2025-03-13 14:08:39,313 file: tool_utils.py func: tool_utils line No: 73 log will be stored in /open_explorer/maptrv2/hb_mapper_makertbin.log 2025-03-13 14:08:39,313 file: hb_mapper.py func: hb_mapper line No: 132 Start hb_mapper.... 2025-03-13 14:08:39,313 file: hb_mapper.py func: hb_mapper line No: 133 hbdk version 3.49.6 2025-03-13 14:08:39,313 file: hb_mapper.py func: hb_mapper line No: 134 horizon_nn version 1.0.6 2025-03-13 14:08:39,313 file: hb_mapper.py func: hb_mapper line No: 135 hb_mapper version 1.23.3 2025-03-13 14:08:39,313 file: hb_mapper_makertbin.py func: hb_mapper_makertbin line No: 526 Start Model Convert.... 2025-03-13 14:08:39,319 file: mapper_conf_parser.py func: mapper_conf_parser line No: 104 validating model_parameters... 2025-03-13 14:08:39,319 file: mapper_conf_parser.py func: mapper_conf_parser line No: 259 Using onnx model file: /open_explorer/maptrv2/a.onnx 2025-03-13 14:08:39,677 file: helper.py func: helper line No: 145 Model input names: ['onnx::Gather_0', 'index.1', 'mask.1', 'idx0', 'onnx::Cast_7'] 2025-03-13 14:08:39,677 file: mapper_conf_parser.py func: mapper_conf_parser line No: 263 Model has 5 inputs according to model file 2025-03-13 14:08:39,677 file: mapper_conf_parser.py func: mapper_conf_parser line No: 1359 Using abs path /open_explorer/maptrv2/model_output 2025-03-13 14:08:39,677 file: mapper_conf_parser.py func: mapper_conf_parser line No: 450 node_dict: {self.node_dict} 2025-03-13 14:08:39,677 file: mapper_conf_parser.py func: mapper_conf_parser line No: 118 validating model_parameters finished 2025-03-13 14:08:39,677 file: mapper_conf_parser.py func: mapper_conf_parser line No: 122 validating input_parameters... 2025-03-13 14:08:39,677 file: mapper_conf_parser.py func: mapper_conf_parser line No: 552 Input shape [1, 1, 6, 3, 480, 800] has length: 6, make sure it is a featuremap input 2025-03-13 14:08:39,678 file: mapper_conf_parser.py func: mapper_conf_parser line No: 552 Input shape [153000, 256] has length: 2, make sure it is a featuremap input 2025-03-13 14:08:39,678 file: mapper_conf_parser.py func: mapper_conf_parser line No: 552 Input shape [153000, 256] has length: 2, make sure it is a featuremap input 2025-03-13 14:08:39,678 file: mapper_conf_parser.py func: mapper_conf_parser line No: 552 Input shape [3] has length: 1, make sure it is a featuremap input 2025-03-13 14:08:39,678 file: mapper_conf_parser.py func: mapper_conf_parser line No: 552 Input shape [1, 6, 22] has length: 3, make sure it is a featuremap input 2025-03-13 14:08:39,678 file: mapper_conf_parser.py func: mapper_conf_parser line No: 134 validating input_parameters finished 2025-03-13 14:08:39,678 file: mapper_conf_parser.py func: mapper_conf_parser line No: 138 validating calibration_parameters... 2025-03-13 14:08:39,678 file: mapper_conf_parser.py func: mapper_conf_parser line No: 956 Parameter calibration_type is skip. cal_data_dir check skipped 2025-03-13 14:08:39,678 file: mapper_conf_parser.py func: mapper_conf_parser line No: 154 validating calibration_parameters finished 2025-03-13 14:08:39,678 file: mapper_conf_parser.py func: mapper_conf_parser line No: 158 validating custom_op... 2025-03-13 14:08:39,678 file: mapper_conf_parser.py func: mapper_conf_parser line No: 1088 custom_op does not exist, skipped 2025-03-13 14:08:39,678 file: mapper_conf_parser.py func: mapper_conf_parser line No: 164 validating custom_op finished 2025-03-13 14:08:39,678 file: mapper_conf_parser.py func: mapper_conf_parser line No: 167 validating compiler_parameters... 2025-03-13 14:08:39,678 file: mapper_conf_parser.py func: mapper_conf_parser line No: 1169 Input node onnx::Gather_0's input_source not set, it will be set to ddr by default 2025-03-13 14:08:39,678 file: mapper_conf_parser.py func: mapper_conf_parser line No: 1169 Input node index.1's input_source not set, it will be set to ddr by default 2025-03-13 14:08:39,678 file: mapper_conf_parser.py func: mapper_conf_parser line No: 1169 Input node mask.1's input_source not set, it will be set to ddr by default 2025-03-13 14:08:39,678 file: mapper_conf_parser.py func: mapper_conf_parser line No: 1169 Input node idx0's input_source not set, it will be set to ddr by default 2025-03-13 14:08:39,678 file: mapper_conf_parser.py func: mapper_conf_parser line No: 1169 Input node onnx::Cast_7's input_source not set, it will be set to ddr by default 2025-03-13 14:08:39,678 file: mapper_conf_parser.py func: mapper_conf_parser line No: 182 validating compiler_parameters finished 2025-03-13 14:08:39,678 file: mapper_conf_parser.py func: mapper_conf_parser line No: 186 validating deprecated parameters... 2025-03-13 14:08:39,678 file: mapper_conf_parser.py func: mapper_conf_parser line No: 192 validating deprecated parameters finished 2025-03-13 14:08:39,678 file: hb_mapper_makertbin.py func: hb_mapper_makertbin line No: 54 Dump config: 2025-03-13 14:08:39,678 file: hb_mapper_makertbin.py func: hb_mapper_makertbin line No: 55 calibration_parameters: calibration_type: skip compiler_parameters: compile_mode: latency debug: true optimize_level: O0 input_parameters: input_layout_rt: NCHW;NCHW;NCHW;NCHW;NCHW; input_layout_train: NCHW;NCHW;NCHW;NCHW;NCHW; input_name: onnx::Gather_0;index.1;mask.1;idx0;onnx::Cast_7; input_shape: 1x1x6x3x480x800;153000x256;153000x256;3;1x6x22 input_type_rt: featuremap;featuremap;featuremap;featuremap;featuremap; input_type_train: featuremap;featuremap;featuremap;featuremap;featuremap; norm_type: no_preprocess;no_preprocess;no_preprocess;no_preprocess;no_preprocess; model_parameters: layer_out_dump: false march: bayes onnx_model: /open_explorer/maptrv2/a.onnx output_model_file_prefix: maptrv2 working_dir: model_output 2025-03-13 14:08:39,679 file: hb_mapper_makertbin.py func: hb_mapper_makertbin line No: 60 input 'onnx::Gather_0' : original model shape: [1, 1, 6, 3, 480, 800] 2025-03-13 14:08:39,680 file: hb_mapper_makertbin.py func: hb_mapper_makertbin line No: 60 input 'index.1' : original model shape: [153000, 256] 2025-03-13 14:08:39,680 file: hb_mapper_makertbin.py func: hb_mapper_makertbin line No: 60 input 'mask.1' : original model shape: [153000, 256] 2025-03-13 14:08:39,680 file: hb_mapper_makertbin.py func: hb_mapper_makertbin line No: 60 input 'idx0' : original model shape: [3] 2025-03-13 14:08:39,680 file: hb_mapper_makertbin.py func: hb_mapper_makertbin line No: 60 input 'onnx::Cast_7' : original model shape: [1, 6, 22] 2025-03-13 14:08:39,680 file: hb_mapper_makertbin.py func: hb_mapper_makertbin line No: 67 can not get nhwc info from shape: [1, 1, 6, 3, 480, 800] 2025-03-13 14:08:39,680 file: hb_mapper_makertbin.py func: hb_mapper_makertbin line No: 67 can not get nhwc info from shape: [153000, 256] 2025-03-13 14:08:39,680 file: hb_mapper_makertbin.py func: hb_mapper_makertbin line No: 67 can not get nhwc info from shape: [153000, 256] 2025-03-13 14:08:39,680 file: hb_mapper_makertbin.py func: hb_mapper_makertbin line No: 67 can not get nhwc info from shape: [3] 2025-03-13 14:08:39,680 file: hb_mapper_makertbin.py func: hb_mapper_makertbin line No: 67 can not get nhwc info from shape: [1, 6, 22] 2025-03-13 14:08:39,680 file: hb_mapper_makertbin.py func: hb_mapper_makertbin line No: 263 The calibration_type you specified is skip, the skip uses max+random data for calibration 2025-03-13 14:08:39,680 file: hb_mapper_makertbin.py func: hb_mapper_makertbin line No: 515 call build params: {'march': 'bayes', 'save_model': True, 'name_prefix': 'maptrv2', 'input_dict': {'onnx::Gather_0': {'input_shape': [1, 1, 6, 3, 480, 800], 'original_input_layout': 'NCHW'}, 'index.1': {'input_shape': [153000, 256], 'original_input_layout': 'NCHW'}, 'mask.1': {'input_shape': [153000, 256], 'original_input_layout': 'NCHW'}, 'idx0': {'input_shape': [3], 'original_input_layout': 'NCHW'}, 'onnx::Cast_7': {'input_shape': [1, 6, 22], 'original_input_layout': 'NCHW'}}, 'cali_dict': {'calibration_type': 'max'}, 'hbdk_dict': {'hbdk_pass_through_params': '--O0 --debug --core-num 1 --fast ', 'input-source': {'onnx::Gather_0': 'ddr', 'index.1': 'ddr', 'mask.1': 'ddr', 'idx0': 'ddr', 'onnx::Cast_7': 'ddr', '_default_value': 'ddr'}}, 'node_dict': {}, 'check_mode': True} 2025-03-13 14:08:39,691 file: build.py func: build line No: 36 Start to Horizon NN Model Convert. 2025-03-13 14:08:39,691 file: model_debug.py func: model_debug line No: 61 Loading horizon_nn debug methods:[] 2025-03-13 14:08:39,691 file: build.py func: build line No: 146 The specified model compilation architecture: bayes. 2025-03-13 14:08:39,691 file: build.py func: build line No: 148 The specified model compilation optimization parameters: []. 2025-03-13 14:08:40,173 file: build.py func: build line No: 36 Start to prepare the onnx model. 2025-03-13 14:08:40,173 file: utils.py func: utils line No: 53 Input ONNX Model Information: ONNX IR version: 6 Opset version: ['ai.onnx v11', 'horizon v1'] Producer: pytorch v1.13.1 Domain: None Model version: None Graph input: ...Gather_0: shape=[1, 1, 6, 3, 480, 800], dtype=FLOAT32 index.1: shape=[153000, 256], dtype=INT64 mask.1: shape=[153000, 256], dtype=BOOL idx0: shape=[3], dtype=INT32 ...::Cast_7: shape=[1, 6, 22], dtype=FLOAT32 Graph output: 6409: shape=[50, 4], dtype=FLOAT32 6282: shape=[50], dtype=FLOAT32 6285: shape=[50], dtype=INT64 6285: shape=[50], dtype=INT64 2025-03-13 14:08:42,051 file: build.py func: build line No: 39 End to prepare the onnx model. 2025-03-13 14:08:42,577 file: build.py func: build line No: 186 Saving model: maptrv2_original_float_model.onnx. 2025-03-13 14:08:42,582 file: build.py func: build line No: 36 Start to optimize the model. 2025-03-13 14:09:07,374 file: build.py func: build line No: 39 End to optimize the model. 2025-03-13 14:09:07,954 file: build.py func: build line No: 186 Saving model: maptrv2_optimized_float_model.onnx. 2025-03-13 14:09:07,954 file: build.py func: build line No: 36 Start to calibrate the model. 2025-03-13 14:09:13,818 file: tool_utils.py func: tool_utils line No: 317 BPU does not support int64 featuremap indices, node name:/pts_bbox_head/transformer/encoder/Gather 2025-03-13 14:09:14,559 file: tool_utils.py func: tool_utils line No: 317 BPU does not support int64 featuremap indices, node name:/Gather_5 2025-03-13 14:09:26,128 file: tool_utils.py func: tool_utils line No: 317 BPU does not support int64 featuremap indices, node name:/pts_bbox_head/transformer/encoder/Gather 2025-03-13 14:09:26,205 file: tool_utils.py func: tool_utils line No: 317 BPU does not support int64 featuremap indices, node name:/Gather_5 2025-03-13 14:09:31,596 file: build.py func: build line No: 39 End to calibrate the model. 2025-03-13 14:09:32,246 file: build.py func: build line No: 186 Saving model: maptrv2_calibrated_model.onnx. 2025-03-13 14:09:32,246 file: build.py func: build line No: 36 Start to quantize the model. 2025-03-13 14:09:59,594 file: tool_utils.py func: tool_utils line No: 317 BPU does not support int64 featuremap indices, node name:/pts_bbox_head/transformer/encoder/Gather 2025-03-13 14:09:59,594 file: tool_utils.py func: tool_utils line No: 317 BPU does not support int64 featuremap indices, node name:/Gather_5 2025-03-13 14:09:59,600 file: tool_utils.py func: tool_utils line No: 317 BPU does not support int64 featuremap indices, node name:/pts_bbox_head/transformer/encoder/Gather 2025-03-13 14:09:59,605 file: tool_utils.py func: tool_utils line No: 317 BPU does not support int64 featuremap indices, node name:/Gather_5 2025-03-13 14:10:01,293 file: tool_utils.py func: tool_utils line No: 317 BPU does not support int64 featuremap indices, node name:/pts_bbox_head/transformer/encoder/Gather 2025-03-13 14:10:01,293 file: build.py func: build line No: 39 End to quantize the model. 2025-03-13 14:10:01,294 file: tool_utils.py func: tool_utils line No: 317 BPU does not support int64 featuremap indices, node name:/Gather_5 2025-03-13 14:10:02,750 file: build.py func: build line No: 186 Saving model: maptrv2_quantized_model.onnx. 2025-03-13 14:10:07,376 file: build.py func: build line No: 36 Start to compile the model with march bayes. 