthwiseConvolution has no valid tactics for this config, skipping [11/26/2020-16:42:07] [V] [TRT] >>>>>>>>>>>>>>> Chose Runner Type: CudaConvolution Tactic: 6 [11/26/2020-16:42:07] [V] [TRT] [11/26/2020-16:42:07] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:42:07] [V] [TRT] Tactic: 0 time 0.015104 [11/26/2020-16:42:07] [V] [TRT] Fastest Tactic: 0 Time: 0.015104 [11/26/2020-16:42:07] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:42:07] [V] [TRT] Tactic: 0 time 0.011104 [11/26/2020-16:42:07] [V] [TRT] Fastest Tactic: 0 Time: 0.011104 [11/26/2020-16:42:07] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:42:07] [V] [TRT] Tactic: 0 time 0.014336 [11/26/2020-16:42:07] [V] [TRT] Fastest Tactic: 0 Time: 0.014336 [11/26/2020-16:42:07] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:42:07] [V] [TRT] Tactic: 0 time 0.012224 [11/26/2020-16:42:07] [V] [TRT] Fastest Tactic: 0 Time: 0.012224 [11/26/2020-16:42:07] [V] [TRT] *************** Autotuning format combination: Float(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2),(MUL_ADD 8 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 8192)) -> Float(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2),(MUL_ADD 8 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 8192)) *************** [11/26/2020-16:42:08] [V] [TRT] --------------- Timing Runner: Conv_56 + Relu_57 (LegacySASSConvolution) [11/26/2020-16:42:08] [V] [TRT] Tactic: 0 time 0.552288 [11/26/2020-16:42:08] [V] [TRT] Tactic: 1 time 0.228896 [11/26/2020-16:42:08] [V] [TRT] Fastest Tactic: 1 Time: 0.228896 [11/26/2020-16:42:08] [V] [TRT] --------------- Timing Runner: Conv_56 + Relu_57 (FusedConvActConvolution) [11/26/2020-16:42:08] [V] [TRT] Tactic: 7 time 0.5672 [11/26/2020-16:42:08] [V] [TRT] Tactic: 10 time 0.52832 [11/26/2020-16:42:08] [V] [TRT] Tactic: 14 time 0.313344 [11/26/2020-16:42:08] [V] [TRT] Tactic: 15 time 0.344064 [11/26/2020-16:42:08] [V] [TRT] Tactic: 25 time 0.34816 [11/26/2020-16:42:08] [V] [TRT] Tactic: 26 time 0.502784 [11/26/2020-16:42:08] [V] [TRT] Tactic: 29 time 0.474592 [11/26/2020-16:42:08] [V] [TRT] Tactic: 30 time 0.491072 [11/26/2020-16:42:08] [V] [TRT] Tactic: 33 time 0.612352 [11/26/2020-16:42:08] [V] [TRT] Tactic: 36 time 0.864256 [11/26/2020-16:42:08] [V] [TRT] Tactic: 39 time 0.379712 [11/26/2020-16:42:08] [V] [TRT] Tactic: 41 time 0.449536 [11/26/2020-16:42:08] [V] [TRT] Tactic: 42 time 0.627712 [11/26/2020-16:42:08] [V] [TRT] Tactic: 43 time 0.611328 [11/26/2020-16:42:08] [V] [TRT] Tactic: 45 time 0.320064 [11/26/2020-16:42:08] [V] [TRT] Tactic: 47 time 0.398336 [11/26/2020-16:42:08] [V] [TRT] Tactic: 52 time 0.388096 [11/26/2020-16:42:08] [V] [TRT] Tactic: 54 time 0.60928 [11/26/2020-16:42:08] [V] [TRT] Tactic: 56 time 0.443136 [11/26/2020-16:42:08] [V] [TRT] Tactic: 66 time 0.690976 [11/26/2020-16:42:08] [V] [TRT] Tactic: 76 time 0.381952 [11/26/2020-16:42:08] [V] [TRT] Tactic: 90 time 0.421856 [11/26/2020-16:42:08] [V] [TRT] Tactic: 93 time 0.388096 [11/26/2020-16:42:08] [V] [TRT] Tactic: 98 time 0.62976 [11/26/2020-16:42:08] [V] [TRT] Tactic: 104 time 0.535552 [11/26/2020-16:42:08] [V] [TRT] Tactic: 110 time 0.453632 [11/26/2020-16:42:08] [V] [TRT] Tactic: 119 time 0.61952 [11/26/2020-16:42:08] [V] [TRT] Tactic: 121 time 0.424416 [11/26/2020-16:42:08] [V] [TRT] Tactic: 130 time 0.571392 [11/26/2020-16:42:08] [V] [TRT] Tactic: 134 time 0.589824 [11/26/2020-16:42:08] [V] [TRT] Tactic: 136 time 0.544288 [11/26/2020-16:42:08] [V] [TRT] Tactic: 137 time 0.572128 [11/26/2020-16:42:08] [V] [TRT] Tactic: 139 time 0.610304 [11/26/2020-16:42:08] [V] [TRT] Tactic: 144 time 0.527744 [11/26/2020-16:42:08] [V] [TRT] Tactic: 149 time 0.386048 [11/26/2020-16:42:08] [V] [TRT] Tactic: 151 time 0.629472 [11/26/2020-16:42:08] [V] [TRT] Tactic: 152 time 0.472064 [11/26/2020-16:42:09] [V] [TRT] Tactic: 153 time 0.696864 [11/26/2020-16:42:09] [V] [TRT] Tactic: 156 time 0.472064 [11/26/2020-16:42:09] [V] [TRT] Tactic: 159 time 0.433152 [11/26/2020-16:42:09] [V] [TRT] Tactic: 162 time 0.49616 [11/26/2020-16:42:09] [V] [TRT] Tactic: 164 time 0.372736 [11/26/2020-16:42:09] [V] [TRT] Fastest Tactic: 14 Time: 0.313344 [11/26/2020-16:42:09] [V] [TRT] --------------- Timing Runner: Conv_56 + Relu_57 (CaskConvolution) [11/26/2020-16:42:09] [V] [TRT] Conv_56 + Relu_57 (scudnn) Set Tactic Name: maxwell_scudnn_128x32_relu_medium_nn_v1 [11/26/2020-16:42:09] [V] [TRT] Tactic: 1062367460111450758 time 0.556032 [11/26/2020-16:42:09] [V] [TRT] Conv_56 + Relu_57 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148n_nt_v1 [11/26/2020-16:42:09] [V] [TRT] Tactic: 3827454225649558724 time 0.195584 [11/26/2020-16:42:09] [V] [TRT] Conv_56 + Relu_57 (scudnn) Set Tactic Name: maxwell_scudnn_128x64_relu_large_nn_v1 [11/26/2020-16:42:09] [V] [TRT] Tactic: 4337000649858996379 time 0.521216 [11/26/2020-16:42:09] [V] [TRT] Conv_56 + Relu_57 (scudnn) Set Tactic Name: maxwell_scudnn_128x128_relu_medium_nn_v1 [11/26/2020-16:42:09] [V] [TRT] Tactic: 4501471010995462441 time 0.488448 [11/26/2020-16:42:09] [V] [TRT] Conv_56 + Relu_57 (scudnn) Set Tactic Name: maxwell_scudnn_128x64_relu_small_nn_v1 [11/26/2020-16:42:09] [V] [TRT] Tactic: 5137655947464784826 time 0.4608 [11/26/2020-16:42:09] [V] [TRT] Conv_56 + Relu_57 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148m_nt_v1 [11/26/2020-16:42:09] [V] [TRT] Tactic: 5921334924264294896 time 0.207616 [11/26/2020-16:42:09] [V] [TRT] Conv_56 + Relu_57 (scudnn) Set Tactic Name: maxwell_scudnn_128x64_relu_medium_nn_v1 [11/26/2020-16:42:09] [V] [TRT] Tactic: 6645123197870846056 time 0.492544 [11/26/2020-16:42:09] [V] [TRT] Conv_56 + Relu_57 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148t_nt_v1 [11/26/2020-16:42:09] [V] [TRT] Tactic: 7852627285308570038 time 0.212992 [11/26/2020-16:42:09] [V] [TRT] Conv_56 + Relu_57 (scudnn) Set Tactic Name: maxwell_scudnn_128x128_relu_large_nn_v1 [11/26/2020-16:42:09] [V] [TRT] Tactic: -9137461792520977713 time 0.510976 [11/26/2020-16:42:09] [V] [TRT] Conv_56 + Relu_57 (scudnn) Set Tactic Name: maxwell_scudnn_128x32_relu_large_nn_v1 [11/26/2020-16:42:09] [V] [TRT] Tactic: -6092040395344634144 time 0.611328 [11/26/2020-16:42:09] [V] [TRT] Conv_56 + Relu_57 (scudnn) Set Tactic Name: maxwell_scudnn_128x32_relu_small_nn_v1 [11/26/2020-16:42:09] [V] [TRT] Tactic: -3456450830548107839 time 0.4904 [11/26/2020-16:42:09] [V] [TRT] Conv_56 + Relu_57 (scudnn) Set Tactic Name: maxwell_scudnn_128x128_relu_small_nn_v1 [11/26/2020-16:42:09] [V] [TRT] Tactic: -410470605513481746 time 0.477184 [11/26/2020-16:42:09] [V] [TRT] Fastest Tactic: 3827454225649558724 Time: 0.195584 [11/26/2020-16:42:09] [V] [TRT] --------------- Timing Runner: Conv_56 + Relu_57 (CudaConvolution) [11/26/2020-16:42:09] [V] [TRT] Tactic: 0 time 0.71984 [11/26/2020-16:42:09] [V] [TRT] Tactic: 1 time 0.548864 [11/26/2020-16:42:09] [V] [TRT] Tactic: 2 time 0.643072 [11/26/2020-16:42:09] [V] [TRT] Tactic: 5 time 15.5453 [11/26/2020-16:42:09] [V] [TRT] Tactic: 6 time 0.2936 [11/26/2020-16:42:09] [V] [TRT] Fastest Tactic: 6 Time: 0.2936 [11/26/2020-16:42:09] [V] [TRT] --------------- Timing Runner: Conv_56 + Relu_57 (CudaDepthwiseConvolution) [11/26/2020-16:42:09] [V] [TRT] CudaDepthwiseConvolution has no valid tactics for this config, skipping [11/26/2020-16:42:09] [V] [TRT] >>>>>>>>>>>>>>> Chose Runner Type: CaskConvolution Tactic: 3827454225649558724 [11/26/2020-16:42:09] [V] [TRT] Conv_56 + Relu_57 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148n_nt_v1 [11/26/2020-16:42:09] [V] [TRT] [11/26/2020-16:42:09] [V] [TRT] *************** Autotuning format combination: Half(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2),(MUL_ADD 8 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 8192)) -> Half(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2),(MUL_ADD 8 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 8192)) *************** [11/26/2020-16:42:09] [V] [TRT] --------------- Timing Runner: Conv_56 + Relu_57 (FusedConvActConvolution) [11/26/2020-16:42:09] [V] [TRT] FusedConvActConvolution has no valid tactics for this config, skipping [11/26/2020-16:42:09] [V] [TRT] --------------- Timing Runner: Conv_56 + Relu_57 (CaskConvolution) [11/26/2020-16:42:09] [V] [TRT] CaskConvolution has no valid tactics for this config, skipping [11/26/2020-16:42:09] [V] [TRT] --------------- Timing Runner: Conv_56 + Relu_57 (CudaConvolution) [11/26/2020-16:42:09] [V] [TRT] Tactic: 0 time 0.777216 [11/26/2020-16:42:09] [V] [TRT] Tactic: 1 time 0.555008 [11/26/2020-16:42:09] [V] [TRT] Tactic: 2 time 0.641024 [11/26/2020-16:42:09] [V] [TRT] Tactic: 5 time 14.2879 [11/26/2020-16:42:09] [V] [TRT] Tactic: 6 time 0.27648 [11/26/2020-16:42:09] [V] [TRT] Fastest Tactic: 6 Time: 0.27648 [11/26/2020-16:42:09] [V] [TRT] --------------- Timing Runner: Conv_56 + Relu_57 (CudaDepthwiseConvolution) [11/26/2020-16:42:09] [V] [TRT] CudaDepthwiseConvolution has no valid tactics for this config, skipping [11/26/2020-16:42:09] [V] [TRT] >>>>>>>>>>>>>>> Chose Runner Type: CudaConvolution Tactic: 6 [11/26/2020-16:42:09] [V] [TRT] [11/26/2020-16:42:09] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:42:09] [V] [TRT] Tactic: 0 time 0.014336 [11/26/2020-16:42:09] [V] [TRT] Fastest Tactic: 0 Time: 0.014336 [11/26/2020-16:42:09] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:42:09] [V] [TRT] Tactic: 0 time 0.011168 [11/26/2020-16:42:09] [V] [TRT] Fastest Tactic: 0 Time: 0.011168 [11/26/2020-16:42:09] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:42:09] [V] [TRT] Tactic: 0 time 0.01488 [11/26/2020-16:42:09] [V] [TRT] Fastest Tactic: 0 Time: 0.01488 [11/26/2020-16:42:09] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:42:09] [V] [TRT] Tactic: 0 time 0.011264 [11/26/2020-16:42:09] [V] [TRT] Fastest Tactic: 0 Time: 0.011264 [11/26/2020-16:42:09] [V] [TRT] *************** Autotuning format combination: Float(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2),(MUL_ADD 8 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 8192)), Float(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2),(MUL_ADD 8 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 8192)) -> Float(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2),(MUL_ADD 8 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 8192)) *************** [11/26/2020-16:42:09] [V] [TRT] --------------- Timing Runner: Conv_58 + Add_59 + Relu_60 (LegacySASSConvolution) [11/26/2020-16:42:09] [V] [TRT] Tactic: 0 time 0.559104 [11/26/2020-16:42:09] [V] [TRT] Tactic: 1 time 0.235424 [11/26/2020-16:42:09] [V] [TRT] Fastest Tactic: 1 Time: 0.235424 [11/26/2020-16:42:09] [V] [TRT] --------------- Timing Runner: Conv_58 + Add_59 + Relu_60 (CaskConvolution) [11/26/2020-16:42:09] [V] [TRT] Conv_58 + Add_59 + Relu_60 (scudnn) Set Tactic Name: maxwell_scudnn_128x32_relu_medium_nn_v1 [11/26/2020-16:42:09] [V] [TRT] Tactic: 1062367460111450758 time 0.55808 [11/26/2020-16:42:09] [V] [TRT] Conv_58 + Add_59 + Relu_60 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148n_nt_v1 [11/26/2020-16:42:09] [V] [TRT] Tactic: 3827454225649558724 time 0.201728 [11/26/2020-16:42:09] [V] [TRT] Conv_58 + Add_59 + Relu_60 (scudnn) Set Tactic Name: maxwell_scudnn_128x64_relu_large_nn_v1 [11/26/2020-16:42:09] [V] [TRT] Tactic: 4337000649858996379 time 0.518144 [11/26/2020-16:42:09] [V] [TRT] Conv_58 + Add_59 + Relu_60 (scudnn) Set Tactic Name: maxwell_scudnn_128x128_relu_medium_nn_v1 [11/26/2020-16:42:09] [V] [TRT] Tactic: 4501471010995462441 time 0.490496 [11/26/2020-16:42:09] [V] [TRT] Conv_58 + Add_59 + Relu_60 (scudnn) Set Tactic Name: maxwell_scudnn_128x64_relu_small_nn_v1 [11/26/2020-16:42:09] [V] [TRT] Tactic: 5137655947464784826 time 0.467968 [11/26/2020-16:42:09] [V] [TRT] Conv_58 + Add_59 + Relu_60 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148m_nt_v1 [11/26/2020-16:42:09] [V] [TRT] Tactic: 5921334924264294896 time 0.20992 [11/26/2020-16:42:09] [V] [TRT] Conv_58 + Add_59 + Relu_60 (scudnn) Set Tactic Name: maxwell_scudnn_128x64_relu_medium_nn_v1 [11/26/2020-16:42:09] [V] [TRT] Tactic: 6645123197870846056 time 0.499712 [11/26/2020-16:42:09] [V] [TRT] Conv_58 + Add_59 + Relu_60 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148t_nt_v1 [11/26/2020-16:42:09] [V] [TRT] Tactic: 7852627285308570038 time 0.218112 [11/26/2020-16:42:09] [V] [TRT] Conv_58 + Add_59 + Relu_60 (scudnn) Set Tactic Name: maxwell_scudnn_128x128_relu_large_nn_v1 [11/26/2020-16:42:09] [V] [TRT] Tactic: -9137461792520977713 time 0.515072 [11/26/2020-16:42:09] [V] [TRT] Conv_58 + Add_59 + Relu_60 (scudnn) Set Tactic Name: maxwell_scudnn_128x32_relu_large_nn_v1 [11/26/2020-16:42:09] [V] [TRT] Tactic: -6092040395344634144 time 0.608256 [11/26/2020-16:42:09] [V] [TRT] Conv_58 + Add_59 + Relu_60 (scudnn) Set Tactic Name: maxwell_scudnn_128x32_relu_small_nn_v1 [11/26/2020-16:42:09] [V] [TRT] Tactic: -3456450830548107839 time 0.50176 [11/26/2020-16:42:09] [V] [TRT] Conv_58 + Add_59 + Relu_60 (scudnn) Set Tactic Name: maxwell_scudnn_128x128_relu_small_nn_v1 [11/26/2020-16:42:09] [V] [TRT] Tactic: -410470605513481746 time 0.479232 [11/26/2020-16:42:09] [V] [TRT] Fastest Tactic: 3827454225649558724 Time: 0.201728 [11/26/2020-16:42:09] [V] [TRT] --------------- Timing Runner: Conv_58 + Add_59 + Relu_60 (CudaConvolution) [11/26/2020-16:42:09] [V] [TRT] Tactic: 0 time 0.728064 [11/26/2020-16:42:09] [V] [TRT] Tactic: 1 time 0.546816 [11/26/2020-16:42:09] [V] [TRT] Tactic: 2 time 0.64976 [11/26/2020-16:42:09] [V] [TRT] Tactic: 5 time 14.0882 [11/26/2020-16:42:09] [V] [TRT] Tactic: 6 time 0.305152 [11/26/2020-16:42:09] [V] [TRT] Fastest Tactic: 6 Time: 0.305152 [11/26/2020-16:42:09] [V] [TRT] >>>>>>>>>>>>>>> Chose Runner Type: CaskConvolution Tactic: 3827454225649558724 [11/26/2020-16:42:09] [V] [TRT] Conv_58 + Add_59 + Relu_60 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148n_nt_v1 [11/26/2020-16:42:09] [V] [TRT] [11/26/2020-16:42:09] [V] [TRT] *************** Autotuning format combination: Half(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2),(MUL_ADD 8 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 8192)), Half(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2),(MUL_ADD 8 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 8192)) -> Half(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2),(MUL_ADD 8 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 8192)) *************** [11/26/2020-16:42:09] [V] [TRT] --------------- Timing Runner: Conv_58 + Add_59 + Relu_60 (CaskConvolution) [11/26/2020-16:42:09] [V] [TRT] CaskConvolution has no valid tactics for this config, skipping [11/26/2020-16:42:09] [V] [TRT] --------------- Timing Runner: Conv_58 + Add_59 + Relu_60 (CudaConvolution) [11/26/2020-16:42:10] [V] [TRT] Tactic: 0 time 0.78448 [11/26/2020-16:42:10] [V] [TRT] Tactic: 1 time 0.570368 [11/26/2020-16:42:10] [V] [TRT] Tactic: 2 time 0.649216 [11/26/2020-16:42:10] [V] [TRT] Tactic: 5 time 15.2177 [11/26/2020-16:42:10] [V] [TRT] Tactic: 6 time 0.279552 [11/26/2020-16:42:10] [V] [TRT] Fastest Tactic: 6 Time: 0.279552 [11/26/2020-16:42:10] [V] [TRT] >>>>>>>>>>>>>>> Chose Runner Type: CudaConvolution Tactic: 6 [11/26/2020-16:42:10] [V] [TRT] [11/26/2020-16:42:10] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:42:10] [V] [TRT] Tactic: 0 time 0.014336 [11/26/2020-16:42:10] [V] [TRT] Fastest Tactic: 0 Time: 0.014336 [11/26/2020-16:42:10] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:42:10] [V] [TRT] Tactic: 0 time 0.011136 [11/26/2020-16:42:10] [V] [TRT] Fastest Tactic: 0 Time: 0.011136 [11/26/2020-16:42:10] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:42:10] [V] [TRT] Tactic: 0 time 0.015168 [11/26/2020-16:42:10] [V] [TRT] Fastest Tactic: 0 Time: 0.015168 [11/26/2020-16:42:10] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:42:10] [V] [TRT] Tactic: 0 time 0.011136 [11/26/2020-16:42:10] [V] [TRT] Fastest Tactic: 0 Time: 0.011136 [11/26/2020-16:42:10] [V] [TRT] *************** Autotuning format combination: Float(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2),(MUL_ADD 8 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 8192)) -> Float(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2),(MUL_ADD 8 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 8192)) *************** [11/26/2020-16:42:10] [V] [TRT] --------------- Timing Runner: Conv_61 + Relu_62 (LegacySASSConvolution) [11/26/2020-16:42:10] [V] [TRT] Tactic: 0 time 0.551936 [11/26/2020-16:42:10] [V] [TRT] Tactic: 1 time 0.232448 [11/26/2020-16:42:10] [V] [TRT] Fastest Tactic: 1 Time: 0.232448 [11/26/2020-16:42:10] [V] [TRT] --------------- Timing Runner: Conv_61 + Relu_62 (FusedConvActConvolution) [11/26/2020-16:42:10] [V] [TRT] Tactic: 7 time 0.567904 [11/26/2020-16:42:10] [V] [TRT] Tactic: 10 time 0.528384 [11/26/2020-16:42:10] [V] [TRT] Tactic: 14 time 0.316416 [11/26/2020-16:42:10] [V] [TRT] Tactic: 15 time 0.346112 [11/26/2020-16:42:10] [V] [TRT] Tactic: 25 time 0.343424 [11/26/2020-16:42:10] [V] [TRT] Tactic: 26 time 0.504736 [11/26/2020-16:42:10] [V] [TRT] Tactic: 29 time 0.47104 [11/26/2020-16:42:10] [V] [TRT] Tactic: 30 time 0.498688 [11/26/2020-16:42:10] [V] [TRT] Tactic: 33 time 0.629568 [11/26/2020-16:42:10] [V] [TRT] Tactic: 36 time 0.866304 [11/26/2020-16:42:10] [V] [TRT] Tactic: 39 time 0.375808 [11/26/2020-16:42:10] [V] [TRT] Tactic: 41 time 0.453632 [11/26/2020-16:42:10] [V] [TRT] Tactic: 