2025-03-13 14:10:10,176 file: hybrid_build.py func: hybrid_build line No: 133 Compile submodel: torch_jit_subgraph_0 2025-03-13 14:10:12,237 file: hbdk_cc.py func: hbdk_cc line No: 115 hbdk-cc parameters:['--O0', '--debug', '--core-num', '1', '--fast', '--input-layout', 'NHWC', '--output-layout', 'NHWC', '--input-source', 'ddr,ddr'] 2025-03-13 14:10:12,237 file: hbdk_cc.py func: hbdk_cc line No: 116 hbdk-cc command used:hbdk-cc -f hbir -m /tmp/tmpuicub4az/torch_jit_subgraph_0.hbir -o /tmp/tmpuicub4az/torch_jit_subgraph_0.hbm --march bayes --progressbar --O0 --debug --core-num 1 --fast --input-layout NHWC --output-layout NHWC --input-source ddr,ddr 2025-03-13 14:10:14,038 file: tool_utils.py func: tool_utils line No: 322 consumed time 1.64801 2025-03-13 14:10:14,386 file: tool_utils.py func: tool_utils line No: 322 FPS=24.75, latency = 242393.7 us, DDR = 298952298 bytes (see torch_jit_subgraph_0.html) 2025-03-13 14:10:14,399 file: hybrid_build.py func: hybrid_build line No: 133 Compile submodel: torch_jit_subgraph_1 2025-03-13 14:10:14,871 file: hbdk_cc.py func: hbdk_cc line No: 115 hbdk-cc parameters:['--O0', '--debug', '--core-num', '1', '--fast', '--input-layout', 'NHWC', '--output-layout', 'NHWC', '--input-source', 'ddr,ddr'] 2025-03-13 14:10:14,872 file: hbdk_cc.py func: hbdk_cc line No: 116 hbdk-cc command used:hbdk-cc -f hbir -m /tmp/tmpuicub4az/torch_jit_subgraph_1.hbir -o /tmp/tmpuicub4az/torch_jit_subgraph_1.hbm --march bayes --progressbar --O0 --debug --core-num 1 --fast --input-layout NHWC --output-layout NHWC --input-source ddr,ddr 2025-03-13 14:10:16,494 file: tool_utils.py func: tool_utils line No: 322 consumed time 1.50355 2025-03-13 14:10:17,183 file: tool_utils.py func: tool_utils line No: 322 FPS=2939.41, latency = 522553.3 us, DDR = 2164879 bytes (see torch_jit_subgraph_1.html) 2025-03-13 14:10:17,284 file: hybrid_build.py func: hybrid_build line No: 133 Compile submodel: torch_jit_subgraph_2 2025-03-13 14:10:17,775 file: hbdk_cc.py func: hbdk_cc line No: 115 hbdk-cc parameters:['--O0', '--debug', '--core-num', '1', '--fast', '--input-layout', 'NHWC', '--output-layout', 'NHWC', '--input-source', 'ddr'] 2025-03-13 14:10:17,776 file: hbdk_cc.py func: hbdk_cc line No: 116 hbdk-cc command used:hbdk-cc -f hbir -m /tmp/tmpuicub4az/torch_jit_subgraph_2.hbir -o /tmp/tmpuicub4az/torch_jit_subgraph_2.hbm --march bayes --progressbar --O0 --debug --core-num 1 --fast --input-layout NHWC --output-layout NHWC --input-source ddr 2025-03-13 14:10:17,905 file: tool_utils.py func: tool_utils line No: 322 consumed time 0.0145381 2025-03-13 14:10:18,050 file: tool_utils.py func: tool_utils line No: 322 FPS=105540.9, latency = 9.5 us, DDR = 21504 bytes (see torch_jit_subgraph_2.html) 2025-03-13 14:10:18,051 file: hybrid_build.py func: hybrid_build line No: 133 Compile submodel: torch_jit_subgraph_3 2025-03-13 14:10:18,897 file: hbdk_cc.py func: hbdk_cc line No: 115 hbdk-cc parameters:['--O0', '--debug', '--core-num', '1', '--fast', '--input-layout', 'NHWC', '--output-layout', 'NHWC', '--input-source', 'ddr'] 2025-03-13 14:10:18,897 file: hbdk_cc.py func: hbdk_cc line No: 116 hbdk-cc command used:hbdk-cc -f hbir -m /tmp/tmpuicub4az/torch_jit_subgraph_3.hbir -o /tmp/tmpuicub4az/torch_jit_subgraph_3.hbm --march bayes --progressbar --O0 --debug --core-num 1 --fast --input-layout NHWC --output-layout NHWC --input-source ddr 2025-03-13 14:10:30,864 file: tool_utils.py func: tool_utils line No: 322 consumed time 11.8475 2025-03-13 14:10:31,440 file: tool_utils.py func: tool_utils line No: 322 FPS=2.37, latency = 422810.4 us, DDR = 3271350528 bytes (see torch_jit_subgraph_3.html) 2025-03-13 14:10:31,474 file: hybrid_build.py func: hybrid_build line No: 133 Compile submodel: torch_jit_subgraph_4 2025-03-13 14:10:32,000 file: hbdk_cc.py func: hbdk_cc line No: 115 hbdk-cc parameters:['--O0', '--debug', '--core-num', '1', '--fast', '--input-layout', 'NHWC', '--output-layout', 'NHWC', '--input-source', 'ddr,ddr,ddr'] 2025-03-13 14:10:32,000 file: hbdk_cc.py func: hbdk_cc line No: 116 hbdk-cc command used:hbdk-cc -f hbir -m /tmp/tmpuicub4az/torch_jit_subgraph_4.hbir -o /tmp/tmpuicub4az/torch_jit_subgraph_4.hbm --march bayes --progressbar --O0 --debug --core-num 1 --fast --input-layout NHWC --output-layout NHWC --input-source ddr,ddr,ddr 2025-03-13 14:10:32,674 file: tool_utils.py func: tool_utils line No: 322 consumed time 0.551385 2025-03-13 14:10:32,834 file: tool_utils.py func: tool_utils line No: 322 FPS=16.04, latency = 62333.8 us, DDR = 201062528 bytes (see torch_jit_subgraph_4.html) 2025-03-13 14:10:32,840 file: hybrid_build.py func: hybrid_build line No: 133 Compile submodel: torch_jit_subgraph_5 2025-03-13 14:10:33,287 file: hbdk_cc.py func: hbdk_cc line No: 115 hbdk-cc parameters:['--O0', '--debug', '--core-num', '1', '--fast', '--input-layout', 'NHWC', '--output-layout', 'NHWC', '--input-source', 'ddr,ddr,ddr'] 2025-03-13 14:10:33,288 file: hbdk_cc.py func: hbdk_cc line No: 116 hbdk-cc command used:hbdk-cc -f hbir -m /tmp/tmpuicub4az/torch_jit_subgraph_5.hbir -o /tmp/tmpuicub4az/torch_jit_subgraph_5.hbm --march bayes --progressbar --O0 --debug --core-num 1 --fast --input-layout NHWC --output-layout NHWC --input-source ddr,ddr,ddr 2025-03-13 14:10:33,931 file: tool_utils.py func: tool_utils line No: 322 consumed time 0.535742 2025-03-13 14:10:34,083 file: tool_utils.py func: tool_utils line No: 322 FPS=5.57, latency = 179562.4 us, DDR = 184008768 bytes (see torch_jit_subgraph_5.html) 2025-03-13 14:10:34,089 file: hybrid_build.py func: hybrid_build line No: 133 Compile submodel: torch_jit_subgraph_6 2025-03-13 14:10:34,518 file: hbdk_cc.py func: hbdk_cc line No: 115 hbdk-cc parameters:['--O0', '--debug', '--core-num', '1', '--fast', '--input-layout', 'NHWC', '--output-layout', 'NHWC', '--input-source', 'ddr'] 2025-03-13 14:10:34,519 file: hbdk_cc.py func: hbdk_cc line No: 116 hbdk-cc command used:hbdk-cc -f hbir -m /tmp/tmpuicub4az/torch_jit_subgraph_6.hbir -o /tmp/tmpuicub4az/torch_jit_subgraph_6.hbm --march bayes --progressbar --O0 --debug --core-num 1 --fast --input-layout NHWC --output-layout NHWC --input-source ddr 2025-03-13 14:10:38,891 file: tool_utils.py func: tool_utils line No: 322 consumed time 4.25528 2025-03-13 14:10:39,078 file: tool_utils.py func: tool_utils line No: 322 FPS=30770.01, latency = 259993.4 us, DDR = 24028 bytes (see torch_jit_subgraph_6.html) 2025-03-13 14:10:39,088 file: hybrid_build.py func: hybrid_build line No: 133 Compile submodel: torch_jit_subgraph_7 2025-03-13 14:10:39,593 file: hbdk_cc.py func: hbdk_cc line No: 115 hbdk-cc parameters:['--O0', '--debug', '--core-num', '1', '--fast', '--input-layout', 'NHWC', '--output-layout', 'NHWC', '--input-source', 'ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr'] 2025-03-13 14:10:39,593 file: hbdk_cc.py func: hbdk_cc line No: 116 hbdk-cc command used:hbdk-cc -f hbir -m /tmp/tmpuicub4az/torch_jit_subgraph_7.hbir -o /tmp/tmpuicub4az/torch_jit_subgraph_7.hbm --march bayes --progressbar --O0 --debug --core-num 1 --fast --input-layout NHWC --output-layout NHWC --input-source ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr 2025-03-13 14:10:39,639 file: tool_utils.py func: tool_utils line No: 317 Can not find the scale for node /pts_bbox_head/transformer/decoder/layers.1/attentions.2/Expand_output_0_/pts_bbox_head/transformer/decoder/layers.1/attentions.2/GatherElements_Cast 2025-03-13 14:10:39,640 file: tool_utils.py func: tool_utils line No: 317 Can not find the scale for node /pts_bbox_head/transformer/decoder/layers.1/attentions.2/Expand_1_output_0_/pts_bbox_head/transformer/decoder/layers.1/attentions.2/GatherElements_1_Cast 2025-03-13 14:10:39,641 file: tool_utils.py func: tool_utils line No: 317 Can not find the scale for node /pts_bbox_head/transformer/decoder/layers.1/attentions.2/Expand_2_output_0_/pts_bbox_head/transformer/decoder/layers.1/attentions.2/GatherElements_2_Cast 2025-03-13 14:10:39,641 file: tool_utils.py func: tool_utils line No: 317 Can not find the scale for node /pts_bbox_head/transformer/decoder/layers.1/attentions.2/Expand_3_output_0_/pts_bbox_head/transformer/decoder/layers.1/attentions.2/GatherElements_3_Cast 2025-03-13 14:10:45,386 file: tool_utils.py func: tool_utils line No: 322 consumed time 5.67275 2025-03-13 14:10:45,612 file: tool_utils.py func: tool_utils line No: 322 FPS=7.69, latency = 130093.7 us, DDR = 978355680 bytes (see torch_jit_subgraph_7.html) 2025-03-13 14:10:45,623 file: hybrid_build.py func: hybrid_build line No: 133 Compile submodel: torch_jit_subgraph_8 2025-03-13 14:10:46,097 file: hbdk_cc.py func: hbdk_cc line No: 115 hbdk-cc parameters:['--O0', '--debug', '--core-num', '1', '--fast', '--input-layout', 'NHWC', '--output-layout', 'NHWC', '--input-source', 'ddr,ddr,ddr'] 2025-03-13 14:10:46,097 file: hbdk_cc.py func: hbdk_cc line No: 116 hbdk-cc command used:hbdk-cc -f hbir -m /tmp/tmpuicub4az/torch_jit_subgraph_8.hbir -o /tmp/tmpuicub4az/torch_jit_subgraph_8.hbm --march bayes --progressbar --O0 --debug --core-num 1 --fast --input-layout NHWC --output-layout NHWC --input-source ddr,ddr,ddr 2025-03-13 14:10:46,802 file: tool_utils.py func: tool_utils line No: 322 consumed time 0.58772 2025-03-13 14:10:46,962 file: tool_utils.py func: tool_utils line No: 322 FPS=16.04, latency = 62333.8 us, DDR = 201062528 bytes (see torch_jit_subgraph_8.html) 2025-03-13 14:10:46,968 file: hybrid_build.py func: hybrid_build line No: 133 Compile submodel: torch_jit_subgraph_9 2025-03-13 14:10:47,470 file: hbdk_cc.py func: hbdk_cc line No: 115 hbdk-cc parameters:['--O0', '--debug', '--core-num', '1', '--fast', '--input-layout', 'NHWC', '--output-layout', 'NHWC', '--input-source', 'ddr,ddr,ddr'] 2025-03-13 14:10:47,470 file: hbdk_cc.py func: hbdk_cc line No: 116 hbdk-cc command used:hbdk-cc -f hbir -m /tmp/tmpuicub4az/torch_jit_subgraph_9.hbir -o /tmp/tmpuicub4az/torch_jit_subgraph_9.hbm --march bayes --progressbar --O0 --debug --core-num 1 --fast --input-layout NHWC --output-layout NHWC --input-source ddr,ddr,ddr 2025-03-13 14:10:48,222 file: tool_utils.py func: tool_utils line No: 322 consumed time 0.614577 2025-03-13 14:10:48,386 file: tool_utils.py func: tool_utils line No: 322 FPS=5.57, latency = 179562.4 us, DDR = 184008768 bytes (see torch_jit_subgraph_9.html) 2025-03-13 14:10:48,391 file: hybrid_build.py func: hybrid_build line No: 133 Compile submodel: torch_jit_subgraph_10 2025-03-13 14:10:48,857 file: hbdk_cc.py func: hbdk_cc line No: 115 hbdk-cc parameters:['--O0', '--debug', '--core-num', '1', '--fast', '--input-layout', 'NHWC', '--output-layout', 'NHWC', '--input-source', 'ddr'] 2025-03-13 14:10:48,857 file: hbdk_cc.py func: hbdk_cc line No: 116 hbdk-cc command used:hbdk-cc -f hbir -m /tmp/tmpuicub4az/torch_jit_subgraph_10.hbir -o /tmp/tmpuicub4az/torch_jit_subgraph_10.hbm --march bayes --progressbar --O0 --debug --core-num 1 --fast --input-layout NHWC --output-layout NHWC --input-source ddr 2025-03-13 14:10:54,044 file: tool_utils.py func: tool_utils line No: 322 consumed time 5.07334 2025-03-13 14:10:54,262 file: tool_utils.py func: tool_utils line No: 322 FPS=30770.01, latency = 259993.4 us, DDR = 24028 bytes (see torch_jit_subgraph_10.html) 2025-03-13 14:10:54,275 file: hybrid_build.py func: hybrid_build line No: 133 Compile submodel: torch_jit_subgraph_11 2025-03-13 14:10:54,807 file: hbdk_cc.py func: hbdk_cc line No: 115 hbdk-cc parameters:['--O0', '--debug', '--core-num', '1', '--fast', '--input-layout', 'NHWC', '--output-layout', 'NHWC', '--input-source', 'ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr'] 2025-03-13 14:10:54,807 file: hbdk_cc.py func: hbdk_cc line No: 116 hbdk-cc command used:hbdk-cc -f hbir -m /tmp/tmpuicub4az/torch_jit_subgraph_11.hbir -o /tmp/tmpuicub4az/torch_jit_subgraph_11.hbm --march bayes --progressbar --O0 --debug --core-num 1 --fast --input-layout NHWC --output-layout NHWC --input-source ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr 2025-03-13 14:10:54,859 file: tool_utils.py func: tool_utils line No: 317 Can not find the scale for node /pts_bbox_head/transformer/decoder/layers.2/attentions.2/Expand_output_0_/pts_bbox_head/transformer/decoder/layers.2/attentions.2/GatherElements_Cast 2025-03-13 14:10:54,860 file: tool_utils.py func: tool_utils line No: 317 Can not find the scale for node /pts_bbox_head/transformer/decoder/layers.2/attentions.2/Expand_1_output_0_/pts_bbox_head/transformer/decoder/layers.2/attentions.2/GatherElements_1_Cast 2025-03-13 14:10:54,860 file: tool_utils.py func: tool_utils line No: 317 Can not find the scale for node /pts_bbox_head/transformer/decoder/layers.2/attentions.2/Expand_2_output_0_/pts_bbox_head/transformer/decoder/layers.2/attentions.2/GatherElements_2_Cast 2025-03-13 14:10:54,860 file: tool_utils.py func: tool_utils line No: 317 Can not find the scale for node /pts_bbox_head/transformer/decoder/layers.2/attentions.2/Expand_3_output_0_/pts_bbox_head/transformer/decoder/layers.2/attentions.2/GatherElements_3_Cast 2025-03-13 14:11:01,025 file: tool_utils.py func: tool_utils line No: 322 consumed time 6.09291 2025-03-13 14:11:01,283 file: tool_utils.py func: tool_utils line No: 322 FPS=7.69, latency = 130093.7 us, DDR = 978355680 bytes (see torch_jit_subgraph_11.html) 2025-03-13 14:11:01,296 file: hybrid_build.py func: hybrid_build line No: 133 Compile submodel: torch_jit_subgraph_12 2025-03-13 14:11:01,793 file: hbdk_cc.py func: hbdk_cc line No: 115 hbdk-cc parameters:['--O0', '--debug', '--core-num', '1', '--fast', '--input-layout', 'NHWC', '--output-layout', 'NHWC', '--input-source', 'ddr,ddr,ddr'] 2025-03-13 14:11:01,793 file: hbdk_cc.py func: hbdk_cc line No: 116 hbdk-cc command used:hbdk-cc -f hbir -m /tmp/tmpuicub4az/torch_jit_subgraph_12.hbir -o /tmp/tmpuicub4az/torch_jit_subgraph_12.hbm --march bayes --progressbar --O0 --debug --core-num 1 --fast --input-layout NHWC --output-layout NHWC --input-source ddr,ddr,ddr 2025-03-13 14:11:02,520 file: tool_utils.py func: tool_utils line No: 322 consumed time 0.614232 2025-03-13 14:11:02,703 file: tool_utils.py func: tool_utils line No: 322 FPS=16.04, latency = 62333.8 us, DDR = 201062528 bytes (see torch_jit_subgraph_12.html) 2025-03-13 14:11:02,711 file: hybrid_build.py func: hybrid_build line No: 133 Compile submodel: torch_jit_subgraph_13 2025-03-13 14:11:03,206 file: hbdk_cc.py func: hbdk_cc line No: 115 hbdk-cc parameters:['--O0', '--debug', '--core-num', '1', '--fast', '--input-layout', 'NHWC', '--output-layout', 'NHWC', '--input-source', 'ddr,ddr,ddr'] 2025-03-13 14:11:03,206 file: hbdk_cc.py func: hbdk_cc line No: 116 hbdk-cc command used:hbdk-cc -f hbir -m /tmp/tmpuicub4az/torch_jit_subgraph_13.hbir -o /tmp/tmpuicub4az/torch_jit_subgraph_13.hbm --march bayes --progressbar --O0 --debug --core-num 1 --fast --input-layout NHWC --output-layout NHWC --input-source ddr,ddr,ddr 2025-03-13 14:11:03,885 file: tool_utils.py func: tool_utils line No: 322 consumed time 0.548405 2025-03-13 14:11:04,054 file: tool_utils.py func: tool_utils line No: 322 FPS=5.57, latency = 179562.4 us, DDR = 184008768 bytes (see torch_jit_subgraph_13.html) 2025-03-13 14:11:04,061 file: hybrid_build.py func: hybrid_build line No: 133 Compile submodel: torch_jit_subgraph_14 2025-03-13 14:11:04,564 file: hbdk_cc.py func: hbdk_cc line No: 115 hbdk-cc parameters:['--O0', '--debug', '--core-num', '1', '--fast', '--input-layout', 'NHWC', '--output-layout', 'NHWC', '--input-source', 'ddr'] 2025-03-13 14:11:04,564 file: hbdk_cc.py func: hbdk_cc line No: 116 hbdk-cc command used:hbdk-cc -f hbir -m /tmp/tmpuicub4az/torch_jit_subgraph_14.hbir -o /tmp/tmpuicub4az/torch_jit_subgraph_14.hbm --march bayes --progressbar --O0 --debug --core-num 1 --fast --input-layout NHWC --output-layout NHWC --input-source ddr 2025-03-13 14:11:09,233 file: tool_utils.py func: tool_utils line No: 322 consumed time 4.54865 2025-03-13 14:11:09,417 file: tool_utils.py func: tool_utils line No: 322 FPS=30770.01, latency = 259993.4 us, DDR = 24028 bytes (see torch_jit_subgraph_14.html) 2025-03-13 14:11:09,428 file: hybrid_build.py func: hybrid_build line No: 133 Compile submodel: torch_jit_subgraph_15 2025-03-13 14:11:09,902 file: hbdk_cc.py func: hbdk_cc line No: 115 hbdk-cc parameters:['--O0', '--debug', '--core-num', '1', '--fast', '--input-layout', 'NHWC', '--output-layout', 'NHWC', '--input-source', 'ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr'] 2025-03-13 14:11:09,902 file: hbdk_cc.py func: hbdk_cc line No: 116 hbdk-cc command used:hbdk-cc -f hbir -m /tmp/tmpuicub4az/torch_jit_subgraph_15.hbir -o /tmp/tmpuicub4az/torch_jit_subgraph_15.hbm --march bayes --progressbar --O0 --debug --core-num 1 --fast --input-layout NHWC --output-layout NHWC --input-source ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr 2025-03-13 14:11:09,946 file: tool_utils.py func: tool_utils line No: 317 Can not find the scale for node /pts_bbox_head/transformer/decoder/layers.3/attentions.2/Expand_output_0_/pts_bbox_head/transformer/decoder/layers.3/attentions.2/GatherElements_Cast 2025-03-13 14:11:09,947 file: tool_utils.py func: tool_utils line No: 317 Can not find the scale for node /pts_bbox_head/transformer/decoder/layers.3/attentions.2/Expand_1_output_0_/pts_bbox_head/transformer/decoder/layers.3/attentions.2/GatherElements_1_Cast 2025-03-13 14:11:09,948 file: tool_utils.py func: tool_utils line No: 317 Can not find the scale for node /pts_bbox_head/transformer/decoder/layers.3/attentions.2/Expand_2_output_0_/pts_bbox_head/transformer/decoder/layers.3/attentions.2/GatherElements_2_Cast 2025-03-13 14:11:09,948 file: tool_utils.py func: tool_utils line No: 317 Can not find the scale for node /pts_bbox_head/transformer/decoder/layers.3/attentions.2/Expand_3_output_0_/pts_bbox_head/transformer/decoder/layers.3/attentions.2/GatherElements_3_Cast 2025-03-13 14:11:15,807 file: tool_utils.py func: tool_utils line No: 322 consumed time 5.78086 2025-03-13 14:11:16,050 file: tool_utils.py func: tool_utils line No: 322 FPS=7.69, latency = 130093.7 us, DDR = 978355680 bytes (see torch_jit_subgraph_15.html) 2025-03-13 14:11:16,068 file: hybrid_build.py func: hybrid_build line No: 133 Compile submodel: torch_jit_subgraph_16 2025-03-13 14:11:16,570 file: hbdk_cc.py func: hbdk_cc line No: 115 hbdk-cc parameters:['--O0', '--debug', '--core-num', '1', '--fast', '--input-layout', 'NHWC', '--output-layout', 'NHWC', '--input-source', 'ddr,ddr,ddr'] 2025-03-13 14:11:16,570 file: hbdk_cc.py func: hbdk_cc line No: 116 hbdk-cc command used:hbdk-cc -f hbir -m /tmp/tmpuicub4az/torch_jit_subgraph_16.hbir -o /tmp/tmpuicub4az/torch_jit_subgraph_16.hbm --march bayes --progressbar --O0 --debug --core-num 1 --fast --input-layout NHWC --output-layout NHWC --input-source ddr,ddr,ddr 2025-03-13 14:11:17,301 file: tool_utils.py func: tool_utils line No: 322 consumed time 0.614036 2025-03-13 14:11:17,469 file: tool_utils.py func: tool_utils line No: 322 FPS=16.04, latency = 62333.8 us, DDR = 201062528 bytes (see torch_jit_subgraph_16.html) 2025-03-13 14:11:17,476 file: hybrid_build.py func: hybrid_build line No: 133 Compile submodel: torch_jit_subgraph_17 2025-03-13 14:11:17,947 file: hbdk_cc.py func: hbdk_cc line No: 115 hbdk-cc parameters:['--O0', '--debug', '--core-num', '1', '--fast', '--input-layout', 'NHWC', '--output-layout', 'NHWC', '--input-source', 'ddr,ddr,ddr'] 2025-03-13 14:11:17,947 file: hbdk_cc.py func: hbdk_cc line No: 116 hbdk-cc command used:hbdk-cc -f hbir -m /tmp/tmpuicub4az/torch_jit_subgraph_17.hbir -o /tmp/tmpuicub4az/torch_jit_subgraph_17.hbm --march bayes --progressbar --O0 --debug --core-num 1 --fast --input-layout NHWC --output-layout NHWC --input-source ddr,ddr,ddr 2025-03-13 14:11:18,600 file: tool_utils.py func: tool_utils line No: 322 consumed time 0.529916 2025-03-13 14:11:18,755 file: tool_utils.py func: tool_utils line No: 322 FPS=5.57, latency = 179562.4 us, DDR = 184008768 bytes (see torch_jit_subgraph_17.html) 2025-03-13 14:11:18,761 file: hybrid_build.py func: hybrid_build line No: 133 Compile submodel: torch_jit_subgraph_18 2025-03-13 14:11:19,236 file: hbdk_cc.py func: hbdk_cc line No: 115 hbdk-cc parameters:['--O0', '--debug', '--core-num', '1', '--fast', '--input-layout', 'NHWC', '--output-layout', 'NHWC', '--input-source', 'ddr'] 2025-03-13 14:11:19,236 file: hbdk_cc.py func: hbdk_cc line No: 116 hbdk-cc command used:hbdk-cc -f hbir -m /tmp/tmpuicub4az/torch_jit_subgraph_18.hbir -o /tmp/tmpuicub4az/torch_jit_subgraph_18.hbm --march bayes --progressbar --O0 --debug --core-num 1 --fast --input-layout NHWC --output-layout NHWC --input-source ddr 2025-03-13 14:11:23,957 file: tool_utils.py func: tool_utils line No: 322 consumed time 4.60163 2025-03-13 14:11:24,165 file: tool_utils.py func: tool_utils line No: 322 FPS=30770.01, latency = 259993.4 us, DDR = 24028 bytes (see torch_jit_subgraph_18.html) 2025-03-13 14:11:24,177 file: hybrid_build.py func: hybrid_build line No: 133 Compile submodel: torch_jit_subgraph_19 2025-03-13 14:11:24,724 file: tool_utils.py func: tool_utils line No: 317 Can not find the scale for node /pts_bbox_head/transformer/decoder/layers.4/attentions.2/Expand_output_0_/pts_bbox_head/transformer/decoder/layers.4/attentions.2/GatherElements_Cast 2025-03-13 14:11:24,733 file: tool_utils.py func: tool_utils line No: 317 Can not find the scale for node /pts_bbox_head/transformer/decoder/layers.4/attentions.2/Expand_1_output_0_/pts_bbox_head/transformer/decoder/layers.4/attentions.2/GatherElements_1_Cast 2025-03-13 14:11:24,734 file: hbdk_cc.py func: hbdk_cc line No: 115 hbdk-cc parameters:['--O0', '--debug', '--core-num', '1', '--fast', '--input-layout', 'NHWC', '--output-layout', 'NHWC', '--input-source', 'ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr'] 2025-03-13 14:11:24,734 file: hbdk_cc.py func: hbdk_cc line No: 116 hbdk-cc command used:hbdk-cc -f hbir -m /tmp/tmpuicub4az/torch_jit_subgraph_19.hbir -o /tmp/tmpuicub4az/torch_jit_subgraph_19.hbm --march bayes --progressbar --O0 --debug --core-num 1 --fast --input-layout NHWC --output-layout NHWC --input-source ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr 2025-03-13 14:11:24,785 file: tool_utils.py func: tool_utils line No: 317 Can not find the scale for node /pts_bbox_head/transformer/decoder/layers.4/attentions.2/Expand_2_output_0_/pts_bbox_head/transformer/decoder/layers.4/attentions.2/GatherElements_2_Cast 2025-03-13 14:11:24,785 file: tool_utils.py func: tool_utils line No: 317 Can not find the scale for node /pts_bbox_head/transformer/decoder/layers.4/attentions.2/Expand_3_output_0_/pts_bbox_head/transformer/decoder/layers.4/attentions.2/GatherElements_3_Cast 2025-03-13 14:11:30,407 file: tool_utils.py func: tool_utils line No: 322 consumed time 5.54607 2025-03-13 14:11:30,650 file: tool_utils.py func: tool_utils line No: 322 FPS=7.69, latency = 130093.7 us, DDR = 978355680 bytes (see torch_jit_subgraph_19.html) 2025-03-13 14:11:30,664 file: hybrid_build.py func: hybrid_build line No: 133 Compile submodel: torch_jit_subgraph_20 2025-03-13 14:11:31,182 file: hbdk_cc.py func: hbdk_cc line No: 115 hbdk-cc parameters:['--O0', '--debug', '--core-num', '1', '--fast', '--input-layout', 'NHWC', '--output-layout', 'NHWC', '--input-source', 'ddr,ddr,ddr'] 2025-03-13 14:11:31,183 file: hbdk_cc.py func: hbdk_cc line No: 116 hbdk-cc command used:hbdk-cc -f hbir -m /tmp/tmpuicub4az/torch_jit_subgraph_20.hbir -o /tmp/tmpuicub4az/torch_jit_subgraph_20.hbm --march bayes --progressbar --O0 --debug --core-num 1 --fast --input-layout NHWC --output-layout NHWC --input-source ddr,ddr,ddr 2025-03-13 14:11:31,937 file: tool_utils.py func: tool_utils line No: 322 consumed time 0.629053 2025-03-13 14:11:32,095 file: tool_utils.py func: tool_utils line No: 322 FPS=16.04, latency = 62333.8 us, DDR = 201062528 bytes (see torch_jit_subgraph_20.html) 2025-03-13 14:11:32,102 file: hybrid_build.py func: hybrid_build line No: 133 Compile submodel: torch_jit_subgraph_21 2025-03-13 14:11:32,543 file: hbdk_cc.py func: hbdk_cc line No: 115 hbdk-cc parameters:['--O0', '--debug', '--core-num', '1', '--fast', '--input-layout', 'NHWC', '--output-layout', 'NHWC', '--input-source', 