42 time 0.623616 [11/26/2020-16:42:10] [V] [TRT] Tactic: 43 time 0.611328 [11/26/2020-16:42:10] [V] [TRT] Tactic: 45 time 0.319488 [11/26/2020-16:42:10] [V] [TRT] Tactic: 47 time 0.398336 [11/26/2020-16:42:10] [V] [TRT] Tactic: 52 time 0.389088 [11/26/2020-16:42:10] [V] [TRT] Tactic: 54 time 0.60912 [11/26/2020-16:42:10] [V] [TRT] Tactic: 56 time 0.437344 [11/26/2020-16:42:10] [V] [TRT] Tactic: 66 time 0.693664 [11/26/2020-16:42:10] [V] [TRT] Tactic: 76 time 0.374784 [11/26/2020-16:42:10] [V] [TRT] Tactic: 90 time 0.422912 [11/26/2020-16:42:10] [V] [TRT] Tactic: 93 time 0.387072 [11/26/2020-16:42:10] [V] [TRT] Tactic: 98 time 0.56832 [11/26/2020-16:42:11] [V] [TRT] Tactic: 104 time 0.538624 [11/26/2020-16:42:11] [V] [TRT] Tactic: 110 time 0.452608 [11/26/2020-16:42:11] [V] [TRT] Tactic: 119 time 0.62464 [11/26/2020-16:42:11] [V] [TRT] Tactic: 121 time 0.412672 [11/26/2020-16:42:11] [V] [TRT] Tactic: 130 time 0.574464 [11/26/2020-16:42:11] [V] [TRT] Tactic: 134 time 0.590848 [11/26/2020-16:42:11] [V] [TRT] Tactic: 136 time 0.544768 [11/26/2020-16:42:11] [V] [TRT] Tactic: 137 time 0.572416 [11/26/2020-16:42:11] [V] [TRT] Tactic: 139 time 0.60928 [11/26/2020-16:42:11] [V] [TRT] Tactic: 144 time 0.534048 [11/26/2020-16:42:11] [V] [TRT] Tactic: 149 time 0.384 [11/26/2020-16:42:11] [V] [TRT] Tactic: 151 time 0.626688 [11/26/2020-16:42:11] [V] [TRT] Tactic: 152 time 0.473088 [11/26/2020-16:42:11] [V] [TRT] Tactic: 153 time 0.693152 [11/26/2020-16:42:11] [V] [TRT] Tactic: 156 time 0.475136 [11/26/2020-16:42:11] [V] [TRT] Tactic: 159 time 0.42496 [11/26/2020-16:42:11] [V] [TRT] Tactic: 162 time 0.515072 [11/26/2020-16:42:11] [V] [TRT] Tactic: 164 time 0.374784 [11/26/2020-16:42:11] [V] [TRT] Fastest Tactic: 14 Time: 0.316416 [11/26/2020-16:42:11] [V] [TRT] --------------- Timing Runner: Conv_61 + Relu_62 (CaskConvolution) [11/26/2020-16:42:11] [V] [TRT] Conv_61 + Relu_62 (scudnn) Set Tactic Name: maxwell_scudnn_128x32_relu_medium_nn_v1 [11/26/2020-16:42:11] [V] [TRT] Tactic: 1062367460111450758 time 0.558976 [11/26/2020-16:42:11] [V] [TRT] Conv_61 + Relu_62 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148n_nt_v1 [11/26/2020-16:42:11] [V] [TRT] Tactic: 3827454225649558724 time 0.196608 [11/26/2020-16:42:11] [V] [TRT] Conv_61 + Relu_62 (scudnn) Set Tactic Name: maxwell_scudnn_128x64_relu_large_nn_v1 [11/26/2020-16:42:11] [V] [TRT] Tactic: 4337000649858996379 time 0.523264 [11/26/2020-16:42:11] [V] [TRT] Conv_61 + Relu_62 (scudnn) Set Tactic Name: maxwell_scudnn_128x128_relu_medium_nn_v1 [11/26/2020-16:42:11] [V] [TRT] Tactic: 4501471010995462441 time 0.4864 [11/26/2020-16:42:11] [V] [TRT] Conv_61 + Relu_62 (scudnn) Set Tactic Name: maxwell_scudnn_128x64_relu_small_nn_v1 [11/26/2020-16:42:11] [V] [TRT] Tactic: 5137655947464784826 time 0.460352 [11/26/2020-16:42:11] [V] [TRT] Conv_61 + Relu_62 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148m_nt_v1 [11/26/2020-16:42:11] [V] [TRT] Tactic: 5921334924264294896 time 0.206848 [11/26/2020-16:42:11] [V] [TRT] Conv_61 + Relu_62 (scudnn) Set Tactic Name: maxwell_scudnn_128x64_relu_medium_nn_v1 [11/26/2020-16:42:11] [V] [TRT] Tactic: 6645123197870846056 time 0.497664 [11/26/2020-16:42:11] [V] [TRT] Conv_61 + Relu_62 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148t_nt_v1 [11/26/2020-16:42:11] [V] [TRT] Tactic: 7852627285308570038 time 0.210944 [11/26/2020-16:42:11] [V] [TRT] Conv_61 + Relu_62 (scudnn) Set Tactic Name: maxwell_scudnn_128x128_relu_large_nn_v1 [11/26/2020-16:42:11] [V] [TRT] Tactic: -9137461792520977713 time 0.526336 [11/26/2020-16:42:11] [V] [TRT] Conv_61 + Relu_62 (scudnn) Set Tactic Name: maxwell_scudnn_128x32_relu_large_nn_v1 [11/26/2020-16:42:11] [V] [TRT] Tactic: -6092040395344634144 time 0.606208 [11/26/2020-16:42:11] [V] [TRT] Conv_61 + Relu_62 (scudnn) Set Tactic Name: maxwell_scudnn_128x32_relu_small_nn_v1 [11/26/2020-16:42:11] [V] [TRT] Tactic: -3456450830548107839 time 0.490496 [11/26/2020-16:42:11] [V] [TRT] Conv_61 + Relu_62 (scudnn) Set Tactic Name: maxwell_scudnn_128x128_relu_small_nn_v1 [11/26/2020-16:42:11] [V] [TRT] Tactic: -410470605513481746 time 0.476704 [11/26/2020-16:42:11] [V] [TRT] Fastest Tactic: 3827454225649558724 Time: 0.196608 [11/26/2020-16:42:11] [V] [TRT] --------------- Timing Runner: Conv_61 + Relu_62 (CudaConvolution) [11/26/2020-16:42:11] [V] [TRT] Tactic: 0 time 0.718816 [11/26/2020-16:42:11] [V] [TRT] Tactic: 1 time 0.549888 [11/26/2020-16:42:11] [V] [TRT] Tactic: 2 time 0.64512 [11/26/2020-16:42:11] [V] [TRT] Tactic: 5 time 15.2125 [11/26/2020-16:42:11] [V] [TRT] Tactic: 6 time 0.295936 [11/26/2020-16:42:11] [V] [TRT] Fastest Tactic: 6 Time: 0.295936 [11/26/2020-16:42:11] [V] [TRT] --------------- Timing Runner: Conv_61 + Relu_62 (CudaDepthwiseConvolution) [11/26/2020-16:42:11] [V] [TRT] CudaDepthwiseConvolution has no valid tactics for this config, skipping [11/26/2020-16:42:11] [V] [TRT] >>>>>>>>>>>>>>> Chose Runner Type: CaskConvolution Tactic: 3827454225649558724 [11/26/2020-16:42:11] [V] [TRT] Conv_61 + Relu_62 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148n_nt_v1 [11/26/2020-16:42:11] [V] [TRT] [11/26/2020-16:42:11] [V] [TRT] *************** Autotuning format combination: Half(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2),(MUL_ADD 8 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 8192)) -> Half(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2),(MUL_ADD 8 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 8192)) *************** [11/26/2020-16:42:11] [V] [TRT] --------------- Timing Runner: Conv_61 + Relu_62 (FusedConvActConvolution) [11/26/2020-16:42:11] [V] [TRT] FusedConvActConvolution has no valid tactics for this config, skipping [11/26/2020-16:42:11] [V] [TRT] --------------- Timing Runner: Conv_61 + Relu_62 (CaskConvolution) [11/26/2020-16:42:11] [V] [TRT] CaskConvolution has no valid tactics for this config, skipping [11/26/2020-16:42:11] [V] [TRT] --------------- Timing Runner: Conv_61 + Relu_62 (CudaConvolution) [11/26/2020-16:42:11] [V] [TRT] Tactic: 0 time 0.780288 [11/26/2020-16:42:11] [V] [TRT] Tactic: 1 time 0.569344 [11/26/2020-16:42:11] [V] [TRT] Tactic: 2 time 0.642048 [11/26/2020-16:42:11] [V] [TRT] Tactic: 5 time 15.2566 [11/26/2020-16:42:11] [V] [TRT] Tactic: 6 time 0.275456 [11/26/2020-16:42:11] [V] [TRT] Fastest Tactic: 6 Time: 0.275456 [11/26/2020-16:42:11] [V] [TRT] --------------- Timing Runner: Conv_61 + Relu_62 (CudaDepthwiseConvolution) [11/26/2020-16:42:11] [V] [TRT] CudaDepthwiseConvolution has no valid tactics for this config, skipping [11/26/2020-16:42:11] [V] [TRT] >>>>>>>>>>>>>>> Chose Runner Type: CudaConvolution Tactic: 6 [11/26/2020-16:42:11] [V] [TRT] [11/26/2020-16:42:11] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:42:11] [V] [TRT] Tactic: 0 time 0.014336 [11/26/2020-16:42:11] [V] [TRT] Fastest Tactic: 0 Time: 0.014336 [11/26/2020-16:42:11] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:42:11] [V] [TRT] Tactic: 0 time 0.012288 [11/26/2020-16:42:11] [V] [TRT] Fastest Tactic: 0 Time: 0.012288 [11/26/2020-16:42:11] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:42:11] [V] [TRT] Tactic: 0 time 0.015136 [11/26/2020-16:42:11] [V] [TRT] Fastest Tactic: 0 Time: 0.015136 [11/26/2020-16:42:11] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:42:11] [V] [TRT] Tactic: 0 time 0.010976 [11/26/2020-16:42:11] [V] [TRT] Fastest Tactic: 0 Time: 0.010976 [11/26/2020-16:42:11] [V] [TRT] *************** Autotuning format combination: Float(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2),(MUL_ADD 8 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 8192)), Float(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2),(MUL_ADD 8 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 8192)) -> Float(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2),(MUL_ADD 8 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 8192)) *************** [11/26/2020-16:42:11] [V] [TRT] --------------- Timing Runner: Conv_63 + Add_64 + Relu_65 (LegacySASSConvolution) [11/26/2020-16:42:11] [V] [TRT] Tactic: 0 time 0.556032 [11/26/2020-16:42:12] [V] [TRT] Tactic: 1 time 0.233408 [11/26/2020-16:42:12] [V] [TRT] Fastest Tactic: 1 Time: 0.233408 [11/26/2020-16:42:12] [V] [TRT] --------------- Timing Runner: Conv_63 + Add_64 + Relu_65 (CaskConvolution) [11/26/2020-16:42:12] [V] [TRT] Conv_63 + Add_64 + Relu_65 (scudnn) Set Tactic Name: maxwell_scudnn_128x32_relu_medium_nn_v1 [11/26/2020-16:42:12] [V] [TRT] Tactic: 1062367460111450758 time 0.564224 [11/26/2020-16:42:12] [V] [TRT] Conv_63 + Add_64 + Relu_65 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148n_nt_v1 [11/26/2020-16:42:12] [V] [TRT] Tactic: 3827454225649558724 