'ddr,ddr,ddr'] 2025-03-13 14:11:32,543 file: hbdk_cc.py func: hbdk_cc line No: 116 hbdk-cc command used:hbdk-cc -f hbir -m /tmp/tmpuicub4az/torch_jit_subgraph_21.hbir -o /tmp/tmpuicub4az/torch_jit_subgraph_21.hbm --march bayes --progressbar --O0 --debug --core-num 1 --fast --input-layout NHWC --output-layout NHWC --input-source ddr,ddr,ddr 2025-03-13 14:11:33,181 file: tool_utils.py func: tool_utils line No: 322 consumed time 0.530463 2025-03-13 14:11:33,337 file: tool_utils.py func: tool_utils line No: 322 FPS=5.57, latency = 179562.4 us, DDR = 184008768 bytes (see torch_jit_subgraph_21.html) 2025-03-13 14:11:33,343 file: hybrid_build.py func: hybrid_build line No: 133 Compile submodel: torch_jit_subgraph_22 2025-03-13 14:11:33,849 file: hbdk_cc.py func: hbdk_cc line No: 115 hbdk-cc parameters:['--O0', '--debug', '--core-num', '1', '--fast', '--input-layout', 'NHWC', '--output-layout', 'NHWC', '--input-source', 'ddr'] 2025-03-13 14:11:33,849 file: hbdk_cc.py func: hbdk_cc line No: 116 hbdk-cc command used:hbdk-cc -f hbir -m /tmp/tmpuicub4az/torch_jit_subgraph_22.hbir -o /tmp/tmpuicub4az/torch_jit_subgraph_22.hbm --march bayes --progressbar --O0 --debug --core-num 1 --fast --input-layout NHWC --output-layout NHWC --input-source ddr 2025-03-13 14:11:38,274 file: tool_utils.py func: tool_utils line No: 322 consumed time 4.30012 2025-03-13 14:11:38,462 file: tool_utils.py func: tool_utils line No: 322 FPS=30770.01, latency = 259993.4 us, DDR = 24028 bytes (see torch_jit_subgraph_22.html) 2025-03-13 14:11:38,473 file: hybrid_build.py func: hybrid_build line No: 133 Compile submodel: torch_jit_subgraph_23 2025-03-13 14:11:39,057 file: hbdk_cc.py func: hbdk_cc line No: 115 hbdk-cc parameters:['--O0', '--debug', '--core-num', '1', '--fast', '--input-layout', 'NHWC', '--output-layout', 'NHWC', '--input-source', 'ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr'] 2025-03-13 14:11:39,057 file: hbdk_cc.py func: hbdk_cc line No: 116 hbdk-cc command used:hbdk-cc -f hbir -m /tmp/tmpuicub4az/torch_jit_subgraph_23.hbir -o /tmp/tmpuicub4az/torch_jit_subgraph_23.hbm --march bayes --progressbar --O0 --debug --core-num 1 --fast --input-layout NHWC --output-layout NHWC --input-source ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr 2025-03-13 14:11:39,102 file: tool_utils.py func: tool_utils line No: 317 Can not find the scale for node /pts_bbox_head/transformer/decoder/layers.5/attentions.2/Expand_output_0_/pts_bbox_head/transformer/decoder/layers.5/attentions.2/GatherElements_Cast 2025-03-13 14:11:39,103 file: tool_utils.py func: tool_utils line No: 317 Can not find the scale for node /pts_bbox_head/transformer/decoder/layers.5/attentions.2/Expand_1_output_0_/pts_bbox_head/transformer/decoder/layers.5/attentions.2/GatherElements_1_Cast 2025-03-13 14:11:39,103 file: tool_utils.py func: tool_utils line No: 317 Can not find the scale for node /pts_bbox_head/transformer/decoder/layers.5/attentions.2/Expand_2_output_0_/pts_bbox_head/transformer/decoder/layers.5/attentions.2/GatherElements_2_Cast 2025-03-13 14:11:39,103 file: tool_utils.py func: tool_utils line No: 317 Can not find the scale for node /pts_bbox_head/transformer/decoder/layers.5/attentions.2/Expand_3_output_0_/pts_bbox_head/transformer/decoder/layers.5/attentions.2/GatherElements_3_Cast 2025-03-13 14:11:45,799 file: tool_utils.py func: tool_utils line No: 322 consumed time 6.61706 2025-03-13 14:11:46,084 file: tool_utils.py func: tool_utils line No: 322 FPS=5.9, latency = 169618.0 us, DDR = 1165159168 bytes (see torch_jit_subgraph_23.html) 2025-03-13 14:11:46,108 file: hybrid_build.py func: hybrid_build line No: 133 Compile submodel: torch_jit_subgraph_24 2025-03-13 14:11:46,530 file: hbdk_cc.py func: hbdk_cc line No: 115 hbdk-cc parameters:['--O0', '--debug', '--core-num', '1', '--fast', '--input-layout', 'NHWC', '--output-layout', 'NHWC', '--input-source', 'ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr'] 2025-03-13 14:11:46,531 file: hbdk_cc.py func: hbdk_cc line No: 116 hbdk-cc command used:hbdk-cc -f hbir -m /tmp/tmpuicub4az/torch_jit_subgraph_24.hbir -o /tmp/tmpuicub4az/torch_jit_subgraph_24.hbm --march bayes --progressbar --O0 --debug --core-num 1 --fast --input-layout NHWC --output-layout NHWC --input-source ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr,ddr 2025-03-13 14:11:46,762 file: tool_utils.py func: tool_utils line No: 322 consumed time 0.120023 2025-03-13 14:11:46,903 file: tool_utils.py func: tool_utils line No: 322 FPS=1135.12, latency = 881.0 us, DDR = 2654560 bytes (see torch_jit_subgraph_24.html) 2025-03-13 14:11:46,905 file: hybrid_build.py func: hybrid_build line No: 133 Compile submodel: torch_jit_subgraph_25 2025-03-13 14:11:47,331 file: hbdk_cc.py func: hbdk_cc line No: 115 hbdk-cc parameters:['--O0', '--debug', '--core-num', '1', '--fast', '--input-layout', 'NHWC', '--output-layout', 'NHWC', '--input-source', 'ddr'] 2025-03-13 14:11:47,331 file: hbdk_cc.py func: hbdk_cc line No: 116 hbdk-cc command used:hbdk-cc -f hbir -m /tmp/tmpuicub4az/torch_jit_subgraph_25.hbir -o /tmp/tmpuicub4az/torch_jit_subgraph_25.hbm --march bayes --progressbar --O0 --debug --core-num 1 --fast --input-layout NHWC --output-layout NHWC --input-source ddr 2025-03-13 14:11:47,465 file: tool_utils.py func: tool_utils line No: 322 consumed time 0.0210528 2025-03-13 14:11:47,600 file: tool_utils.py func: tool_utils line No: 322 FPS=14669.93, latency = 68.2 us, DDR = 73216 bytes (see torch_jit_subgraph_25.html) 2025-03-13 14:11:47,602 file: hybrid_build.py func: hybrid_build line No: 133 Compile submodel: torch_jit_subgraph_26 2025-03-13 14:11:48,042 file: hbdk_cc.py func: hbdk_cc line No: 115 hbdk-cc parameters:['--O0', '--debug', '--core-num', '1', '--fast', '--input-layout', 'NHWC', '--output-layout', 'NHWC', '--input-source', 'ddr'] 2025-03-13 14:11:48,043 file: hbdk_cc.py func: hbdk_cc line No: 116 hbdk-cc command used:hbdk-cc -f hbir -m /tmp/tmpuicub4az/torch_jit_subgraph_26.hbir -o /tmp/tmpuicub4az/torch_jit_subgraph_26.hbm --march bayes --progressbar --O0 --debug --core-num 1 --fast --input-layout NHWC --output-layout NHWC --input-source ddr 2025-03-13 14:11:48,163 file: tool_utils.py func: tool_utils line No: 322 consumed time 0.0127919 2025-03-13 14:11:48,292 file: tool_utils.py func: tool_utils line No: 322 FPS=64143.68, latency = 15.6 us, DDR = 16096 bytes (see torch_jit_subgraph_26.html) 2025-03-13 14:11:55,427 file: build.py func: build line No: 39 End to compile the model with march bayes. 2025-03-13 14:11:55,567 file: print_node_info.py func: print_node_info line No: 57 The converted model node information: ================================================================================================================== Node ON Subgraph Type In/Out DataType ------------------------------------------------------------------------------------------------------------------- /Gather_reshape BPU id(0) Reshape int8/int8 /img_backbone/conv1/Conv BPU id(0) HzSQuantizedConv int8/int8 /img_backbone/maxpool/MaxPool BPU id(0) HzQuantizedMaxPool int8/int8 /img_backbone/layer1/layer1.0/conv1/Conv BPU id(0) HzSQuantizedConv int8/int8 /img_backbone/layer1/layer1.0/conv2/Conv BPU id(0) HzSQuantizedConv int8/int8 /img_backbone/layer1/layer1.0/conv3/Conv BPU id(0) HzSQuantizedConv int8/int8 ...ne/layer1/layer1.0/downsample/downsample.0/Conv BPU id(0) HzSQuantizedConv int8/int8 /img_backbone/layer1/layer1.1/conv1/Conv BPU id(0) HzSQuantizedConv int8/int8 /img_backbone/layer1/layer1.1/conv2/Conv BPU id(0) HzSQuantizedConv int8/int8 /img_backbone/layer1/layer1.1/conv3/Conv BPU id(0) HzSQuantizedConv int8/int8 /img_backbone/layer1/layer1.2/conv1/Conv BPU id(0) HzSQuantizedConv int8/int8 /img_backbone/layer1/layer1.2/conv2/Conv BPU id(0) HzSQuantizedConv int8/int8 /img_backbone/layer1/layer1.2/conv3/Conv BPU id(0) HzSQuantizedConv int8/int8 /img_backbone/layer2/layer2.0/conv1/Conv BPU id(0) HzSQuantizedConv int8/int8 /img_backbone/layer2/layer2.0/conv2/Conv BPU id(0) HzSQuantizedConv int8/int8 /img_backbone/layer2/layer2.0/conv3/Conv BPU id(0) HzSQuantizedConv int8/int8 ...ne/layer2/layer2.0/downsample/downsample.0/Conv BPU id(0) HzSQuantizedConv int8/int8 /img_backbone/layer2/layer2.1/conv1/Conv BPU id(0) HzSQuantizedConv int8/int8 /img_backbone/layer2/layer2.1/conv2/Conv BPU id(0) HzSQuantizedConv int8/int8 /img_backbone/layer2/layer2.1/conv3/Conv BPU id(0) HzSQuantizedConv int8/int8 /img_backbone/layer2/layer2.2/conv1/Conv BPU id(0) HzSQuantizedConv int8/int8 /img_backbone/layer2/layer2.2/conv2/Conv BPU id(0) HzSQuantizedConv int8/int8 /img_backbone/layer2/layer2.2/conv3/Conv BPU id(0) HzSQuantizedConv int8/int8 /img_backbone/layer2/layer2.3/conv1/Conv BPU id(0) HzSQuantizedConv int8/int8 /img_backbone/layer2/layer2.3/conv2/Conv BPU id(0) HzSQuantizedConv int8/int8 /img_backbone/layer2/layer2.3/conv3/Conv BPU id(0) HzSQuantizedConv int8/int8 /img_backbone/layer3/layer3.0/conv1/Conv BPU id(0) HzSQuantizedConv int8/int8 /img_backbone/layer3/layer3.0/conv2/Conv BPU id(0) HzSQuantizedConv int8/int8 /img_backbone/layer3/layer3.0/conv3/Conv BPU id(0) HzSQuantizedConv int8/int8 ...ne/layer3/layer3.0/downsample/downsample.0/Conv BPU id(0) HzSQuantizedConv int8/int8 /img_backbone/layer3/layer3.1/conv1/Conv BPU id(0) HzSQuantizedConv int8/int8 /img_backbone/layer3/layer3.1/conv2/Conv BPU id(0) HzSQuantizedConv int8/int8 /img_backbone/layer3/layer3.1/conv3/Conv BPU id(0) HzSQuantizedConv int8/int8 /img_backbone/layer3/layer3.2/conv1/Conv BPU id(0) HzSQuantizedConv int8/int8 /img_backbone/layer3/layer3.2/conv2/Conv BPU id(0) HzSQuantizedConv int8/int8 /img_backbone/layer3/layer3.2/conv3/Conv BPU id(0) HzSQuantizedConv int8/int8 /img_backbone/layer3/layer3.3/conv1/Conv BPU id(0) HzSQuantizedConv int8/int8 /img_backbone/layer3/layer3.3/conv2/Conv BPU id(0) HzSQuantizedConv int8/int8 /img_backbone/layer3/layer3.3/conv3/Conv BPU id(0) HzSQuantizedConv int8/int8 /img_backbone/layer3/layer3.4/conv1/Conv BPU id(0) HzSQuantizedConv int8/int8 /img_backbone/layer3/layer3.4/conv2/Conv BPU id(0) HzSQuantizedConv int8/int8 /img_backbone/layer3/layer3.4/conv3/Conv BPU id(0) HzSQuantizedConv int8/int8 /img_backbone/layer3/layer3.5/conv1/Conv BPU id(0) HzSQuantizedConv int8/int8 /img_backbone/layer3/layer3.5/conv2/Conv BPU id(0) HzSQuantizedConv int8/int8 /img_backbone/layer3/layer3.5/conv3/Conv BPU id(0) HzSQuantizedConv int8/int8 /img_backbone/layer4/layer4.0/conv1/Conv BPU id(0) HzSQuantizedConv int8/int8 /img_backbone/layer4/layer4.0/conv2/Conv BPU id(0) HzSQuantizedConv int8/int8 /img_backbone/layer4/layer4.0/conv3/Conv BPU id(0) HzSQuantizedConv int8/int8 ...ne/layer4/layer4.0/downsample/downsample.0/Conv BPU id(0) HzSQuantizedConv int8/int8 /img_backbone/layer4/layer4.1/conv1/Conv BPU id(0) HzSQuantizedConv int8/int8 /img_backbone/layer4/layer4.1/conv2/Conv BPU id(0) HzSQuantizedConv int8/int8 /img_backbone/layer4/layer4.1/conv3/Conv BPU id(0) HzSQuantizedConv int8/int8 /img_backbone/layer4/layer4.2/conv1/Conv BPU id(0) HzSQuantizedConv int8/int8 /img_backbone/layer4/layer4.2/conv2/Conv BPU id(0) HzSQuantizedConv int8/int8 /img_backbone/layer4/layer4.2/conv3/Conv BPU id(0) HzSQuantizedConv int8/int8 /img_neck/lateral_convs.0/conv/Conv BPU id(0) HzSQuantizedConv int8/int8 /img_neck/fpn_convs.0/conv/Conv BPU id(0) HzSQuantizedConv int8/int8 ...bbox_head/transformer/encoder/depth_net/Reshape CPU -- Reshape float/float ...sformer/encoder/depth_net/bn/BatchNormalization CPU -- BatchNormalization float/float ...ncoder/depth_net/reduce_conv/reduce_conv.0/Conv BPU id(0) HzSQuantizedConv int8/int8 ...oder/depth_net/context_mlp/fc1/Gemm_pre_reshape BPU id(0) Reshape int8/int8 ...nsformer/encoder/depth_net/context_mlp/fc1/Gemm BPU id(0) HzSQuantizedConv int8/int8 ...nsformer/encoder/depth_net/context_mlp/fc2/Gemm BPU id(0) HzSQuantizedConv int8/int8 ...r/encoder/depth_net/context_se/conv_reduce/Conv BPU id(0) HzSQuantizedConv int8/int8 ...r/encoder/depth_net/context_se/conv_expand/Conv BPU id(0) HzSQuantizedConv int8/int8 ...ormer/encoder/depth_net/context_se/gate/Sigmoid BPU id(0) HzLut int8/int8 ...ad/transformer/encoder/depth_net/context_se/Mul BPU id(0) HzSElementwiseMul int8/int8 ...transformer/encoder/depth_net/context_conv/Conv BPU id(0) HzSQuantizedConv int8/int8 ...ransformer/encoder/depth_net/depth_mlp/fc1/Gemm BPU id(0) HzSQuantizedConv int8/int8 ...ransformer/encoder/depth_net/depth_mlp/fc2/Gemm BPU id(0) HzSQuantizedConv int8/int8 ...mer/encoder/depth_net/depth_se/conv_reduce/Conv BPU id(0) HzSQuantizedConv int8/int8 ...mer/encoder/depth_net/depth_se/conv_expand/Conv BPU id(0) HzSQuantizedConv int8/int8 ...sformer/encoder/depth_net/depth_se/gate/Sigmoid BPU id(0) HzLut int8/int8 ...head/transformer/encoder/depth_net/depth_se/Mul BPU id(0) HzSElementwiseMul int8/int8 ...er/depth_net/depth_conv/depth_conv.0/conv1/Conv BPU id(0) HzSQuantizedConv int8/int8 ...er/depth_net/depth_conv/depth_conv.0/conv2/Conv BPU id(0) HzSQuantizedConv int8/int8 ...er/depth_net/depth_conv/depth_conv.1/conv1/Conv BPU id(0) HzSQuantizedConv int8/int8 ...er/depth_net/depth_conv/depth_conv.1/conv2/Conv BPU id(0) HzSQuantizedConv int8/int8 ...er/depth_net/depth_conv/depth_conv.2/conv1/Conv BPU id(0) HzSQuantizedConv int8/int8 ...er/depth_net/depth_conv/depth_conv.2/conv2/Conv BPU id(0) HzSQuantizedConv int8/int8 .../depth_conv/depth_conv.3/aspp1/atrous_conv/Conv BPU id(0) HzSQuantizedConv int8/int8 .../depth_conv/depth_conv.3/aspp2/atrous_conv/Conv BPU id(0) HzSQuantizedConv int8/int8 .../depth_conv/depth_conv.3/aspp3/atrous_conv/Conv BPU id(0) HzSQuantizedConv int8/int8 .../depth_conv/depth_conv.3/aspp4/atrous_conv/Conv BPU id(0) HzSQuantizedConv int8/int8 ...al_avg_pool/global_avg_pool.0/GlobalAveragePool BPU id(0) HzSQuantizedGlobalAveragePool int8/int8 ...h_conv.3/global_avg_pool/global_avg_pool.1/Conv BPU id(0) HzSQuantizedConv int8/int8 ...ncoder/depth_net/depth_conv/depth_conv.3/Resize BPU id(0) HzQuantizedRoiResize int8/int8 ...oder/depth_net/depth_conv/depth_conv.3/Concat_2 BPU id(0) Concat int8/int8 ...er/depth_net/depth_conv/depth_conv.3/conv1/Conv BPU id(0) HzSQuantizedConv int8/int8 .../encoder/depth_net/depth_conv/depth_conv.4/Conv BPU id(0) HzSQuantizedConv int8/int8 ...box_head/transformer/encoder/depth_net/Concat_1 BPU id(0) Concat int8/int8 /pts_bbox_head/transformer/encoder/Slice BPU id(0) Slice int8/int8 /pts_bbox_head/transformer/encoder/Softmax CPU -- HzSoftmax float/float /pts_bbox_head/transformer/encoder/Slice_1 BPU id(0) Slice int8/int8 /pts_bbox_head/transformer/encoder/Reshape_1 BPU id(0) Reshape int8/int8 /pts_bbox_head/transformer/encoder/Reshape_2 BPU id(1) Reshape int8/int8 /pts_bbox_head/transformer/encoder/Mul BPU id(1) HzSElementwiseMul int8/int8 /pts_bbox_head/transformer/encoder/Transpose_4 BPU id(1) Transpose int8/int8 /pts_bbox_head/transformer/encoder/Flatten CPU -- Reshape float/float /pts_bbox_head/transformer/encoder/Transpose_5 CPU -- Transpose float/float /pts_bbox_head/transformer/encoder/Cast_2 CPU -- Cast int8/int32 /pts_bbox_head/transformer/encoder/ScatterElements CPU -- ScatterElements float/float /pts_bbox_head/transformer/encoder/Not_cast CPU -- Cast int8/float /pts_bbox_head/transformer/encoder/Not_to_equal CPU -- Equal float/int8 /pts_bbox_head/transformer/encoder/Where_cast CPU -- Cast int8/float /pts_bbox_head/transformer/encoder/Where_equal CPU -- Equal float/int8 ...ead/transformer/encoder/Where_equal_output_cast CPU -- Cast int8/float /pts_bbox_head/transformer/encoder/Where_mul_y CPU -- Mul float/float /pts_bbox_head/transformer/encoder/Cast_4 CPU -- Cast int8/int8 /pts_bbox_head/transformer/encoder/Gather CPU -- Gather float/float /pts_bbox_head/transformer/encoder/ReduceSum BPU id(2) HzSQuantizedReduceSum int8/int16 /pts_bbox_head/transformer/encoder/ScatterND CPU -- ScatterND float/float /pts_bbox_head/transformer/encoder/Slice_3 CPU -- Slice float/float /pts_bbox_head/transformer/encoder/Transpose_6 CPU -- Transpose float/float /pts_bbox_head/transformer/encoder/Reshape_4 CPU -- Reshape float/float ...ransformer/encoder/downsample/downsample.0/Conv BPU id(3) HzSQuantizedConv int8/int8 ...ransformer/encoder/downsample/downsample.3/Conv BPU id(3) HzSQuantizedConv int8/int8 ...ransformer/encoder/downsample/downsample.6/Conv BPU id(3) HzSQuantizedConv int8/int8 ...decoder/layers.0/attentions.2/value_proj/MatMul BPU id(3) HzSQuantizedConv int8/int8 ...tentions.2/value_proj/MatMul_output_0_Reshape_0 BPU id(3) Reshape int8/int8 ...d/transformer/decoder/layers.0/attentions.2/Pad BPU id(3) HzPad int8/int8 ...former/decoder/layers.0/attentions.2/Reshape_11 BPU id(3) Reshape int8/int8 ...er/decoder/layers.0/attentions.2/GatherElements BPU id(3) GatherElements int8/int8 .../decoder/layers.0/attentions.2/GatherElements_1 BPU id(3) GatherElements int8/int8 .../decoder/layers.0/attentions.2/GatherElements_2 BPU id(3) GatherElements int8/int8 .../decoder/layers.0/attentions.2/GatherElements_3 BPU id(3) GatherElements int8/int8 ...ransformer/decoder/layers.0/attentions.2/Mul_14 BPU id(3) HzSElementwiseMul int8/int8 ...ransformer/decoder/layers.0/attentions.2/Mul_15 BPU id(3) HzSElementwiseMul int8/int8 ...ransformer/decoder/layers.0/attentions.2/Add_12 BPU id(3) HzSElementwiseAdd int8/int8 ...ransformer/decoder/layers.0/attentions.2/Mul_16 BPU id(3) HzSElementwiseMul int8/int8 ...ransformer/decoder/layers.0/attentions.2/Add_13 BPU id(3) HzSElementwiseAdd int8/int8 ...ransformer/decoder/layers.0/attentions.2/Mul_17 BPU id(3) HzSElementwiseMul int8/int8 ...ransformer/decoder/layers.0/attentions.2/Add_14 BPU id(3) HzSElementwiseAdd int8/int8 ...ransformer/decoder/layers.0/attentions.2/Mul_18 BPU id(3) HzSElementwiseMul int8/int8 ...layers.0/attentions.2/Mul_18_output_0_Reshape_0 BPU id(3) Reshape int8/int8 ...sformer/decoder/layers.0/attentions.2/ReduceSum BPU id(3) HzSQuantizedReduceSum int8/int8 ...decoder/layers.0/attentions.2/ReduceSum_reshape BPU id(3) Reshape int8/int8 ...ecoder/layers.0/attentions.2/output_proj/MatMul BPU id(3) HzSQuantizedConv int8/int8 .../attentions.2/output_proj/MatMul_reshape_output BPU id(3) Reshape int8/int8 ...ransformer/decoder/layers.0/attentions.2/Add_15 BPU id(3) HzSElementwiseAdd int8/int8 ...transformer/decoder/layers.0/norms.2/ReduceMean BPU id(3) HzSQuantizedReduceMean int8/int8 ...x_head/transformer/decoder/layers.0/norms.2/Sub BPU id(3) HzSElementwiseSub int8/int8 ...x_head/transformer/decoder/layers.0/norms.2/Pow BPU id(3) HzLut int8/int8 ...ansformer/decoder/layers.0/norms.2/ReduceMean_1 BPU id(3) HzSQuantizedReduceMean int8/int8 ...sformer/decoder/layers.0/norms.2/Div_reciprocal BPU id(3) HzLut int8/int8 ...ad/transformer/decoder/layers.0/norms.2/Div_mul BPU id(3) HzSElementwiseMul int8/int8 ...x_head/transformer/decoder/layers.0/norms.2/Mul BPU id(3) HzSElementwiseMul int8/int8 ...head/transformer/decoder/layers.0/norms.2/Add_1 BPU id(3) HzSElementwiseAdd int8/int8 ...layers/layers.0/layers.0.0/MatMul_reshape_input BPU id(3) Reshape int8/int8 ...yers.0/ffns.0/layers/layers.0/layers.0.0/MatMul BPU id(3) HzSQuantizedConv int8/int8 .../decoder/layers.0/ffns.0/layers/layers.1/MatMul BPU id(3) HzSQuantizedConv int8/int8 ....0/ffns.0/layers/layers.1/MatMul_reshape_output BPU id(3) Reshape int8/int8 ...ox_head/transformer/decoder/layers.0/ffns.0/Add BPU id(3) HzSElementwiseAdd int8/int8 ...transformer/decoder/layers.0/norms.3/ReduceMean BPU id(3) HzSQuantizedReduceMean int8/int8 ...x_head/transformer/decoder/layers.0/norms.3/Sub BPU id(3) HzSElementwiseSub int8/int8 ...x_head/transformer/decoder/layers.0/norms.3/Pow BPU id(3) HzLut int8/int8 ...ansformer/decoder/layers.0/norms.3/ReduceMean_1 BPU id(3) HzSQuantizedReduceMean int8/int8 ...sformer/decoder/layers.0/norms.3/Div_reciprocal BPU id(3) HzLut int8/int8 ...ad/transformer/decoder/layers.0/norms.3/Div_mul BPU id(3) HzSElementwiseMul int8/int8 ...x_head/transformer/decoder/layers.0/norms.3/Mul BPU id(3) HzSElementwiseMul int8/int8 ...head/transformer/decoder/layers.0/norms.3/Add_1 BPU id(3) HzSElementwiseAdd int8/int8 ...bbox_head/transformer/decoder/Transpose_reshape BPU id(3) Reshape int8/int8 .../decoder/reg_branches.0/reg_branches.0.0/MatMul BPU id(3) HzSQuantizedConv int8/int8 .../decoder/reg_branches.0/reg_branches.0.2/MatMul BPU id(3) HzSQuantizedConv int8/int8 .../decoder/reg_branches.0/reg_branches.0.4/MatMul BPU id(3) HzSQuantizedConv int8/int8 /pts_bbox_head/transformer/decoder/Sigmoid BPU id(3) HzLut int8/int8 .../transformer/decoder/Sigmoid_output_0_Reshape_0 BPU id(3) Reshape int8/int8 /pts_bbox_head/transformer/decoder/Unsqueeze_1 BPU id(3) Reshape int8/int8 ..._bbox_head/transformer/decoder/layers.1/Reshape BPU id(3) Reshape int8/int8 ...d/transformer/decoder/layers.1/attentions.0/Add BPU id(3) HzSElementwiseAdd int8/int8 ...er/layers.1/attentions.0/Add_output_0_Reshape_0 BPU id(3) Reshape int8/int8 ...ormer/decoder/layers.1/attentions.0/attn/MatMul BPU id(3) HzSQuantizedConv int8/int8 ...ayers.1/attentions.0/attn/MatMul_reshape_output BPU id(3) Reshape int8/int8 ...mer/decoder/layers.1/attentions.0/attn/MatMul_1 BPU id(3) HzSQuantizedConv int8/int8 ...ers.1/attentions.0/attn/MatMul_1_reshape_output BPU id(3) Reshape int8/int8 ...yers.1/attentions.0/attn/MatMul_2_reshape_input BPU id(3) Reshape int8/int8 ...mer/decoder/layers.1/attentions.0/attn/MatMul_2 BPU id(3) HzSQuantizedConv int8/int8 ...ers.1/attentions.0/attn/MatMul_2_reshape_output BPU id(3) Reshape int8/int8 .../decoder/layers.1/attentions.0/attn/Transpose_4 BPU id(3) Transpose int8/int8 ...er/decoder/layers.1/attentions.0/attn/Div_2_mul BPU id(3) HzSElementwiseMul int8/int8 ...rs.1/attentions.0/attn/Div_2_output_0_Transpose BPU id(3) Transpose int8/int8 .../decoder/layers.1/attentions.0/attn/Transpose_5 BPU id(3) Transpose int8/int8 ...mer/decoder/layers.1/attentions.0/attn/MatMul_3 BPU id(3) HzSQuantizedMatmul int8/int8 ...former/decoder/layers.1/attentions.0/attn/Mul_6 BPU id(3) HzSElementwiseMul int8/int16 ...rmer/decoder/layers.1/attentions.0/attn/Softmax CPU -- Softmax float/float ...mer/decoder/layers.1/attentions.0/attn/MatMul_4 BPU id(4) HzSQuantizedMatmul int8/int8 .../decoder/layers.1/attentions.0/attn/Transpose_6 BPU id(4) Transpose int8/int8 ...er/decoder/layers.1/attentions.0/attn/Reshape_3 BPU id(4) Reshape int8/int8 ...sformer/decoder/layers.1/attentions.0/attn/Gemm BPU id(4) HzSQuantizedConv int8/int8 ...transformer/decoder/layers.1/attentions.0/Add_2 BPU id(4) HzSElementwiseAdd int8/int8 ...box_head/transformer/decoder/layers.1/Reshape_4 BPU id(4) Reshape int8/int8 ...transformer/decoder/layers.1/norms.0/ReduceMean BPU id(4) HzSQuantizedReduceMean int8/int8 ...x_head/transformer/decoder/layers.1/norms.0/Sub BPU id(4) HzSElementwiseSub int8/int8 ...x_head/transformer/decoder/layers.1/norms.0/Pow BPU id(4) HzLut int8/int8 ...ansformer/decoder/layers.1/norms.0/ReduceMean_1 BPU id(4) HzSQuantizedReduceMean int8/int8 ...sformer/decoder/layers.1/norms.0/Div_reciprocal BPU id(4) HzLut int8/int8 ...ad/transformer/decoder/layers.1/norms.0/Div_mul BPU id(4) HzSElementwiseMul int8/int8 ...x_head/transformer/decoder/layers.1/norms.0/Mul BPU id(4) HzSElementwiseMul int8/int8 ...head/transformer/decoder/layers.1/norms.0/Add_1 BPU id(4) HzSElementwiseAdd int8/int8 ...box_head/transformer/decoder/layers.1/Transpose BPU id(4) Transpose int8/int8 ...box_head/transformer/decoder/layers.1/Reshape_9 BPU id(4) Reshape int8/int8 ...d/transformer/decoder/layers.1/attentions.1/Add BPU id(4) HzSElementwiseAdd int8/int8 ...layers.1/attentions.1/attn/MatMul_reshape_input BPU id(4) Reshape int8/int8 ...ormer/decoder/layers.1/attentions.1/attn/MatMul BPU id(4) HzSQuantizedConv int8/int8 ...ayers.1/attentions.1/attn/MatMul_reshape_output BPU id(4) Reshape int8/int8 ...mer/decoder/layers.1/attentions.1/attn/MatMul_1 BPU id(4) HzSQuantizedConv int8/int8 ...ers.1/attentions.1/attn/MatMul_1_reshape_output BPU id(4) Reshape int8/int8 ...yers.1/attentions.1/attn/MatMul_2_reshape_input BPU id(4) Reshape int8/int8 ...mer/decoder/layers.1/attentions.1/attn/MatMul_2 BPU id(4) HzSQuantizedConv int8/int8 ...ers.1/attentions.1/attn/MatMul_2_reshape_output BPU id(4) Reshape int8/int8 .../decoder/layers.1/attentions.1/attn/Transpose_4 BPU id(4) Transpose