time 0.217024 [11/26/2020-16:42:12] [V] [TRT] Conv_63 + Add_64 + Relu_65 (scudnn) Set Tactic Name: maxwell_scudnn_128x64_relu_large_nn_v1 [11/26/2020-16:42:12] [V] [TRT] Tactic: 4337000649858996379 time 0.521216 [11/26/2020-16:42:12] [V] [TRT] Conv_63 + Add_64 + Relu_65 (scudnn) Set Tactic Name: maxwell_scudnn_128x128_relu_medium_nn_v1 [11/26/2020-16:42:12] [V] [TRT] Tactic: 4501471010995462441 time 0.488448 [11/26/2020-16:42:12] [V] [TRT] Conv_63 + Add_64 + Relu_65 (scudnn) Set Tactic Name: maxwell_scudnn_128x64_relu_small_nn_v1 [11/26/2020-16:42:12] [V] [TRT] Tactic: 5137655947464784826 time 0.468992 [11/26/2020-16:42:12] [V] [TRT] Conv_63 + Add_64 + Relu_65 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148m_nt_v1 [11/26/2020-16:42:12] [V] [TRT] Tactic: 5921334924264294896 time 0.210496 [11/26/2020-16:42:12] [V] [TRT] Conv_63 + Add_64 + Relu_65 (scudnn) Set Tactic Name: maxwell_scudnn_128x64_relu_medium_nn_v1 [11/26/2020-16:42:12] [V] [TRT] Tactic: 6645123197870846056 time 0.50688 [11/26/2020-16:42:12] [V] [TRT] Conv_63 + Add_64 + Relu_65 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148t_nt_v1 [11/26/2020-16:42:12] [V] [TRT] Tactic: 7852627285308570038 time 0.216192 [11/26/2020-16:42:12] [V] [TRT] Conv_63 + Add_64 + Relu_65 (scudnn) Set Tactic Name: maxwell_scudnn_128x128_relu_large_nn_v1 [11/26/2020-16:42:12] [V] [TRT] Tactic: -9137461792520977713 time 0.508928 [11/26/2020-16:42:12] [V] [TRT] Conv_63 + Add_64 + Relu_65 (scudnn) Set Tactic Name: maxwell_scudnn_128x32_relu_large_nn_v1 [11/26/2020-16:42:12] [V] [TRT] Tactic: -6092040395344634144 time 0.612352 [11/26/2020-16:42:12] [V] [TRT] Conv_63 + Add_64 + Relu_65 (scudnn) Set Tactic Name: maxwell_scudnn_128x32_relu_small_nn_v1 [11/26/2020-16:42:12] [V] [TRT] Tactic: -3456450830548107839 time 0.502784 [11/26/2020-16:42:12] [V] [TRT] Conv_63 + Add_64 + Relu_65 (scudnn) Set Tactic Name: maxwell_scudnn_128x128_relu_small_nn_v1 [11/26/2020-16:42:12] [V] [TRT] Tactic: -410470605513481746 time 0.479776 [11/26/2020-16:42:12] [V] [TRT] Fastest Tactic: 5921334924264294896 Time: 0.210496 [11/26/2020-16:42:12] [V] [TRT] --------------- Timing Runner: Conv_63 + Add_64 + Relu_65 (CudaConvolution) [11/26/2020-16:42:12] [V] [TRT] Tactic: 0 time 0.730272 [11/26/2020-16:42:12] [V] [TRT] Tactic: 1 time 0.547744 [11/26/2020-16:42:12] [V] [TRT] Tactic: 2 time 0.65024 [11/26/2020-16:42:12] [V] [TRT] Tactic: 5 time 15.3303 [11/26/2020-16:42:12] [V] [TRT] Tactic: 6 time 0.306176 [11/26/2020-16:42:12] [V] [TRT] Fastest Tactic: 6 Time: 0.306176 [11/26/2020-16:42:12] [V] [TRT] >>>>>>>>>>>>>>> Chose Runner Type: CaskConvolution Tactic: 5921334924264294896 [11/26/2020-16:42:12] [V] [TRT] Conv_63 + Add_64 + Relu_65 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148m_nt_v1 [11/26/2020-16:42:12] [V] [TRT] [11/26/2020-16:42:12] [V] [TRT] *************** Autotuning format combination: Half(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2),(MUL_ADD 8 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 8192)), Half(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2),(MUL_ADD 8 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 8192)) -> Half(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2),(MUL_ADD 8 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 8192)) *************** [11/26/2020-16:42:12] [V] [TRT] --------------- Timing Runner: Conv_63 + Add_64 + Relu_65 (CaskConvolution) [11/26/2020-16:42:12] [V] [TRT] CaskConvolution has no valid tactics for this config, skipping [11/26/2020-16:42:12] [V] [TRT] --------------- Timing Runner: Conv_63 + Add_64 + Relu_65 (CudaConvolution) [11/26/2020-16:42:12] [V] [TRT] Tactic: 0 time 0.785408 [11/26/2020-16:42:12] [V] [TRT] Tactic: 1 time 0.575968 [11/26/2020-16:42:12] [V] [TRT] Tactic: 2 time 0.64512 [11/26/2020-16:42:12] [V] [TRT] Tactic: 5 time 15.557 [11/26/2020-16:42:12] [V] [TRT] Tactic: 6 time 0.279552 [11/26/2020-16:42:12] [V] [TRT] Fastest Tactic: 6 Time: 0.279552 [11/26/2020-16:42:12] [V] [TRT] >>>>>>>>>>>>>>> Chose Runner Type: CudaConvolution Tactic: 6 [11/26/2020-16:42:12] [V] [TRT] [11/26/2020-16:42:12] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:42:12] [V] [TRT] Tactic: 0 time 0.01536 [11/26/2020-16:42:12] [V] [TRT] Fastest Tactic: 0 Time: 0.01536 [11/26/2020-16:42:12] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:42:12] [V] [TRT] Tactic: 0 time 0.011104 [11/26/2020-16:42:12] [V] [TRT] Fastest Tactic: 0 Time: 0.011104 [11/26/2020-16:42:12] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:42:12] [V] [TRT] Tactic: 0 time 0.014912 [11/26/2020-16:42:12] [V] [TRT] Fastest Tactic: 0 Time: 0.014912 [11/26/2020-16:42:12] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:42:12] [V] [TRT] Tactic: 0 time 0.0112 [11/26/2020-16:42:12] [V] [TRT] Fastest Tactic: 0 Time: 0.0112 [11/26/2020-16:42:12] [V] [TRT] *************** Autotuning format combination: Float(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2),(MUL_ADD 8 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 8192)) -> Float(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2),(MUL_ADD 8 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 8192)) *************** [11/26/2020-16:42:12] [V] [TRT] --------------- Timing Runner: Conv_66 + Relu_67 (LegacySASSConvolution) [11/26/2020-16:42:12] [V] [TRT] Tactic: 0 time 0.55296 [11/26/2020-16:42:12] [V] [TRT] Tactic: 1 time 0.240672 [11/26/2020-16:42:12] [V] [TRT] Fastest Tactic: 1 Time: 0.240672 [11/26/2020-16:42:12] [V] [TRT] --------------- Timing Runner: Conv_66 + Relu_67 (FusedConvActConvolution) [11/26/2020-16:42:12] [V] [TRT] Tactic: 7 time 0.567232 [11/26/2020-16:42:12] [V] [TRT] Tactic: 10 time 0.5344 [11/26/2020-16:42:12] [V] [TRT] Tactic: 14 time 0.315264 [11/26/2020-16:42:12] [V] [TRT] Tactic: 15 time 0.344064 [11/26/2020-16:42:12] [V] [TRT] Tactic: 25 time 0.351808 [11/26/2020-16:42:12] [V] [TRT] Tactic: 26 time 0.505856 [11/26/2020-16:42:12] [V] [TRT] Tactic: 29 time 0.475264 [11/26/2020-16:42:12] [V] [TRT] Tactic: 30 time 0.492544 [11/26/2020-16:42:12] [V] [TRT] Tactic: 33 time 0.617472 [11/26/2020-16:42:12] [V] [TRT] Tactic: 36 time 0.86528 [11/26/2020-16:42:12] [V] [TRT] Tactic: 39 time 0.377728 [11/26/2020-16:42:12] [V] [TRT] Tactic: 41 time 0.448512 [11/26/2020-16:42:12] [V] [TRT] Tactic: 42 time 0.63488 [11/26/2020-16:42:13] [V] [TRT] Tactic: 43 time 0.616448 [11/26/2020-16:42:13] [V] [TRT] Tactic: 45 time 0.318464 [11/26/2020-16:42:13] [V] [TRT] Tactic: 47 time 0.400384 [11/26/2020-16:42:13] [V] [TRT] Tactic: 52 time 0.392192 [11/26/2020-16:42:13] [V] [TRT] Tactic: 54 time 0.60928 [11/26/2020-16:42:13] [V] [TRT] Tactic: 56 time 0.44032 [11/26/2020-16:42:13] [V] [TRT] Tactic: 66 time 0.69232 [11/26/2020-16:42:13] [V] [TRT] Tactic: 76 time 0.382976 [11/26/2020-16:42:13] [V] [TRT] Tactic: 90 time 0.42448 [11/26/2020-16:42:13] [V] [TRT] Tactic: 93 time 0.387072 [11/26/2020-16:42:13] [V] [TRT] Tactic: 98 time 0.632832 [11/26/2020-16:42:13] [V] [TRT] Tactic: 104 time 0.544768 [11/26/2020-16:42:13] [V] [TRT] Tactic: 110 time 0.452608 [11/26/2020-16:42:13] [V] [TRT] Tactic: 119 time 0.623616 [11/26/2020-16:42:13] [V] [TRT] Tactic: 121 time 0.410624 [11/26/2020-16:42:13] [V] [TRT] Tactic: 130 time 0.572416 [11/26/2020-16:42:13] [V] [TRT] Tactic: 134 time 0.591872 [11/26/2020-16:42:13] [V] [TRT] Tactic: 136 time 0.539648 [11/26/2020-16:42:13] [V] [TRT] Tactic: 137 time 0.579584 [11/26/2020-16:42:13] [V] [TRT] Tactic: 139 time 0.610304 [11/26/2020-16:42:13] [V] [TRT] Tactic: 144 time 0.5376 [11/26/2020-16:42:13] [V] [TRT] Tactic: 149 time 0.382528 [11/26/2020-16:42:13] [V] [TRT] Tactic: 151 time 0.627712 [11/26/2020-16:42:13] [V] [TRT] Tactic: 152 time 0.479232 [11/26/2020-16:42:13] [V] [TRT] Tactic: 153 time 0.697344 [11/26/2020-16:42:13] [V] [TRT] Tactic: 156 time 0.472064 [11/26/2020-16:42:13] [V] [TRT] Tactic: 159 time 0.42496 [11/26/2020-16:42:13] [V] [TRT] Tactic: 162 time 0.497504 [11/26/2020-16:42:13] [V] [TRT] Tactic: 164 time 0.37376 [11/26/2020-16:42:13] [V] [TRT] Fastest Tactic: 14 Time: 0.315264 [11/26/2020-16:42:13] [V] [TRT] --------------- Timing Runner: Conv_66 + Relu_67 (CaskConvolution) [11/26/2020-16:42:13] [V] [TRT] Conv_66 + Relu_67 (scudnn) Set Tactic Name: maxwell_scudnn_128x32_relu_medium_nn_v1 [11/26/2020-16:42:13] [V] [TRT] Tactic: 1062367460111450758 time 0.558752 [11/26/2020-16:42:13] [V] [TRT] Conv_66 + Relu_67 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148n_nt_v1 [11/26/2020-16:42:13] [V] [TRT] Tactic: 3827454225649558724 time 0.197632 [11/26/2020-16:42:13] [V] [TRT] Conv_66 + Relu_67 (scudnn) Set Tactic Name: maxwell_scudnn_128x64_relu_large_nn_v1 [11/26/2020-16:42:13] [V] [TRT] Tactic: 4337000649858996379 time 0.512 [11/26/2020-16:42:13] [V] [TRT] Conv_66 + Relu_67 (scudnn) Set Tactic Name: maxwell_scudnn_128x128_relu_medium_nn_v1 [11/26/2020-16:42:13] [V] [TRT] Tactic: 4501471010995462441 time 0.487328 [11/26/2020-16:42:13] [V] [TRT] Conv_66 + Relu_67 (scudnn) Set Tactic Name: maxwell_scudnn_128x64_relu_small_nn_v1 [11/26/2020-16:42:13] [V] [TRT] Tactic: 5137655947464784826 time 0.461824 [11/26/2020-16:42:13] [V] [TRT] Conv_66 + Relu_67 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148m_nt_v1 [11/26/2020-16:42:13] [V] [TRT] Tactic: 5921334924264294896 time 0.216064 [11/26/2020-16:42:13] [V] [TRT] Conv_66 + Relu_67 (scudnn) Set Tactic Name: maxwell_scudnn_128x64_relu_medium_nn_v1 [11/26/2020-16:42:13] [V] [TRT] Tactic: 6645123197870846056 time 0.503648 [11/26/2020-16:42:13] [V] [TRT] Conv_66 + Relu_67 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148t_nt_v1 [11/26/2020-16:42:13] [V] [TRT] Tactic: 7852627285308570038 time 0.214016 [11/26/2020-16:42:13] [V] [TRT] Conv_66 + Relu_67 (scudnn) Set Tactic Name: maxwell_scudnn_128x128_relu_large_nn_v1 [11/26/2020-16:42:13] [V] [TRT] Tactic: -9137461792520977713 time 0.514048 [11/26/2020-16:42:13] [V] [TRT] Conv_66 + Relu_67 (scudnn) Set Tactic Name: maxwell_scudnn_128x32_relu_large_nn_v1 [11/26/2020-16:42:13] [V] [TRT] Tactic: -6092040395344634144 time 0.602112 [11/26/2020-16:42:13] [V] [TRT] Conv_66 + Relu_67 (scudnn) Set Tactic Name: maxwell_scudnn_128x32_relu_small_nn_v1 [11/26/2020-16:42:13] [V] [TRT] Tactic: -3456450830548107839 time 0.4912 [11/26/2020-16:42:13] [V] [TRT] Conv_66 + Relu_67 (scudnn) Set Tactic Name: maxwell_scudnn_128x128_relu_small_nn_v1 [11/26/2020-16:42:13] [V] [TRT] Tactic: -410470605513481746 time 0.47504 [11/26/2020-16:42:13] [V] [TRT] Fastest Tactic: 3827454225649558724 Time: 0.197632 [11/26/2020-16:42:13] [V] [TRT] --------------- Timing Runner: Conv_66 + Relu_67 (CudaConvolution) [11/26/2020-16:42:13] [V] [TRT] Tactic: 0 time 0.723968 [11/26/2020-16:42:13] [V] [TRT] Tactic: 1 time 0.55296 [11/26/2020-16:42:13] [V] [TRT] Tactic: 2 time 0.646144 [11/26/2020-16:42:13] [V] [TRT] Tactic: 5 time 15.0979 [11/26/2020-16:42:13] [V] [TRT] Tactic: 6 time 0.294912 [11/26/2020-16:42:13] [V] [TRT] Fastest Tactic: 6 Time: 0.294912 [11/26/2020-16:42:13] [V] [TRT] --------------- Timing Runner: Conv_66 + Relu_67 (CudaDepthwiseConvolution) [11/26/2020-16:42:13] [V] [TRT] CudaDepthwiseConvolution has no valid tactics for this config, skipping [11/26/2020-16:42:13] [V] [TRT] >>>>>>>>>>>>>>> Chose Runner Type: CaskConvolution Tactic: 3827454225649558724 [11/26/2020-16:42:13] [V] [TRT] Conv_66 + Relu_67 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148n_nt_v1 [11/26/2020-16:42:13] [V] [TRT] [11/26/2020-16:42:13] [V] [TRT] *************** Autotuning format combination: Half(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2),(MUL_ADD 8 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 8192)) -> Half(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2),(MUL_ADD 8 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 8192)) *************** [11/26/2020-16:42:13] [V] [TRT] --------------- Timing Runner: Conv_66 + Relu_67 (FusedConvActConvolution) [11/26/2020-16:42:13] [V] [TRT] FusedConvActConvolution has no valid tactics for this config, skipping [11/26/2020-16:42:13] [V] [TRT] --------------- Timing Runner: Conv_66 + Relu_67 (CaskConvolution) [11/26/2020-16:42:13] [V] [TRT] CaskConvolution has no valid tactics for this config, skipping [11/26/2020-16:42:13] [V] [TRT] --------------- Timing Runner: Conv_66 + Relu_67 (CudaConvolution) [11/26/2020-16:42:13] [V] [TRT] Tactic: 0 time 0.779264 [11/26/2020-16:42:13] [V] [TRT] Tactic: 1 time 0.5632 [11/26/2020-16:42:13] [V] [TRT] Tactic: 2 time 0.641024 [11/26/2020-16:42:14] [V] [TRT] Tactic: 5 time 15.3139 [11/26/2020-16:42:14] [V] [TRT] Tactic: 6 time 0.27648 [11/26/2020-16:42:14] [V] [TRT] Fastest Tactic: 6 Time: 0.27648 [11/26/2020-16:42:14] [V] [TRT] --------------- Timing Runner: Conv_66 + Relu_67 (CudaDepthwiseConvolution) [11/26/2020-16:42:14] [V] [TRT] CudaDepthwiseConvolution has no valid tactics for this config, skipping [11/26/2020-16:42:14] [V] [TRT] >>>>>>>>>>>>>>> Chose Runner Type: CudaConvolution Tactic: 6 [11/26/2020-16:42:14] [V] [TRT] [11/26/2020-16:42:14] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:42:14] [V] [TRT] Tactic: 0 time 0.017408 [11/26/2020-16:42:14] [V] [TRT] Fastest Tactic: 0 Time: 0.017408 [11/26/2020-16:42:14] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:42:14] [V] [TRT] Tactic: 0 time 0.0112 [11/26/2020-16:42:14] [V] [TRT] Fastest Tactic: 0 Time: 0.0112 [11/26/2020-16:42:14] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:42:14] [V] [TRT] Tactic: 0 time 0.015008 [11/26/2020-16:42:14] [V] [TRT] Fastest Tactic: 0 Time: 0.015008 [11/26/2020-16:42:14] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:42:14] [V] [TRT] Tactic: 0 time 0.010976 [11/26/2020-16:42:14] [V] [TRT] Fastest Tactic: 0 Time: 0.010976 [11/26/2020-16:42:14] [V] [TRT] *************** Autotuning format combination: Float(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2),(MUL_ADD 8 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 8192)), Float(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2),(MUL_ADD 8 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 8192)) -> Float(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2),(MUL_ADD 8 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 8192)) *************** [11/26/2020-16:42:14] [V] [TRT] --------------- Timing Runner: Conv_68 + Add_69 + Relu_70 (LegacySASSConvolution) [11/26/2020-16:42:14] [V] [TRT] Tactic: 0 time 0.555008 [11/26/2020-16:42:14] [V] [TRT] Tactic: 1 time 0.234208 [11/26/2020-16:42:14] [V] [TRT] Fastest Tactic: 1 Time: 0.234208 [11/26/2020-16:42:14] [V] [TRT] --------------- Timing Runner: Conv_68 + Add_69 + Relu_70 (CaskConvolution) [11/26/2020-16:42:14] [V] [TRT] Conv_68 + Add_69 + Relu_70 (scudnn) Set Tactic Name: maxwell_scudnn_128x32_relu_medium_nn_v1 [11/26/2020-16:42:14] [V] [TRT] Tactic: 1062367460111450758 time 0.566272 [11/26/2020-16:42:14] [V] [TRT] Conv_68 + Add_69 + Relu_70 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148n_nt_v1 [11/26/2020-16:42:14] [V] [TRT] Tactic: 3827454225649558724 time 0.22304 [11/26/2020-16:42:14] [V] [TRT] Conv_68 + Add_69 + Relu_70 (scudnn) Set Tactic Name: maxwell_scudnn_128x64_relu_large_nn_v1 [11/26/2020-16:42:14] [V] [TRT] Tactic: 4337000649858996379 time 0.52736 [11/26/2020-16:42:14] [V] [TRT] Conv_68 + Add_69 + Relu_70 (scudnn) Set Tactic Name: maxwell_scudnn_128x128_relu_medium_nn_v1 [11/26/2020-16:42:14] [V] [TRT] Tactic: 4501471010995462441 time 0.493568 [11/26/2020-16:42:14] [V] [TRT] Conv_68 + Add_69 + Relu_70 (scudnn) Set Tactic Name: maxwell_scudnn_128x64_relu_small_nn_v1 [11/26/2020-16:42:14] [V] [TRT] Tactic: 5137655947464784826 time 0.467968 [11/26/2020-16:42:14] [V] [TRT] Conv_68 + Add_69 + Relu_70 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148m_nt_v1 [11/26/2020-16:42:14] [V] [TRT] Tactic: 5921334924264294896 time 0.209856 [11/26/2020-16:42:14] [V] [TRT] Conv_68 + Add_69 + Relu_70 (scudnn) Set Tactic Name: maxwell_scudnn_128x64_relu_medium_nn_v1 [11/26/2020-16:42:14] [V] [TRT] Tactic: 6645123197870846056 time 0.49664 [11/26/2020-16:42:14] [V] [TRT] Conv_68 + Add_69 + Relu_70 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148t_nt_v1 [11/26/2020-16:42:14] [V] [TRT] Tactic: 7852627285308570038 time 0.216064 [11/26/2020-16:42:14] [V] [TRT] Conv_68 + Add_69 + Relu_70 (scudnn) Set Tactic Name: maxwell_scudnn_128x128_relu_large_nn_v1 [11/26/2020-16:42:14] [V] [TRT] Tactic: -9137461792520977713 time 0.515072 [11/26/2020-16:42:14] [V] [TRT] Conv_68 + Add_69 + Relu_70 (scudnn) Set Tactic Name: maxwell_scudnn_128x32_relu_large_nn_v1 [11/26/2020-16:42:14] [V] [TRT] Tactic: -6092040395344634144 time 0.606208 [11/26/2020-16:42:14] [V] [TRT] Conv_68 + Add_69 + Relu_70 (scudnn) Set Tactic Name: maxwell_scudnn_128x32_relu_small_nn_v1 [11/26/2020-16:42:14] [V] [TRT] Tactic: -3456450830548107839 time 0.500736 [11/26/2020-16:42:14] [V] [TRT] Conv_68 + Add_69 + Relu_70 (scudnn) Set Tactic Name: maxwell_scudnn_128x128_relu_small_nn_v1 [11/26/2020-16:42:14] [V] [TRT] Tactic: -410470605513481746 time 0.484352 [11/26/2020-16:42:14] [V] [TRT] Fastest Tactic: 5921334924264294896 Time: 0.209856 [11/26/2020-16:42:14] [V] [TRT] --------------- Timing Runner: Conv_68 + Add_69 + Relu_70 (CudaConvolution) [11/26/2020-16:42:14] [V] [TRT] Tactic: 0 time 0.730112 [11/26/2020-16:42:14] [V] [TRT] Tactic: 1 time 0.565248 [11/26/2020-16:42:14] [V] [TRT] Tactic: 2 time 0.649216 [11/26/2020-16:42:14] [V] [TRT] Tactic: 5 