int8/int8 ...er/decoder/layers.1/attentions.1/attn/Div_2_mul BPU id(4) HzSElementwiseMul int8/int8 ...rs.1/attentions.1/attn/Div_2_output_0_Transpose BPU id(4) Transpose int8/int8 .../decoder/layers.1/attentions.1/attn/Transpose_5 BPU id(4) Transpose int8/int8 ...mer/decoder/layers.1/attentions.1/attn/MatMul_3 BPU id(4) HzSQuantizedMatmul int8/int32 ...rmer/decoder/layers.1/attentions.1/attn/Softmax CPU -- Softmax float/float ...mer/decoder/layers.1/attentions.1/attn/MatMul_4 BPU id(5) HzSQuantizedMatmul int8/int8 .../decoder/layers.1/attentions.1/attn/Transpose_6 BPU id(5) Transpose int8/int8 ...er/decoder/layers.1/attentions.1/attn/Reshape_3 BPU id(5) Reshape int8/int8 ...sformer/decoder/layers.1/attentions.1/attn/Gemm BPU id(5) HzSQuantizedConv int8/int8 ...transformer/decoder/layers.1/attentions.1/Add_2 BPU id(5) HzSElementwiseAdd int8/int8 ...ox_head/transformer/decoder/layers.1/Reshape_12 BPU id(5) Reshape int8/int8 ...transformer/decoder/layers.1/norms.1/ReduceMean BPU id(5) HzSQuantizedReduceMean int8/int8 ...x_head/transformer/decoder/layers.1/norms.1/Sub BPU id(5) HzSElementwiseSub int8/int8 ...x_head/transformer/decoder/layers.1/norms.1/Pow BPU id(5) HzLut int8/int8 ...ansformer/decoder/layers.1/norms.1/ReduceMean_1 BPU id(5) HzSQuantizedReduceMean int8/int8 ...sformer/decoder/layers.1/norms.1/Div_reciprocal BPU id(5) HzLut int8/int8 ...ad/transformer/decoder/layers.1/norms.1/Div_mul BPU id(5) HzSElementwiseMul int8/int8 ...x_head/transformer/decoder/layers.1/norms.1/Mul BPU id(5) HzSElementwiseMul int8/int8 ...head/transformer/decoder/layers.1/norms.1/Add_1 BPU id(5) HzSElementwiseAdd int8/int8 ...yers.1/norms.1/Add_1_output_0_reshape_Transpose BPU id(5) Transpose int8/int8 ...coder/layers.1/norms.1/Add_1_output_0_Reshape_0 BPU id(5) Reshape int8/int8 ...d/transformer/decoder/layers.1/attentions.2/Add BPU id(5) HzSQuantizedConv int8/int8 ...er/layers.1/attentions.2/Add_output_0_Reshape_0 BPU id(5) Reshape int8/int8 ...decoder/layers.1/attentions.2/value_proj/MatMul BPU id(3) HzSQuantizedConv int8/int8 ...tentions.2/value_proj/MatMul_output_0_Reshape_0 BPU id(3) Reshape int8/int8 ...r/layers.1/attentions.2/sampling_offsets/MatMul BPU id(5) HzSQuantizedConv int8/int8 ...ntions.2/sampling_offsets/MatMul_reshape_output BPU id(5) Reshape int8/int8 .../layers.1/attentions.2/attention_weights/MatMul BPU id(5) HzSQuantizedConv int8/int32 ...tions.2/attention_weights/MatMul_reshape_output CPU -- Reshape float/float ...ansformer/decoder/layers.1/attentions.2/Softmax CPU -- Softmax float/float ...sformer/decoder/layers.1/attentions.2/Reshape_3 BPU id(7) Reshape int8/int8 ...ansformer/decoder/layers.1/attentions.2/Div_mul BPU id(5) HzSElementwiseMul int8/int16 ...transformer/decoder/layers.1/attentions.2/Add_1 CPU -- Add float/float ...d/transformer/decoder/layers.1/attentions.2/Mul BPU id(6) HzSElementwiseMul int8/int8 ...d/transformer/decoder/layers.1/attentions.2/Sub BPU id(6) HzSElementwiseSub int8/int8 ...er/decoder/layers.1/attentions.2/Gather_reshape BPU id(6) Reshape int8/int8 ...ormer/decoder/layers.1/attentions.2/Transpose_3 BPU id(6) Transpose int8/int8 ...sformer/decoder/layers.1/attentions.2/Reshape_6 BPU id(6) Reshape int8/int8 ...nsformer/decoder/layers.1/attentions.2/Gather_1 BPU id(6) Gather int8/int8 ...nsformer/decoder/layers.1/attentions.2/Gather_2 BPU id(6) Gather int8/int8 ...transformer/decoder/layers.1/attentions.2/Add_2 BPU id(6) HzSElementwiseAdd int8/int8 ...transformer/decoder/layers.1/attentions.2/Mul_1 BPU id(6) HzSElementwiseMul int8/int8 ...transformer/decoder/layers.1/attentions.2/Sub_1 BPU id(6) HzSElementwiseSub int8/int8 ...sformer/decoder/layers.1/attentions.2/Div_1_mul BPU id(6) HzSElementwiseMul int8/int8 .../layers.1/attentions.2/Div_1_output_0_Reshape_0 BPU id(6) Reshape int8/int8 ...transformer/decoder/layers.1/attentions.2/Add_3 BPU id(6) HzSElementwiseAdd int8/int8 ...transformer/decoder/layers.1/attentions.2/Mul_2 BPU id(6) HzSElementwiseMul int8/int8 ...transformer/decoder/layers.1/attentions.2/Sub_2 BPU id(6) HzSElementwiseSub int8/int8 ...sformer/decoder/layers.1/attentions.2/Div_2_mul BPU id(6) HzSElementwiseMul int8/int8 .../layers.1/attentions.2/Div_2_output_0_Reshape_0 BPU id(6) Reshape int8/int8 ...transformer/decoder/layers.1/attentions.2/Floor BPU id(6) HzLut int8/int8 .../transformer/decoder/layers.1/attentions.2/Cast CPU -- Cast float/int8 ...ansformer/decoder/layers.1/attentions.2/Floor_1 BPU id(6) HzLut int8/int8 ...ransformer/decoder/layers.1/attentions.2/Cast_1 CPU -- Cast float/int8 ...transformer/decoder/layers.1/attentions.2/Add_4 CPU -- Add int8/int8 ...transformer/decoder/layers.1/attentions.2/Add_5 CPU -- Add int8/int8 ...ransformer/decoder/layers.1/attentions.2/Cast_2 CPU -- Cast int8/float ...transformer/decoder/layers.1/attentions.2/Sub_3 BPU id(7) HzSElementwiseSub int8/int8 ...ransformer/decoder/layers.1/attentions.2/Cast_3 CPU -- Cast int8/float ...transformer/decoder/layers.1/attentions.2/Sub_4 BPU id(7) HzSElementwiseSub int8/int8 ...transformer/decoder/layers.1/attentions.2/Mul_3 BPU id(7) HzSElementwiseMul int8/int8 ...rmer/decoder/layers.1/attentions.2/Unsqueeze_25 BPU id(7) Reshape int8/int8 ...ransformer/decoder/layers.1/attentions.2/Cast_4 CPU -- Cast int8/float ...transformer/decoder/layers.1/attentions.2/Sub_5 BPU id(7) HzSElementwiseSub int8/int8 ...transformer/decoder/layers.1/attentions.2/Mul_4 BPU id(7) HzSElementwiseMul int8/int8 ...rmer/decoder/layers.1/attentions.2/Unsqueeze_26 BPU id(7) Reshape int8/int8 ...ransformer/decoder/layers.1/attentions.2/Cast_5 CPU -- Cast int8/float ...transformer/decoder/layers.1/attentions.2/Sub_6 BPU id(7) HzSElementwiseSub int8/int8 ...transformer/decoder/layers.1/attentions.2/Mul_5 BPU id(7) HzSElementwiseMul int8/int8 ...rmer/decoder/layers.1/attentions.2/Unsqueeze_27 BPU id(7) Reshape int8/int8 ...transformer/decoder/layers.1/attentions.2/Mul_6 BPU id(7) HzSElementwiseMul int8/int8 ...rmer/decoder/layers.1/attentions.2/Unsqueeze_28 BPU id(7) Reshape int8/int8 ...d/transformer/decoder/layers.1/attentions.2/Pad BPU id(3) HzPad int8/int8 ...transformer/decoder/layers.1/attentions.2/Add_6 CPU -- Add int8/int8 ...transformer/decoder/layers.1/attentions.2/Add_7 CPU -- Add int8/int8 .../transformer/decoder/layers.1/attentions.2/Less CPU -- Less int8/int8 ...former/decoder/layers.1/attentions.2/Where_cast CPU -- Cast int8/int8 ...ormer/decoder/layers.1/attentions.2/Where_mul_x CPU -- Mul int8/int8 ...ormer/decoder/layers.1/attentions.2/Where_equal CPU -- Equal int8/int8 ...r/layers.1/attentions.2/Where_equal_output_cast CPU -- Cast int8/int8 ...ormer/decoder/layers.1/attentions.2/Where_mul_y CPU -- Mul int8/int8 ...sformer/decoder/layers.1/attentions.2/Where_add CPU -- Add int8/int8 ...ansformer/decoder/layers.1/attentions.2/Greater CPU -- Greater int8/int8 ...rmer/decoder/layers.1/attentions.2/Where_1_cast CPU -- Cast int8/int8 ...mer/decoder/layers.1/attentions.2/Where_1_mul_x CPU -- Mul int8/int8 ...mer/decoder/layers.1/attentions.2/Where_1_equal CPU -- Equal int8/int8 ...layers.1/attentions.2/Where_1_equal_output_cast CPU -- Cast int8/int8 ...mer/decoder/layers.1/attentions.2/Where_1_mul_y CPU -- Mul int8/int8 ...ormer/decoder/layers.1/attentions.2/Where_1_add CPU -- Add int8/int8 ...ransformer/decoder/layers.1/attentions.2/Less_1 CPU -- Less int8/int8 ...rmer/decoder/layers.1/attentions.2/Where_2_cast CPU -- Cast int8/int8 ...mer/decoder/layers.1/attentions.2/Where_2_mul_x CPU -- Mul int8/int8 ...mer/decoder/layers.1/attentions.2/Where_2_equal CPU -- Equal int8/int8 ...layers.1/attentions.2/Where_2_equal_output_cast CPU -- Cast int8/int8 ...mer/decoder/layers.1/attentions.2/Where_2_mul_y CPU -- Mul int8/int8 ...ormer/decoder/layers.1/attentions.2/Where_2_add CPU -- Add int8/int8 ...sformer/decoder/layers.1/attentions.2/Greater_1 CPU -- Greater int8/int8 ...rmer/decoder/layers.1/attentions.2/Where_3_cast CPU -- Cast int8/int8 ...mer/decoder/layers.1/attentions.2/Where_3_mul_x CPU -- Mul int8/int8 ...mer/decoder/layers.1/attentions.2/Where_3_equal CPU -- Equal int8/int8 ...layers.1/attentions.2/Where_3_equal_output_cast CPU -- Cast int8/int8 ...mer/decoder/layers.1/attentions.2/Where_3_mul_y CPU -- Mul int8/int8 ...ormer/decoder/layers.1/attentions.2/Where_3_add CPU -- Add int8/int8 ...ransformer/decoder/layers.1/attentions.2/Less_2 CPU -- Less int8/int8 ...rmer/decoder/layers.1/attentions.2/Where_4_cast CPU -- Cast int8/int8 ...mer/decoder/layers.1/attentions.2/Where_4_mul_x CPU -- Mul int8/int8 ...mer/decoder/layers.1/attentions.2/Where_4_equal CPU -- Equal int8/int8 ...layers.1/attentions.2/Where_4_equal_output_cast CPU -- Cast int8/int8 ...mer/decoder/layers.1/attentions.2/Where_4_mul_y CPU -- Mul int8/int8 ...ormer/decoder/layers.1/attentions.2/Where_4_add CPU -- Add int8/int8 ...sformer/decoder/layers.1/attentions.2/Greater_2 CPU -- Greater int8/int8 ...rmer/decoder/layers.1/attentions.2/Where_5_cast CPU -- Cast int8/int8 ...mer/decoder/layers.1/attentions.2/Where_5_mul_x CPU -- Mul int8/int8 ...mer/decoder/layers.1/attentions.2/Where_5_equal CPU -- Equal int8/int8 ...layers.1/attentions.2/Where_5_equal_output_cast CPU -- Cast int8/int8 ...mer/decoder/layers.1/attentions.2/Where_5_mul_y CPU -- Mul int8/int8 ...ormer/decoder/layers.1/attentions.2/Where_5_add CPU -- Add int8/int8 ...ransformer/decoder/layers.1/attentions.2/Less_3 CPU -- Less int8/int8 ...rmer/decoder/layers.1/attentions.2/Where_6_cast CPU -- Cast int8/int8 ...mer/decoder/layers.1/attentions.2/Where_6_mul_x CPU -- Mul int8/int8 ...mer/decoder/layers.1/attentions.2/Where_6_equal CPU -- Equal int8/int8 ...layers.1/attentions.2/Where_6_equal_output_cast CPU -- Cast int8/int8 ...mer/decoder/layers.1/attentions.2/Where_6_mul_y CPU -- Mul int8/int8 ...ormer/decoder/layers.1/attentions.2/Where_6_add CPU -- Add int8/int8 ...sformer/decoder/layers.1/attentions.2/Greater_3 CPU -- Greater int8/int8 ...rmer/decoder/layers.1/attentions.2/Where_7_cast CPU -- Cast int8/int8 ...mer/decoder/layers.1/attentions.2/Where_7_mul_x CPU -- Mul int8/int8 ...mer/decoder/layers.1/attentions.2/Where_7_equal CPU -- Equal int8/int8 ...layers.1/attentions.2/Where_7_equal_output_cast CPU -- Cast int8/int8 ...mer/decoder/layers.1/attentions.2/Where_7_mul_y CPU -- Mul int8/int8 ...ormer/decoder/layers.1/attentions.2/Where_7_add CPU -- Add int8/int8 ...former/decoder/layers.1/attentions.2/Reshape_11 BPU id(3) Reshape int8/int8 ...transformer/decoder/layers.1/attentions.2/Mul_8 CPU -- Mul int8/int8 ...transformer/decoder/layers.1/attentions.2/Add_8 CPU -- Add int8/int8 ...rmer/decoder/layers.1/attentions.2/Unsqueeze_32 CPU -- Reshape int8/int8 ...ransformer/decoder/layers.1/attentions.2/Expand CPU -- Expand int8/int8 ...ransformer/decoder/layers.1/attentions.2/Mul_10 CPU -- Mul int8/int8 ...transformer/decoder/layers.1/attentions.2/Add_9 CPU -- Add int8/int8 ...rmer/decoder/layers.1/attentions.2/Unsqueeze_36 CPU -- Reshape int8/int8 ...nsformer/decoder/layers.1/attentions.2/Expand_1 CPU -- Expand int8/int8 ...ransformer/decoder/layers.1/attentions.2/Add_10 CPU -- Add int8/int8 ...rmer/decoder/layers.1/attentions.2/Unsqueeze_40 CPU -- Reshape int8/int8 ...nsformer/decoder/layers.1/attentions.2/Expand_2 CPU -- Expand int8/int8 ...ransformer/decoder/layers.1/attentions.2/Add_11 CPU -- Add int8/int8 ...rmer/decoder/layers.1/attentions.2/Unsqueeze_44 CPU -- Reshape int8/int8 ...nsformer/decoder/layers.1/attentions.2/Expand_3 CPU -- Expand int8/int8 ...coder/layers.1/attentions.2/GatherElements_Cast CPU -- Cast int8/int32 ...er/decoder/layers.1/attentions.2/GatherElements BPU id(7) GatherElements int8/int8 ...der/layers.1/attentions.2/GatherElements_1_Cast CPU -- Cast int8/int32 .../decoder/layers.1/attentions.2/GatherElements_1 BPU id(7) GatherElements int8/int8 ...der/layers.1/attentions.2/GatherElements_2_Cast CPU -- Cast int8/int32 .../decoder/layers.1/attentions.2/GatherElements_2 BPU id(7) GatherElements int8/int8 ...der/layers.1/attentions.2/GatherElements_3_Cast CPU -- Cast int8/int32 .../decoder/layers.1/attentions.2/GatherElements_3 BPU id(7) GatherElements