time 15.3805 [11/26/2020-16:42:14] [V] [TRT] Tactic: 6 time 0.3048 [11/26/2020-16:42:14] [V] [TRT] Fastest Tactic: 6 Time: 0.3048 [11/26/2020-16:42:14] [V] [TRT] >>>>>>>>>>>>>>> Chose Runner Type: CaskConvolution Tactic: 5921334924264294896 [11/26/2020-16:42:14] [V] [TRT] Conv_68 + Add_69 + Relu_70 (scudnn_winograd) Set Tactic Name: maxwell_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148m_nt_v1 [11/26/2020-16:42:14] [V] [TRT] [11/26/2020-16:42:14] [V] [TRT] *************** Autotuning format combination: Half(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2),(MUL_ADD 8 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 8192)), Half(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2),(MUL_ADD 8 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 8192)) -> Half(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2),(MUL_ADD 8 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 8192)) *************** [11/26/2020-16:42:14] [V] [TRT] --------------- Timing Runner: Conv_68 + Add_69 + Relu_70 (CaskConvolution) [11/26/2020-16:42:14] [V] [TRT] CaskConvolution has no valid tactics for this config, skipping [11/26/2020-16:42:14] [V] [TRT] --------------- Timing Runner: Conv_68 + Add_69 + Relu_70 (CudaConvolution) [11/26/2020-16:42:14] [V] [TRT] Tactic: 0 time 0.784384 [11/26/2020-16:42:14] [V] [TRT] Tactic: 1 time 0.567296 [11/26/2020-16:42:14] [V] [TRT] Tactic: 2 time 0.646144 [11/26/2020-16:42:14] [V] [TRT] Tactic: 5 time 14.421 [11/26/2020-16:42:14] [V] [TRT] Tactic: 6 time 0.280576 [11/26/2020-16:42:14] [V] [TRT] Fastest Tactic: 6 Time: 0.280576 [11/26/2020-16:42:14] [V] [TRT] >>>>>>>>>>>>>>> Chose Runner Type: CudaConvolution Tactic: 6 [11/26/2020-16:42:14] [V] [TRT] [11/26/2020-16:42:14] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:42:14] [V] [TRT] Tactic: 0 time 0.014336 [11/26/2020-16:42:14] [V] [TRT] Fastest Tactic: 0 Time: 0.014336 [11/26/2020-16:42:14] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:42:14] [V] [TRT] Tactic: 0 time 0.011264 [11/26/2020-16:42:14] [V] [TRT] Fastest Tactic: 0 Time: 0.011264 [11/26/2020-16:42:14] [V] [TRT] *************** Autotuning format combination: Float(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2),(MUL_ADD 8 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 8192)) -> Float(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 3),(MUL_ADD 4 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 12),(MUL_ADD 2048 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 6144)) *************** [11/26/2020-16:42:14] [V] [TRT] --------------- Timing Runner: Conv_71 + Relu_72 (LegacySASSConvolution) [11/26/2020-16:42:14] [V] [TRT] Tactic: 0 time 0.249856 [11/26/2020-16:42:14] [V] [TRT] Fastest Tactic: 0 Time: 0.249856 [11/26/2020-16:42:14] [V] [TRT] --------------- Timing Runner: Conv_71 + Relu_72 (FusedConvActConvolution) [11/26/2020-16:42:14] [V] [TRT] FusedConvActConvolution has no valid tactics for this config, skipping [11/26/2020-16:42:14] [V] [TRT] --------------- Timing Runner: Conv_71 + Relu_72 (CaskConvolution) [11/26/2020-16:42:14] [V] [TRT] Conv_71 + Relu_72 (scudnn) Set Tactic Name: maxwell_scudnn_128x32_relu_medium_nn_v1 [11/26/2020-16:42:14] [V] [TRT] Tactic: 1062367460111450758 time 0.249856 [11/26/2020-16:42:14] [V] [TRT] Conv_71 + Relu_72 (scudnn) Set Tactic Name: maxwell_scudnn_128x128_relu_medium_nn_v1 [11/26/2020-16:42:14] [V] [TRT] Tactic: 4501471010995462441 time 0.211872 [11/26/2020-16:42:14] [V] [TRT] Conv_71 + Relu_72 (scudnn) Set Tactic Name: maxwell_scudnn_128x64_relu_small_nn_v1 [11/26/2020-16:42:14] [V] [TRT] Tactic: 5137655947464784826 time 0.169408 [11/26/2020-16:42:14] [V] [TRT] Conv_71 + Relu_72 (scudnn) Set Tactic Name: maxwell_scudnn_128x64_relu_medium_nn_v1 [11/26/2020-16:42:14] [V] [TRT] Tactic: 6645123197870846056 time 0.185344 [11/26/2020-16:42:14] [V] [TRT] Conv_71 + Relu_72 (scudnn) Set Tactic Name: maxwell_scudnn_128x32_relu_small_nn_v1 [11/26/2020-16:42:14] [V] [TRT] Tactic: -3456450830548107839 time 0.201664 [11/26/2020-16:42:14] [V] [TRT] Conv_71 + Relu_72 (scudnn) Set Tactic Name: maxwell_scudnn_128x128_relu_small_nn_v1 [11/26/2020-16:42:14] [V] [TRT] Tactic: -410470605513481746 time 0.205952 [11/26/2020-16:42:14] [V] [TRT] Fastest Tactic: 5137655947464784826 Time: 0.169408 [11/26/2020-16:42:14] [V] [TRT] --------------- Timing Runner: Conv_71 + Relu_72 (CudaConvolution) [11/26/2020-16:42:14] [V] [TRT] Tactic: 0 time 0.251904 [11/26/2020-16:42:14] [V] [TRT] Tactic: 1 time 0.232448 [11/26/2020-16:42:14] [V] [TRT] Tactic: 2 time 0.247808 [11/26/2020-16:42:14] [V] [TRT] Tactic: 5 time 14.2418 [11/26/2020-16:42:14] [V] [TRT] Fastest Tactic: 1 Time: 0.232448 [11/26/2020-16:42:14] [V] [TRT] --------------- Timing Runner: Conv_71 + Relu_72 (CudaDepthwiseConvolution) [11/26/2020-16:42:14] [V] [TRT] CudaDepthwiseConvolution has no valid tactics for this config, skipping [11/26/2020-16:42:14] [V] [TRT] >>>>>>>>>>>>>>> Chose Runner Type: CaskConvolution Tactic: 5137655947464784826 [11/26/2020-16:42:14] [V] [TRT] Conv_71 + Relu_72 (scudnn) Set Tactic Name: maxwell_scudnn_128x64_relu_small_nn_v1 [11/26/2020-16:42:14] [V] [TRT] [11/26/2020-16:42:14] [V] [TRT] *************** Autotuning format combination: Half(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2),(MUL_ADD 8 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 16),(MUL_ADD 4096 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 8192)) -> Half(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 3),(MUL_ADD 4 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 12),(MUL_ADD 2048 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 6144)) *************** [11/26/2020-16:42:14] [V] [TRT] --------------- Timing Runner: Conv_71 + Relu_72 (FusedConvActConvolution) [11/26/2020-16:42:14] [V] [TRT] FusedConvActConvolution has no valid tactics for this config, skipping [11/26/2020-16:42:14] [V] [TRT] --------------- Timing Runner: Conv_71 + Relu_72 (CaskConvolution) [11/26/2020-16:42:14] [V] [TRT] CaskConvolution has no valid tactics for this config, skipping [11/26/2020-16:42:14] [V] [TRT] --------------- Timing Runner: Conv_71 + Relu_72 (CudaConvolution) [11/26/2020-16:42:14] [V] [TRT] Tactic: 0 time 0.35952 [11/26/2020-16:42:15] [V] [TRT] Tactic: 1 time 0.247808 [11/26/2020-16:42:15] [V] [TRT] Tactic: 2 time 0.344064 [11/26/2020-16:42:15] [V] [TRT] Tactic: 5 time 15.0548 [11/26/2020-16:42:15] [V] [TRT] Fastest Tactic: 1 Time: 0.247808 [11/26/2020-16:42:15] [V] [TRT] --------------- Timing Runner: Conv_71 + Relu_72 (CudaDepthwiseConvolution) [11/26/2020-16:42:15] [V] [TRT] CudaDepthwiseConvolution has no valid tactics for this config, skipping [11/26/2020-16:42:15] [V] [TRT] >>>>>>>>>>>>>>> Chose Runner Type: CudaConvolution Tactic: 1 [11/26/2020-16:42:15] [V] [TRT] [11/26/2020-16:42:15] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:42:15] [V] [TRT] Tactic: 0 time 0.009216 [11/26/2020-16:42:15] [V] [TRT] Fastest Tactic: 0 Time: 0.009216 [11/26/2020-16:42:15] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:42:15] [V] [TRT] Tactic: 0 time 0.007904 [11/26/2020-16:42:15] [V] [TRT] Fastest Tactic: 0 Time: 0.007904 [11/26/2020-16:42:15] [V] [TRT] *************** Autotuning format combination: Float(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 3),(MUL_ADD 4 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 12),(MUL_ADD 2048 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 6144)) -> Float(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2),(MUL_ADD 3 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 6),(MUL_ADD 1536 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 3072)) *************** [11/26/2020-16:42:15] [V] [TRT] --------------- Timing Runner: Conv_73 + Relu_74 (LegacySASSConvolution) [11/26/2020-16:42:15] [V] [TRT] Tactic: 0 time 0.246784 [11/26/2020-16:42:15] [V] [TRT] Fastest Tactic: 0 Time: 0.246784 [11/26/2020-16:42:15] [V] [TRT] --------------- Timing Runner: Conv_73 + Relu_74 (FusedConvActConvolution) [11/26/2020-16:42:15] [V] [TRT] FusedConvActConvolution has no valid tactics for this config, skipping [11/26/2020-16:42:15] [V] [TRT] --------------- Timing Runner: Conv_73 + Relu_74 (CaskConvolution) [11/26/2020-16:42:15] [V] [TRT] Conv_73 + Relu_74 (scudnn) Set Tactic Name: maxwell_scudnn_128x32_relu_medium_nn_v1 [11/26/2020-16:42:15] [V] [TRT] Tactic: 1062367460111450758 time 0.239552 [11/26/2020-16:42:15] [V] [TRT] Conv_73 + Relu_74 (scudnn) Set Tactic Name: maxwell_scudnn_128x128_relu_medium_nn_v1 [11/26/2020-16:42:15] [V] [TRT] Tactic: 4501471010995462441 time 0.21904 [11/26/2020-16:42:15] [V] [TRT] Conv_73 + Relu_74 (scudnn) Set Tactic Name: maxwell_scudnn_128x64_relu_small_nn_v1 [11/26/2020-16:42:15] [V] [TRT] Tactic: 5137655947464784826 time 0.17056 [11/26/2020-16:42:15] [V] [TRT] Conv_73 + Relu_74 (scudnn) Set Tactic Name: maxwell_scudnn_128x128_relu_interior_nn_v1 [11/26/2020-16:42:15] [V] [TRT] Tactic: 5326823351883942011 time 0.211744 [11/26/2020-16:42:15] [V] [TRT] Conv_73 + Relu_74 (scudnn) Set Tactic Name: maxwell_scudnn_128x64_relu_medium_nn_v1 [11/26/2020-16:42:15] [V] [TRT] Tactic: 6645123197870846056 time 0.17408 [11/26/2020-16:42:15] [V] [TRT] Conv_73 + Relu_74 (scudnn) Set Tactic Name: maxwell_scudnn_128x32_relu_interior_nn_v1 [11/26/2020-16:42:15] [V] [TRT] Tactic: -6576203419454146580 time 0.192512 [11/26/2020-16:42:15] [V] [TRT] Conv_73 + Relu_74 (scudnn) Set Tactic Name: maxwell_scudnn_128x32_relu_small_nn_v1 [11/26/2020-16:42:15] [V] [TRT] Tactic: -3456450830548107839 time 0.193536 [11/26/2020-16:42:15] [V] [TRT] Conv_73 + Relu_74 (scudnn) Set Tactic Name: maxwell_scudnn_128x128_relu_small_nn_v1 [11/26/2020-16:42:15] [V] [TRT] Tactic: -410470605513481746 time 0.206848 [11/26/2020-16:42:15] [V] [TRT] Conv_73 + Relu_74 (scudnn) Set Tactic Name: maxwell_scudnn_128x64_relu_interior_nn_v1 [11/26/2020-16:42:15] [V] [TRT] Tactic: -37215280111360163 time 0.166912 [11/26/2020-16:42:15] [V] [TRT] Fastest Tactic: -37215280111360163 Time: 0.166912 [11/26/2020-16:42:15] [V] [TRT] --------------- Timing Runner: Conv_73 + Relu_74 (CudaConvolution) [11/26/2020-16:42:15] [V] [TRT] Tactic: 0 time 0.230368 [11/26/2020-16:42:15] [V] [TRT] Tactic: 1 time 0.24576 [11/26/2020-16:42:15] [V] [TRT] Tactic: 2 time 0.218112 [11/26/2020-16:42:15] [V] [TRT] Tactic: 5 time 14.2715 [11/26/2020-16:42:15] [V] [TRT] Fastest Tactic: 2 Time: 0.218112 [11/26/2020-16:42:15] [V] [TRT] --------------- Timing Runner: Conv_73 + Relu_74 (CudaDepthwiseConvolution) [11/26/2020-16:42:15] [V] [TRT] CudaDepthwiseConvolution has no valid tactics for this config, skipping [11/26/2020-16:42:15] [V] [TRT] >>>>>>>>>>>>>>> Chose Runner Type: CaskConvolution Tactic: -37215280111360163 [11/26/2020-16:42:15] [V] [TRT] Conv_73 + Relu_74 (scudnn) Set Tactic Name: maxwell_scudnn_128x64_relu_interior_nn_v1 [11/26/2020-16:42:15] [V] [TRT] [11/26/2020-16:42:15] [V] [TRT] *************** Autotuning format combination: Half(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 3),(MUL_ADD 4 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 12),(MUL_ADD 2048 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 6144)) -> Half(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2),(MUL_ADD 3 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 6),(MUL_ADD 1536 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 3072)) *************** [11/26/2020-16:42:15] [V] [TRT] --------------- Timing Runner: Conv_73 + Relu_74 (FusedConvActConvolution) [11/26/2020-16:42:15] [V] [TRT] FusedConvActConvolution has no valid tactics for this config, skipping [11/26/2020-16:42:15] [V] [TRT] --------------- Timing Runner: Conv_73 + Relu_74 (CaskConvolution) [11/26/2020-16:42:15] [V] [TRT] CaskConvolution has no valid tactics for this config, skipping [11/26/2020-16:42:15] [V] [TRT] --------------- Timing Runner: Conv_73 + Relu_74 (CudaConvolution) [11/26/2020-16:42:15] [V] [TRT] Tactic: 0 time 0.36352 [11/26/2020-16:42:15] [V] [TRT] Tactic: 1 time 0.234496 [11/26/2020-16:42:15] [V] [TRT] Tactic: 2 time 0.352256 [11/26/2020-16:42:15] [V] [TRT] Tactic: 5 time 15.0702 [11/26/2020-16:42:15] [V] [TRT] Fastest Tactic: 1 Time: 0.234496 [11/26/2020-16:42:15] [V] [TRT] --------------- Timing Runner: Conv_73 + Relu_74 (CudaDepthwiseConvolution) [11/26/2020-16:42:15] [V] [TRT] CudaDepthwiseConvolution has no valid tactics for this config, skipping [11/26/2020-16:42:15] [V] [TRT] >>>>>>>>>>>>>>> Chose Runner Type: CudaConvolution Tactic: 1 [11/26/2020-16:42:15] [V] [TRT] [11/26/2020-16:42:15] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:42:15] [V] [TRT] Tactic: 0 time 0.01024 [11/26/2020-16:42:15] [V] [TRT] Fastest Tactic: 0 Time: 0.01024 [11/26/2020-16:42:15] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:42:15] [V] [TRT] Tactic: 0 time 0.008192 [11/26/2020-16:42:15] [V] [TRT] Fastest Tactic: 0 Time: 0.008192 [11/26/2020-16:42:15] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:42:15] [V] [TRT] Tactic: 0 time 0.007776 [11/26/2020-16:42:15] [V] [TRT] Fastest Tactic: 0 Time: 0.007776 [11/26/2020-16:42:15] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:42:15] [V] [TRT] Tactic: 0 time 0.008 [11/26/2020-16:42:15] [V] [TRT] Fastest Tactic: 0 Time: 0.008 [11/26/2020-16:42:15] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:42:15] [V] [TRT] Tactic: 0 time 0.006784 [11/26/2020-16:42:15] [V] [TRT] Fastest Tactic: 0 Time: 0.006784 [11/26/2020-16:42:15] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:42:15] [V] [TRT] Tactic: 0 time 0.011072 [11/26/2020-16:42:15] [V] [TRT] Fastest Tactic: 0 Time: 0.011072 [11/26/2020-16:42:15] [V] [TRT] *************** Autotuning format combination: Float(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2),(MUL_ADD 3 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 6),(MUL_ADD 1536 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 3072)) -> Float(1,3,1536,(MUL_ADD 1536 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 3072)) *************** [11/26/2020-16:42:15] [V] [TRT] --------------- Timing Runner: Transpose_75 (Shuffle) [11/26/2020-16:42:15] [V] [TRT] Tactic: 0 time 0.007904 [11/26/2020-16:42:15] [V] [TRT] Fastest Tactic: 0 Time: 0.007904 [11/26/2020-16:42:15] [V] [TRT] *************** Autotuning format combination: Float(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2),(MUL_ADD 3 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 6):32,(MUL_ADD 48 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 96)) -> Float(1,3,1536:32,(* 1536 (CEIL_DIV (+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2) 32))) *************** [11/26/2020-16:42:15] [V] [TRT] --------------- Timing Runner: Transpose_75 (Shuffle) [11/26/2020-16:42:15] [V] [TRT] Tactic: 0 time 0.011136 [11/26/2020-16:42:15] [V] [TRT] Fastest Tactic: 0 Time: 0.011136 [11/26/2020-16:42:15] [V] [TRT] *************** Autotuning format combination: Half(1,(+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2),(MUL_ADD 3 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 6),(MUL_ADD 1536 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 3072)) -> Half(1,3,1536,(MUL_ADD 1536 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 3072)) *************** [11/26/2020-16:42:15] [V] [TRT] --------------- Timing Runner: Transpose_75 (Shuffle) [11/26/2020-16:42:15] [V] [TRT] Tactic: 0 time 0.007072 [11/26/2020-16:42:15] [V] [TRT] Fastest Tactic: 0 Time: 0.007072 [11/26/2020-16:42:15] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:42:15] [V] [TRT] Tactic: 0 time 0.008096 [11/26/2020-16:42:15] [V] [TRT] Fastest Tactic: 0 Time: 0.008096 [11/26/2020-16:42:15] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:42:15] [V] [TRT] Tactic: 0 time 0.007872 [11/26/2020-16:42:15] [V] [TRT] Fastest Tactic: 0 Time: 0.007872 [11/26/2020-16:42:15] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:42:15] [V] [TRT] Tactic: 0 time 0.007168 [11/26/2020-16:42:15] [V] [TRT] Fastest Tactic: 0 Time: 0.007168 [11/26/2020-16:42:16] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:42:16] [V] [TRT] Tactic: 0 time 0.00592 [11/26/2020-16:42:16] [V] [TRT] Fastest Tactic: 0 Time: 0.00592 [11/26/2020-16:42:16] [V] [TRT] *************** Autotuning format combination: Float(1,3,1536,(MUL_ADD 1536 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 3072)) -> Float(1,1,512,(MUL_ADD 512 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 1024)) *************** [11/26/2020-16:42:16] [V] [TRT] --------------- Timing Runner: AveragePool_76 (Pooling) [11/26/2020-16:42:16] [V] [TRT] Tactic: -1 time 0.008896 [11/26/2020-16:42:16] [V] [TRT] Fastest Tactic: -1 Time: 0.008896 [11/26/2020-16:42:16] [V] [TRT] --------------- Timing Runner: AveragePool_76 (TiledPooling) [11/26/2020-16:42:16] [V] [TRT] TiledPooling has no valid tactics for this config, skipping [11/26/2020-16:42:16] [V] [TRT] >>>>>>>>>>>>>>> Chose Runner Type: Pooling Tactic: -1 [11/26/2020-16:42:16] [V] [TRT] [11/26/2020-16:42:16] [V] [TRT] *************** Autotuning format combination: Half(1,3,1536,(MUL_ADD 1536 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 3072)) -> Half(1,1,512,(MUL_ADD 512 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 1024)) *************** [11/26/2020-16:42:16] [V] [TRT] --------------- Timing Runner: AveragePool_76 (Pooling) [11/26/2020-16:42:16] [V] [TRT] Tactic: -1 time 0.007968 [11/26/2020-16:42:16] [V] [TRT] Fastest Tactic: -1 Time: 0.007968 [11/26/2020-16:42:16] [V] [TRT] --------------- Timing Runner: AveragePool_76 (TiledPooling) [11/26/2020-16:42:16] [V] [TRT] TiledPooling has no valid tactics for this config, skipping [11/26/2020-16:42:16] [V] [TRT] >>>>>>>>>>>>>>> Chose Runner Type: Pooling Tactic: -1 [11/26/2020-16:42:16] [V] [TRT] [11/26/2020-16:42:16] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:42:16] [V] [TRT] Tactic: 0 time 0.005984 [11/26/2020-16:42:16] [V] [TRT] Fastest Tactic: 0 Time: 0.005984 [11/26/2020-16:42:16] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:42:16] [V] [TRT] Tactic: 0 time 0.004928 [11/26/2020-16:42:16] [V] [TRT] Fastest Tactic: 0 Time: 0.004928 [11/26/2020-16:42:16] [V] [TRT] *************** Autotuning format combination: Float(1,1,512,(MUL_ADD 512 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 1024)) -> Float(1,512,(* 512 (# 0 (SHAPE input)))) *************** [11/26/2020-16:42:16] [V] [TRT] --------------- Timing Runner: Reshape_90 + Transpose_97 (Shuffle) [11/26/2020-16:42:16] [V] [TRT] Tactic: 0 time 0.004992 [11/26/2020-16:42:16] [V] [TRT] Fastest Tactic: 0 Time: 0.004992 [11/26/2020-16:42:16] [V] [TRT] *************** Autotuning format combination: Half(1,1,512,(MUL_ADD 512 (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 1024)) -> Half(1,512,(* 512 (# 0 (SHAPE input)))) *************** [11/26/2020-16:42:16] [V] [TRT] --------------- Timing Runner: Reshape_90 + Transpose_97 (Shuffle) [11/26/2020-16:42:16] [V] [TRT] Tactic: 0 time 0.004096 [11/26/2020-16:42:16] [V] [TRT] Fastest Tactic: 0 Time: 0.004096 [11/26/2020-16:42:16] [V] [TRT] *************** Autotuning format combination: -> Int32(1) *************** [11/26/2020-16:42:16] [V] [TRT] --------------- Timing Runner: (Unnamed Layer* 114) [Shape][HostToDeviceCopy] (ShapeHostToDevice) [11/26/2020-16:42:16] [V] [TRT] Tactic: 0 is the only option, timing skipped [11/26/2020-16:42:16] [V] [TRT] Fastest Tactic: 0 Time: 0 [11/26/2020-16:42:16] [V] [TRT] *************** Autotuning format combination: Int32(1), Int32() -> Int32() *************** [11/26/2020-16:42:16] [V] [TRT] --------------- Timing Runner: (Unnamed Layer* 116) [Gather] (Gather) [11/26/2020-16:42:16] [V] [TRT] Tactic: 1 time 0.004096 [11/26/2020-16:42:16] [V] [TRT] Tactic: 2 time 0.004672 [11/26/2020-16:42:16] [V] [TRT] Tactic: 3 time 0.005728 [11/26/2020-16:42:16] [V] [TRT] Tactic: 4 time 0.004096 [11/26/2020-16:42:16] [V] [TRT] Tactic: 6 time 0.00512 [11/26/2020-16:42:16] [V] [TRT] Fastest Tactic: 1 Time: 0.004096 [11/26/2020-16:42:16] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:42:16] [V] [TRT] Tactic: 0 time 0.003072 [11/26/2020-16:42:16] [V] [TRT] Fastest Tactic: 0 Time: 0.003072 [11/26/2020-16:42:16] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:42:16] [V] [TRT] Tactic: 0 time 0.003072 [11/26/2020-16:42:16] [V] [TRT] Fastest Tactic: 0 Time: 0.003072 [11/26/2020-16:42:16] [V] [TRT] *************** Autotuning format combination: Float(1,1,1) -> Float(1,512,(* 512 (# 0 (SHAPE input)))) *************** [11/26/2020-16:42:16] [V] [TRT] --------------- Timing Runner: (Unnamed Layer* 100) [Slice] (Slice) [11/26/2020-16:42:16] [V] [TRT] Tactic: 0 time 0.004064 [11/26/2020-16:42:16] [V] [TRT] Fastest Tactic: 0 Time: 0.004064 [11/26/2020-16:42:16] [V] [TRT] *************** Autotuning format combination: Half(1,1,1) -> Half(1,512,(* 512 (# 0 (SHAPE input)))) *************** [11/26/2020-16:42:16] [V] [TRT] --------------- Timing Runner: (Unnamed Layer* 100) [Slice] (Slice) [11/26/2020-16:42:16] [V] [TRT] Tactic: 0 time 0.003616 [11/26/2020-16:42:16] [V] [TRT] Fastest Tactic: 0 Time: 0.003616 [11/26/2020-16:42:16] [V] [TRT] *************** Autotuning format combination: -> Int32(1) *************** [11/26/2020-16:42:16] [V] [TRT] --------------- Timing Runner: (Unnamed Layer* 202) [Shape][HostToDeviceCopy] (ShapeHostToDevice) [11/26/2020-16:42:16] [V] [TRT] Tactic: 0 is the only option, timing skipped [11/26/2020-16:42:16] [V] [TRT] Fastest Tactic: 0 Time: 0 [11/26/2020-16:42:16] [V] [TRT] *************** Autotuning format combination: Int32(1), Int32() -> Int32() *************** [11/26/2020-16:42:16] [V] [TRT] --------------- Timing Runner: (Unnamed Layer* 204) [Gather] (Gather) [11/26/2020-16:42:16] [V] [TRT] Tactic: 1 time 0.004928 [11/26/2020-16:42:16] [V] [TRT] Tactic: 2 time 0.004096 [11/26/2020-16:42:16] [V] [TRT] Tactic: 3 time 0.007168 [11/26/2020-16:42:16] [V] [TRT] Tactic: 4 time 0.005216 [11/26/2020-16:42:16] [V] [TRT] Tactic: 6 time 0.004928 [11/26/2020-16:42:16] [V] [TRT] Fastest Tactic: 2 Time: 0.004096 [11/26/2020-16:42:16] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:42:16] [V] [TRT] Tactic: 0 time 0.002976 [11/26/2020-16:42:16] [V] [TRT] Fastest Tactic: 0 Time: 0.002976 [11/26/2020-16:42:16] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:42:17] [V] [TRT] Tactic: 0 time 0.003072 [11/26/2020-16:42:17] [V] [TRT] Fastest Tactic: 0 Time: 0.003072 [11/26/2020-16:42:17] [V] [TRT] *************** Autotuning format combination: Float(1,1,1) -> Float(1,512,(* 512 (# 0 (SHAPE input)))) *************** [11/26/2020-16:42:17] [V] [TRT] --------------- Timing Runner: (Unnamed Layer* 188) [Slice] (Slice) [11/26/2020-16:42:17] [V] [TRT] Tactic: 0 time 0.003904 [11/26/2020-16:42:17] [V] [TRT] Fastest Tactic: 0 Time: 0.003904 [11/26/2020-16:42:17] [V] [TRT] *************** Autotuning format combination: Half(1,1,1) -> Half(1,512,(* 512 (# 0 (SHAPE input)))) *************** [11/26/2020-16:42:17] [V] [TRT] --------------- Timing Runner: (Unnamed Layer* 188) [Slice] (Slice) [11/26/2020-16:42:17] [V] [TRT] Tactic: 0 time 0.003072 [11/26/2020-16:42:17] [V] [TRT] Fastest Tactic: 0 Time: 0.003072 [11/26/2020-16:42:17] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:42:17] [V] [TRT] Tactic: 0 time 0.004768 [11/26/2020-16:42:17] [V] [TRT] Fastest Tactic: 0 Time: 0.004768 [11/26/2020-16:42:17] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:42:17] [V] [TRT] Tactic: 0 time 0.003904 [11/26/2020-16:42:17] [V] [TRT] Fastest Tactic: 0 Time: 0.003904 [11/26/2020-16:42:17] [V] [TRT] --------------- Timing Runner: (Reformat) [11/26/2020-16:42:17] [V] [TRT] Tactic: 0 time 0.004096 [11/26/2020-16:42:17] [V] [TRT] Fastest Tactic: 0 Time: 0.004096 [11/26/2020-16:42:17] [V] [TRT] *************** Autotuning format combination: Float(1,512,(* 512 (# 0 (SHAPE input)))), Int32(), Int32(), Float(1,512,(* 512 (# 0 (SHAPE input)))), Float(1,2048,2048), Int32() -> Float(1,512,(* 512 (# 0 (SHAPE input))),(* 512 (# 0 (SHAPE input)))), Float(1,512,(* 512 (# 0 (SHAPE input))),(* 512 (# 0 (SHAPE input)))) *************** [11/26/2020-16:42:17] [V] [TRT] --------------- Timing Runner: {(Unnamed Layer* 119) [Iterator],(Unnamed Layer* 118) [Iterator],(Unnamed Layer* 117) [TripLimit],(Unnamed Layer* 131) [Shuffle],(Unnamed Layer* 125) [Shuffle],(Unnamed Layer* 134) [Recurrence],(Unnamed Layer* 133) [Recurrence],(Unnamed Layer* 132) [Concatenation],(Unnamed Layer* 136) [Matrix Multiply],(Unnamed Layer* 135) [Matrix Multiply],(Unnamed Layer* 137) [ElementWise],(Unnamed Layer* 138) [ElementWise],(Unnamed Layer* 151) [Slice],(Unnamed Layer* 145) [Slice],(Unnamed Layer* 139) [Slice],(Unnamed Layer* 142) [Slice],(Unnamed Layer* 153) [Activation],(Unnamed Layer* 147) [Activation],(Unnamed Layer* 141) [Activation],(Unnamed Layer* 144) [Activation],(Unnamed Layer* 148) [ElementWise],(Unnamed Layer* 149) [ElementWise],(Unnamed Layer* 150) [ElementWise],(Unnamed Layer* 154) [Activation],(Unnamed Layer* 155) [ElementWise],(Unnamed Layer* 164) [Slice],(Unnamed Layer* 163) [Slice],(Unnamed Layer* 166) [LoopOutput],(Unnamed Layer* 165) [LoopOutput]} (Myelin) [11/26/2020-16:42:17] [W] [TRT] Myelin graph with multiple dynamic values may have poor performance if they differ. Dynamic values are: (# 0 (SHAPE input)) (+ (FLOOR_DIV (+ (FLOOR_DIV (+ (# 3 (SHAPE input)) -2) 2) -1) 2) 2) [11/26/2020-16:42:24] [E] [TRT] ../builder/myelin/codeGenerator.cpp (338) - Myelin Error in compileGraph: 69 (myelinExceededMemBudget : Exceeded mem budget of 4294967296. Need 5338390656 ) [11/26/2020-16:42:24] [E] [TRT] ../builder/myelin/codeGenerator.cpp (338) - Myelin Error in compileGraph: 69 (myelinExceededMemBudget : Exceeded mem budget of 4294967296. Need 5338390656 ) [11/26/2020-16:42:24] [E] Engine creation failed [11/26/2020-16:42:24] [E] Engine set up failed &&&& FAILED TensorRT.trtexec # /home/sai/ncai/TensorRT-7.0.0.11/bin/trtexec --fp16 --explicitBatch --workspace=2048 --onnx=../models/text_recognition/latin.onnx --saveEngine=../models/text_recognition/latin.onnx.trt --minShapes='input':1x1x64x16 --optShapes='input':1x1x64x320 --maxShapes='input':2x1x64x3200 --verbose