int8/int8 ...ransformer/decoder/layers.1/attentions.2/Mul_14 BPU id(7) HzSElementwiseMul int8/int8 ...ransformer/decoder/layers.1/attentions.2/Mul_15 BPU id(7) HzSElementwiseMul int8/int8 ...ransformer/decoder/layers.1/attentions.2/Add_12 BPU id(7) HzSElementwiseAdd int8/int8 ...ransformer/decoder/layers.1/attentions.2/Mul_16 BPU id(7) HzSElementwiseMul int8/int8 ...ransformer/decoder/layers.1/attentions.2/Add_13 BPU id(7) HzSElementwiseAdd int8/int8 ...ransformer/decoder/layers.1/attentions.2/Mul_17 BPU id(7) HzSElementwiseMul int8/int8 ...ransformer/decoder/layers.1/attentions.2/Add_14 BPU id(7) HzSElementwiseAdd int8/int8 ...former/decoder/layers.1/attentions.2/Reshape_16 BPU id(7) Reshape int8/int8 ...ormer/decoder/layers.1/attentions.2/Transpose_5 BPU id(7) Transpose int8/int8 ...former/decoder/layers.1/attentions.2/Reshape_17 BPU id(7) Reshape int8/int8 ...ransformer/decoder/layers.1/attentions.2/Mul_18 BPU id(7) HzSElementwiseMul int8/int8 ...sformer/decoder/layers.1/attentions.2/ReduceSum BPU id(7) HzSQuantizedReduceSum int8/int8 ...decoder/layers.1/attentions.2/ReduceSum_reshape BPU id(7) Reshape int8/int8 ...ecoder/layers.1/attentions.2/output_proj/MatMul BPU id(7) HzSQuantizedConv int8/int8 .../attentions.2/output_proj/MatMul_reshape_output BPU id(7) Reshape int8/int8 ...ransformer/decoder/layers.1/attentions.2/Add_15 BPU id(7) HzSElementwiseAdd int8/int8 ...transformer/decoder/layers.1/norms.2/ReduceMean BPU id(7) HzSQuantizedReduceMean int8/int8 ...x_head/transformer/decoder/layers.1/norms.2/Sub BPU id(7) HzSElementwiseSub int8/int8 ...x_head/transformer/decoder/layers.1/norms.2/Pow BPU id(7) HzLut int8/int8 ...ansformer/decoder/layers.1/norms.2/ReduceMean_1 BPU id(7) HzSQuantizedReduceMean int8/int8 ...sformer/decoder/layers.1/norms.2/Div_reciprocal BPU id(7) HzLut int8/int8 ...ad/transformer/decoder/layers.1/norms.2/Div_mul BPU id(7) HzSElementwiseMul int8/int8 ...x_head/transformer/decoder/layers.1/norms.2/Mul BPU id(7) HzSElementwiseMul int8/int8 ...head/transformer/decoder/layers.1/norms.2/Add_1 BPU id(7) HzSElementwiseAdd int8/int8 ...layers/layers.0/layers.0.0/MatMul_reshape_input BPU id(7) Reshape int8/int8 ...yers.1/ffns.0/layers/layers.0/layers.0.0/MatMul BPU id(7) HzSQuantizedConv int8/int8 .../decoder/layers.1/ffns.0/layers/layers.1/MatMul BPU id(7) HzSQuantizedConv int8/int8 ....1/ffns.0/layers/layers.1/MatMul_reshape_output BPU id(7) Reshape int8/int8 ...ox_head/transformer/decoder/layers.1/ffns.0/Add BPU id(7) HzSElementwiseAdd int8/int8 ...transformer/decoder/layers.1/norms.3/ReduceMean BPU id(7) HzSQuantizedReduceMean int8/int8 ...x_head/transformer/decoder/layers.1/norms.3/Sub BPU id(7) HzSElementwiseSub int8/int8 ...x_head/transformer/decoder/layers.1/norms.3/Pow BPU id(7) HzLut int8/int8 ...ansformer/decoder/layers.1/norms.3/ReduceMean_1 BPU id(7) HzSQuantizedReduceMean int8/int8 ...sformer/decoder/layers.1/norms.3/Div_reciprocal BPU id(7) HzLut int8/int8 ...ad/transformer/decoder/layers.1/norms.3/Div_mul BPU id(7) HzSElementwiseMul int8/int8 ...x_head/transformer/decoder/layers.1/norms.3/Mul BPU id(7) HzSElementwiseMul int8/int8 ...head/transformer/decoder/layers.1/norms.3/Add_1 BPU id(7) HzSElementwiseAdd int8/int8 ...ox_head/transformer/decoder/Transpose_1_reshape BPU id(7) Reshape int8/int8 .../decoder/reg_branches.1/reg_branches.1.0/MatMul BPU id(7) HzSQuantizedConv int8/int8 .../decoder/reg_branches.1/reg_branches.1.2/MatMul BPU id(7) HzSQuantizedConv int8/int8 ...put_0_reshape_transpose_calibrated_TO_FUSE_RELU BPU id(3) HzSQuantizedConv int8/int8 /pts_bbox_head/transformer/decoder/Clip_4 BPU id(3) HzLut int8/int8 /pts_bbox_head/transformer/decoder/Sub_1 BPU id(3) HzSElementwiseSub int8/int8 /pts_bbox_head/transformer/decoder/Clip_5 BPU id(3) HzLut int8/int8 ..._bbox_head/transformer/decoder/Div_1_reciprocal BPU id(3) HzLut int8/int8 /pts_bbox_head/transformer/decoder/Div_1_mul BPU id(3) HzSElementwiseMul int8/int8 /pts_bbox_head/transformer/decoder/Log_1 BPU id(3) HzLut int8/int8 .../decoder/reg_branches.1/reg_branches.1.4/MatMul BPU id(7) HzSQuantizedConv int8/int8 /pts_bbox_head/transformer/decoder/Sigmoid_1 BPU id(7) HzLut int8/int8 ...ransformer/decoder/Sigmoid_1_output_0_Reshape_0 BPU id(7) Reshape int8/int8 /pts_bbox_head/transformer/decoder/Unsqueeze_2 BPU id(7) Reshape int8/int8 ..._bbox_head/transformer/decoder/layers.2/Reshape BPU id(7) Reshape int8/int8 ...d/transformer/decoder/layers.2/attentions.0/Add BPU id(7) HzSElementwiseAdd int8/int8 ...er/layers.2/attentions.0/Add_output_0_Reshape_0 BPU id(7) Reshape int8/int8 ...ormer/decoder/layers.2/attentions.0/attn/MatMul BPU id(7) HzSQuantizedConv int8/int8 ...ayers.2/attentions.0/attn/MatMul_reshape_output BPU id(7) Reshape int8/int8 ...mer/decoder/layers.2/attentions.0/attn/MatMul_1 BPU id(7) HzSQuantizedConv int8/int8 ...ers.2/attentions.0/attn/MatMul_1_reshape_output BPU id(7) Reshape int8/int8 ...yers.2/attentions.0/attn/MatMul_2_reshape_input BPU id(7) Reshape int8/int8 ...mer/decoder/layers.2/attentions.0/attn/MatMul_2 BPU id(7) HzSQuantizedConv int8/int8 ...ers.2/attentions.0/attn/MatMul_2_reshape_output BPU id(7) Reshape int8/int8 .../decoder/layers.2/attentions.0/attn/Transpose_4 BPU id(7) Transpose int8/int8 ...er/decoder/layers.2/attentions.0/attn/Div_2_mul BPU id(7) HzSElementwiseMul int8/int8 ...rs.2/attentions.0/attn/Div_2_output_0_Transpose BPU id(7) Transpose int8/int8 .../decoder/layers.2/attentions.0/attn/Transpose_5 BPU id(7) Transpose int8/int8 ...mer/decoder/layers.2/attentions.0/attn/MatMul_3 BPU id(7) HzSQuantizedMatmul int8/int8 ...former/decoder/layers.2/attentions.0/attn/Mul_6 BPU id(7) HzSElementwiseMul int8/int16 ...rmer/decoder/layers.2/attentions.0/attn/Softmax CPU -- Softmax float/float ...mer/decoder/layers.2/attentions.0/attn/MatMul_4 BPU id(8) HzSQuantizedMatmul int8/int8 .../decoder/layers.2/attentions.0/attn/Transpose_6 BPU id(8) Transpose int8/int8 ...er/decoder/layers.2/attentions.0/attn/Reshape_3 BPU id(8) Reshape int8/int8 ...sformer/decoder/layers.2/attentions.0/attn/Gemm BPU id(8) HzSQuantizedConv int8/int8 ...transformer/decoder/layers.2/attentions.0/Add_2 BPU id(8) HzSElementwiseAdd int8/int8 ...box_head/transformer/decoder/layers.2/Reshape_4 BPU id(8) Reshape int8/int8 ...transformer/decoder/layers.2/norms.0/ReduceMean BPU id(8) HzSQuantizedReduceMean int8/int8 ...x_head/transformer/decoder/layers.2/norms.0/Sub BPU id(8) HzSElementwiseSub int8/int8 ...x_head/transformer/decoder/layers.2/norms.0/Pow BPU id(8) HzLut int8/int8 ...ansformer/decoder/layers.2/norms.0/ReduceMean_1 BPU id(8) HzSQuantizedReduceMean int8/int8 ...sformer/decoder/layers.2/norms.0/Div_reciprocal BPU id(8) HzLut int8/int8 ...ad/transformer/decoder/layers.2/norms.0/Div_mul BPU id(8) HzSElementwiseMul int8/int8 ...x_head/transformer/decoder/layers.2/norms.0/Mul BPU id(8) HzSElementwiseMul int8/int8 ...head/transformer/decoder/layers.2/norms.0/Add_1 BPU id(8) HzSElementwiseAdd int8/int8 ...box_head/transformer/decoder/layers.2/Transpose BPU id(8) Transpose int8/int8 ...box_head/transformer/decoder/layers.2/Reshape_9 BPU id(8) Reshape int8/int8 ...d/transformer/decoder/layers.2/attentions.1/Add BPU id(8) HzSElementwiseAdd int8/int8 ...layers.2/attentions.1/attn/MatMul_reshape_input BPU id(8) Reshape int8/int8 ...ormer/decoder/layers.2/attentions.1/attn/MatMul BPU id(8) HzSQuantizedConv int8/int8 ...ayers.2/attentions.1/attn/MatMul_reshape_output BPU id(8) Reshape int8/int8 ...mer/decoder/layers.2/attentions.1/attn/MatMul_1 BPU id(8) HzSQuantizedConv int8/int8 ...ers.2/attentions.1/attn/MatMul_1_reshape_output BPU id(8) Reshape int8/int8 ...yers.2/attentions.1/attn/MatMul_2_reshape_input BPU id(8) Reshape int8/int8 ...mer/decoder/layers.2/attentions.1/attn/MatMul_2 BPU id(8) HzSQuantizedConv int8/int8 ...ers.2/attentions.1/attn/MatMul_2_reshape_output BPU id(8) Reshape int8/int8 .../decoder/layers.2/attentions.1/attn/Transpose_4 BPU id(8) Transpose int8/int8 ...er/decoder/layers.2/attentions.1/attn/Div_2_mul BPU id(8) HzSElementwiseMul int8/int8 ...rs.2/attentions.1/attn/Div_2_output_0_Transpose BPU id(8) Transpose int8/int8 .../decoder/layers.2/attentions.1/attn/Transpose_5 BPU id(8) Transpose int8/int8 ...mer/decoder/layers.2/attentions.1/attn/MatMul_3 BPU id(8) HzSQuantizedMatmul int8/int32 ...rmer/decoder/layers.2/attentions.1/attn/Softmax CPU -- Softmax float/float ...mer/decoder/layers.2/attentions.1/attn/MatMul_4 BPU id(9) HzSQuantizedMatmul int8/int8 .../decoder/layers.2/attentions.1/attn/Transpose_6 BPU id(9) Transpose int8/int8 ...er/decoder/layers.2/attentions.1/attn/Reshape_3 BPU id(9) Reshape int8/int8 ...sformer/decoder/layers.2/attentions.1/attn/Gemm BPU id(9) HzSQuantizedConv int8/int8 ...transformer/decoder/layers.2/attentions.1/Add_2 BPU id(9) HzSElementwiseAdd int8/int8 ...ox_head/transformer/decoder/layers.2/Reshape_12 BPU id(9) Reshape int8/int8 ...transformer/decoder/layers.2/norms.1/ReduceMean BPU id(9) HzSQuantizedReduceMean int8/int8 ...x_head/transformer/decoder/layers.2/norms.1/Sub BPU id(9) HzSElementwiseSub int8/int8 ...x_head/transformer/decoder/layers.2/norms.1/Pow BPU id(9) HzLut int8/int8 ...ansformer/decoder/layers.2/norms.1/ReduceMean_1 BPU id(9) HzSQuantizedReduceMean int8/int8 ...sformer/decoder/layers.2/norms.1/Div_reciprocal BPU id(9) HzLut int8/int8 ...ad/transformer/decoder/layers.2/norms.1/Div_mul BPU id(9) HzSElementwiseMul int8/int8 ...x_head/transformer/decoder/layers.2/norms.1/Mul BPU id(9) HzSElementwiseMul int8/int8 ...head/transformer/decoder/layers.2/norms.1/Add_1 BPU id(9) HzSElementwiseAdd int8/int8 ...yers.2/norms.1/Add_1_output_0_reshape_Transpose BPU id(9) Transpose int8/int8 ...coder/layers.2/norms.1/Add_1_output_0_Reshape_0 BPU id(9) Reshape int8/int8 ...d/transformer/decoder/layers.2/attentions.2/Add BPU id(9) HzSQuantizedConv int8/int8 ...er/layers.2/attentions.2/Add_output_0_Reshape_0 BPU id(9) Reshape int8/int8 ...decoder/layers.2/attentions.2/value_proj/MatMul BPU id(3) HzSQuantizedConv int8/int8 ...tentions.2/value_proj/MatMul_output_0_Reshape_0 BPU id(3) Reshape int8/int8 ...r/layers.2/attentions.2/sampling_offsets/MatMul BPU id(9) HzSQuantizedConv int8/int8 ...ntions.2/sampling_offsets/MatMul_reshape_output BPU id(9) Reshape int8/int8 .../layers.2/attentions.2/attention_weights/MatMul BPU id(9) HzSQuantizedConv int8/int32 ...tions.2/attention_weights/MatMul_reshape_output CPU -- Reshape float/float ...ansformer/decoder/layers.2/attentions.2/Softmax CPU -- Softmax float/float ...sformer/decoder/layers.2/attentions.2/Reshape_3 BPU id(11) Reshape int8/int8 ...ansformer/decoder/layers.2/attentions.2/Div_mul BPU id(9) HzSElementwiseMul int8/int16 ...transformer/decoder/layers.2/attentions.2/Add_1 CPU -- Add float/float ...d/transformer/decoder/layers.2/attentions.2/Mul BPU id(10) HzSElementwiseMul int8/int8 ...d/transformer/decoder/layers.2/attentions.2/Sub BPU id(10) HzSElementwiseSub int8/int8 ...er/decoder/layers.2/attentions.2/Gather_reshape BPU id(10) Reshape int8/int8 ...ormer/decoder/layers.2/attentions.2/Transpose_3 BPU id(10) Transpose int8/int8 ...sformer/decoder/layers.2/attentions.2/Reshape_6 BPU id(10) Reshape int8/int8 ...nsformer/decoder/layers.2/attentions.2/Gather_1 BPU id(10) Gather int8/int8 ...nsformer/decoder/layers.2/attentions.2/Gather_2 BPU id(10) Gather int8/int8 ...transformer/decoder/layers.2/attentions.2/Add_2 BPU id(10) HzSElementwiseAdd int8/int8 ...transformer/decoder/layers.2/attentions.2/Mul_1 BPU id(10) HzSElementwiseMul int8/int8 ...transformer/decoder/layers.2/attentions.2/Sub_1 BPU id(10) HzSElementwiseSub int8/int8 ...sformer/decoder/layers.2/attentions.2/Div_1_mul BPU id(10) HzSElementwiseMul int8/int8 .../layers.2/attentions.2/Div_1_output_0_Reshape_0 BPU id(10) Reshape int8/int8 ...transformer/decoder/layers.2/attentions.2/Add_3 BPU id(10) HzSElementwiseAdd int8/int8 ...transformer/decoder/layers.2/attentions.2/Mul_2 BPU id(10) HzSElementwiseMul int8/int8 ...transformer/decoder/layers.2/attentions.2/Sub_2 BPU id(10) HzSElementwiseSub int8/int8 ...sformer/decoder/layers.2/attentions.2/Div_2_mul BPU id(10) HzSElementwiseMul int8/int8 .../layers.2/attentions.2/Div_2_output_0_Reshape_0 BPU id(10) Reshape int8/int8 ...transformer/decoder/layers.2/attentions.2/Floor BPU id(10) HzLut int8/int8 .../transformer/decoder/layers.2/attentions.2/Cast CPU -- Cast float/int8 ...ansformer/decoder/layers.2/attentions.2/Floor_1 BPU id(10) HzLut int8/int8 ...ransformer/decoder/layers.2/attentions.2/Cast_1 CPU -- Cast float/int8 ...transformer/decoder/layers.2/attentions.2/Add_4 CPU -- Add int8/int8 ...transformer/decoder/layers.2/attentions.2/Add_5 CPU -- Add int8/int8 ...ransformer/decoder/layers.2/attentions.2/Cast_2 CPU -- Cast int8/float ...transformer/decoder/layers.2/attentions.2/Sub_3 BPU id(11) HzSElementwiseSub int8/int8 ...ransformer/decoder/layers.2/attentions.2/Cast_3 CPU -- Cast int8/float ...transformer/decoder/layers.2/attentions.2/Sub_4 BPU id(11) HzSElementwiseSub int8/int8 ...transformer/decoder/layers.2/attentions.2/Mul_3 BPU id(11) HzSElementwiseMul int8/int8 ...rmer/decoder/layers.2/attentions.2/Unsqueeze_25 BPU id(11) Reshape int8/int8 ...ransformer/decoder/layers.2/attentions.2/Cast_4 CPU -- Cast int8/float ...transformer/decoder/layers.2/attentions.2/Sub_5 BPU id(11) HzSElementwiseSub int8/int8 ...transformer/decoder/layers.2/attentions.2/Mul_4 BPU id(11) HzSElementwiseMul int8/int8 ...rmer/decoder/layers.2/attentions.2/Unsqueeze_26 BPU id(11) Reshape int8/int8 ...ransformer/decoder/layers.2/attentions.2/Cast_5 CPU -- Cast int8/float ...transformer/decoder/layers.2/attentions.2/Sub_6 BPU id(11) HzSElementwiseSub int8/int8 ...transformer/decoder/layers.2/attentions.2/Mul_5 BPU id(11) HzSElementwiseMul int8/int8 ...rmer/decoder/layers.2/attentions.2/Unsqueeze_27 BPU id(11) Reshape int8/int8 ...transformer/decoder/layers.2/attentions.2/Mul_6 BPU id(11) HzSElementwiseMul int8/int8 ...rmer/decoder/layers.2/attentions.2/Unsqueeze_28 BPU id(11) Reshape int8/int8 ...d/transformer/decoder/layers.2/attentions.2/Pad BPU id(3) HzPad int8/int8 ...transformer/decoder/layers.2/attentions.2/Add_6 CPU -- Add int8/int8 ...transformer/decoder/layers.2/attentions.2/Add_7 CPU -- Add int8/int8 .../transformer/decoder/layers.2/attentions.2/Less CPU -- Less int8/int8 ...former/decoder/layers.2/attentions.2/Where_cast CPU -- Cast int8/int8 ...ormer/decoder/layers.2/attentions.2/Where_mul_x CPU -- Mul int8/int8 ...ormer/decoder/layers.2/attentions.2/Where_equal CPU -- Equal int8/int8 ...r/layers.2/attentions.2/Where_equal_output_cast CPU -- Cast int8/int8 ...ormer/decoder/layers.2/attentions.2/Where_mul_y CPU -- Mul int8/int8 ...sformer/decoder/layers.2/attentions.2/Where_add CPU -- Add int8/int8 ...ansformer/decoder/layers.2/attentions.2/Greater CPU -- Greater int8/int8 ...rmer/decoder/layers.2/attentions.2/Where_1_cast CPU -- Cast int8/int8 ...mer/decoder/layers.2/attentions.2/Where_1_mul_x CPU -- Mul int8/int8 ...mer/decoder/layers.2/attentions.2/Where_1_equal CPU -- Equal int8/int8 ...layers.2/attentions.2/Where_1_equal_output_cast CPU -- Cast int8/int8 ...mer/decoder/layers.2/attentions.2/Where_1_mul_y CPU -- Mul int8/int8 ...ormer/decoder/layers.2/attentions.2/Where_1_add CPU -- Add int8/int8 ...ransformer/decoder/layers.2/attentions.2/Less_1 CPU -- Less int8/int8 ...rmer/decoder/layers.2/attentions.2/Where_2_cast CPU -- Cast int8/int8 ...mer/decoder/layers.2/attentions.2/Where_2_mul_x CPU -- Mul int8/int8 ...mer/decoder/layers.2/attentions.2/Where_2_equal CPU -- Equal int8/int8 ...layers.2/attentions.2/Where_2_equal_output_cast CPU -- Cast int8/int8 ...mer/decoder/layers.2/attentions.2/Where_2_mul_y CPU -- Mul int8/int8 ...ormer/decoder/layers.2/attentions.2/Where_2_add CPU -- Add int8/int8 ...sformer/decoder/layers.2/attentions.2/Greater_1 CPU -- Greater int8/int8 ...rmer/decoder/layers.2/attentions.2/Where_3_cast CPU -- Cast int8/int8 ...mer/decoder/layers.2/attentions.2/Where_3_mul_x CPU -- Mul int8/int8 ...mer/decoder/layers.2/attentions.2/Where_3_equal CPU -- Equal int8/int8 ...layers.2/attentions.2/Where_3_equal_output_cast CPU -- Cast int8/int8 ...mer/decoder/layers.2/attentions.2/Where_3_mul_y CPU -- Mul int8/int8 ...ormer/decoder/layers.2/attentions.2/Where_3_add CPU -- Add int8/int8 ...ransformer/decoder/layers.2/attentions.2/Less_2 CPU -- Less int8/int8 ...rmer/decoder/layers.2/attentions.2/Where_4_cast CPU -- Cast int8/int8 ...mer/decoder/layers.2/attentions.2/Where_4_mul_x CPU -- Mul int8/int8 ...mer/decoder/layers.2/attentions.2/Where_4_equal CPU -- Equal int8/int8 ...layers.2/attentions.2/Where_4_equal_output_cast CPU -- Cast int8/int8 ...mer/decoder/layers.2/attentions.2/Where_4_mul_y CPU -- Mul int8/int8 ...ormer/decoder/layers.2/attentions.2/Where_4_add CPU -- Add int8/int8 ...sformer/decoder/layers.2/attentions.2/Greater_2 CPU -- Greater int8/int8 ...rmer/decoder/layers.2/attentions.2/Where_5_cast CPU -- Cast int8/int8 ...mer/decoder/layers.2/attentions.2/Where_5_mul_x CPU -- Mul int8/int8 ...mer/decoder/layers.2/attentions.2/Where_5_equal CPU -- Equal int8/int8 ...layers.2/attentions.2/Where_5_equal_output_cast CPU -- Cast int8/int8 ...mer/decoder/layers.2/attentions.2/Where_5_mul_y CPU -- Mul int8/int8 ...ormer/decoder/layers.2/attentions.2/Where_5_add CPU -- Add int8/int8 ...ransformer/decoder/layers.2/attentions.2/Less_3 CPU -- Less int8/int8 ...rmer/decoder/layers.2/attentions.2/Where_6_cast CPU -- Cast int8/int8 ...mer/decoder/layers.2/attentions.2/Where_6_mul_x CPU -- Mul int8/int8 ...mer/decoder/layers.2/attentions.2/Where_6_equal CPU -- Equal int8/int8 ...layers.2/attentions.2/Where_6_equal_output_cast CPU -- Cast int8/int8 ...mer/decoder/layers.2/attentions.2/Where_6_mul_y CPU -- Mul int8/int8 ...ormer/decoder/layers.2/attentions.2/Where_6_add CPU -- Add int8/int8 ...sformer/decoder/layers.2/attentions.2/Greater_3 CPU -- Greater int8/int8 ...rmer/decoder/layers.2/attentions.2/Where_7_cast CPU -- Cast int8/int8 ...mer/decoder/layers.2/attentions.2/Where_7_mul_x CPU -- Mul int8/int8 ...mer/decoder/layers.2/attentions.2/Where_7_equal CPU -- Equal int8/int8 ...layers.2/attentions.2/Where_7_equal_output_cast CPU -- Cast int8/int8 ...mer/decoder/layers.2/attentions.2/Where_7_mul_y CPU -- Mul int8/int8 ...ormer/decoder/layers.2/attentions.2/Where_7_add CPU -- Add int8/int8 ...former/decoder/layers.2/attentions.2/Reshape_11 BPU id(3) Reshape int8/int8 ...transformer/decoder/layers.2/attentions.2/Mul_8 CPU -- Mul int8/int8 ...transformer/decoder/layers.2/attentions.2/Add_8 CPU -- Add int8/int8 ...rmer/decoder/layers.2/attentions.2/Unsqueeze_32 CPU -- Reshape int8/int8 ...ransformer/decoder/layers.2/attentions.2/Expand CPU -- Expand int8/int8 ...ransformer/decoder/layers.2/attentions.2/Mul_10 CPU -- Mul int8/int8 ...transformer/decoder/layers.2/attentions.2/Add_9 CPU -- Add int8/int8 ...rmer/decoder/layers.2/attentions.2/Unsqueeze_36 CPU -- Reshape int8/int8 ...nsformer/decoder/layers.2/attentions.2/Expand_1 CPU -- Expand int8/int8 ...ransformer/decoder/layers.2/attentions.2/Add_10 CPU -- Add int8/int8 ...rmer/decoder/layers.2/attentions.2/Unsqueeze_40 CPU -- Reshape int8/int8 ...nsformer/decoder/layers.2/attentions.2/Expand_2 CPU -- Expand int8/int8 ...ransformer/decoder/layers.2/attentions.2/Add_11 CPU -- Add int8/int8 ...rmer/decoder/layers.2/attentions.2/Unsqueeze_44 CPU -- Reshape int8/int8 ...nsformer/decoder/layers.2/attentions.2/Expand_3 CPU -- Expand int8/int8 ...coder/layers.2/attentions.2/GatherElements_Cast CPU -- Cast int8/int32 ...er/decoder/layers.2/attentions.2/GatherElements BPU id(11) GatherElements int8/int8 ...der/layers.2/attentions.2/GatherElements_1_Cast CPU -- Cast int8/int32 .../decoder/layers.2/attentions.2/GatherElements_1 BPU id(11) GatherElements int8/int8 ...der/layers.2/attentions.2/GatherElements_2_Cast CPU -- Cast int8/int32 .../decoder/layers.2/attentions.2/GatherElements_2 BPU id(11) GatherElements int8/int8 ...der/layers.2/attentions.2/GatherElements_3_Cast CPU -- Cast int8/int32 .../decoder/layers.2/attentions.2/GatherElements_3 BPU id(11) GatherElements int8/int8 ...ransformer/decoder/layers.2/attentions.2/Mul_14 BPU id(11) HzSElementwiseMul int8/int8 ...ransformer/decoder/layers.2/attentions.2/Mul_15 BPU id(11) HzSElementwiseMul int8/int8 ...ransformer/decoder/layers.2/attentions.2/Add_12 BPU id(11) HzSElementwiseAdd int8/int8 ...ransformer/decoder/layers.2/attentions.2/Mul_16 BPU id(11) HzSElementwiseMul int8/int8 ...ransformer/decoder/layers.2/attentions.2/Add_13 BPU id(11) HzSElementwiseAdd int8/int8 ...ransformer/decoder/layers.2/attentions.2/Mul_17 BPU id(11) HzSElementwiseMul int8/int8 ...ransformer/decoder/layers.2/attentions.2/Add_14 BPU id(11) HzSElementwiseAdd int8/int8 ...former/decoder/layers.2/attentions.2/Reshape_16 BPU id(11) Reshape int8/int8 ...ormer/decoder/layers.2/attentions.2/Transpose_5 BPU id(11) Transpose int8/int8 ...former/decoder/layers.2/attentions.2/Reshape_17 BPU id(11) Reshape int8/int8 ...ransformer/decoder/layers.2/attentions.2/Mul_18 BPU id(11) HzSElementwiseMul int8/int8 ...sformer/decoder/layers.2/attentions.2/ReduceSum BPU id(11) HzSQuantizedReduceSum int8/int8 ...decoder/layers.2/attentions.2/ReduceSum_reshape BPU id(11) Reshape int8/int8 ...ecoder/layers.2/attentions.2/output_proj/MatMul BPU id(11) HzSQuantizedConv int8/int8 .../attentions.2/output_proj/MatMul_reshape_output BPU id(11) Reshape int8/int8 ...ransformer/decoder/layers.2/attentions.2/Add_15 BPU id(11) HzSElementwiseAdd int8/int8 ...transformer/decoder/layers.2/norms.2/ReduceMean BPU id(11) HzSQuantizedReduceMean int8/int8 ...x_head/transformer/decoder/layers.2/norms.2/Sub BPU id(11) HzSElementwiseSub int8/int8 ...x_head/transformer/decoder/layers.2/norms.2/Pow BPU id(11) HzLut int8/int8 ...ansformer/decoder/layers.2/norms.2/ReduceMean_1 BPU id(11) HzSQuantizedReduceMean int8/int8 ...sformer/decoder/layers.2/norms.2/Div_reciprocal BPU id(11) HzLut int8/int8 ...ad/transformer/decoder/layers.2/norms.2/Div_mul BPU id(11) HzSElementwiseMul int8/int8 ...x_head/transformer/decoder/layers.2/norms.2/Mul BPU id(11) HzSElementwiseMul int8/int8 ...head/transformer/decoder/layers.2/norms.2/Add_1 BPU id(11) HzSElementwiseAdd int8/int8 ...layers/layers.0/layers.0.0/MatMul_reshape_input BPU id(11) Reshape int8/int8 ...yers.2/ffns.0/layers/layers.0/layers.0.0/MatMul BPU id(11) HzSQuantizedConv int8/int8 .../decoder/layers.2/ffns.0/layers/layers.1/MatMul BPU id(11) HzSQuantizedConv int8/int8 ....2/ffns.0/layers/layers.1/MatMul_reshape_output BPU id(11) Reshape int8/int8 ...ox_head/transformer/decoder/layers.2/ffns.0/Add BPU id(11) HzSElementwiseAdd int8/int8 ...transformer/decoder/layers.2/norms.3/ReduceMean BPU id(11) HzSQuantizedReduceMean int8/int8 ...x_head/transformer/decoder/layers.2/norms.3/Sub BPU id(11) HzSElementwiseSub int8/int8 ...x_head/transformer/decoder/layers.2/norms.3/Pow BPU id(11) HzLut int8/int8 ...ansformer/decoder/layers.2/norms.3/ReduceMean_1 BPU id(11) HzSQuantizedReduceMean int8/int8 ...sformer/decoder/layers.2/norms.3/Div_reciprocal BPU id(11) HzLut int8/int8 ...ad/transformer/decoder/layers.2/norms.3/Div_mul BPU id(11) HzSElementwiseMul int8/int8 ...x_head/transformer/decoder/layers.2/norms.3/Mul BPU id(11) HzSElementwiseMul int8/int8 ...head/transformer/decoder/layers.2/norms.3/Add_1 BPU id(11) HzSElementwiseAdd int8/int8 ...ox_head/transformer/decoder/Transpose_2_reshape BPU id(11) Reshape int8/int8 .../decoder/reg_branches.2/reg_branches.2.0/MatMul BPU id(11) HzSQuantizedConv int8/int8 .../decoder/reg_branches.2/reg_branches.2.2/MatMul BPU id(11) HzSQuantizedConv int8/int8 ...put_0_reshape_transpose_calibrated_TO_FUSE_RELU BPU id(7) HzSQuantizedConv int8/int8 /pts_bbox_head/transformer/decoder/Clip_7 BPU id(7) HzLut int8/int8 /pts_bbox_head/transformer/decoder/Sub_2 BPU id(7) HzSElementwiseSub int8/int8 /pts_bbox_head/transformer/decoder/Clip_8 BPU id(7) HzLut int8/int8 ..._bbox_head/transformer/decoder/Div_2_reciprocal BPU id(7) HzLut int8/int8 /pts_bbox_head/transformer/decoder/Div_2_mul BPU id(7) HzSElementwiseMul int8/int8 /pts_bbox_head/transformer/decoder/Log_2 BPU id(7) HzLut int8/int8 .../decoder/reg_branches.2/reg_branches.2.4/MatMul BPU id(11) HzSQuantizedConv int8/int8 /pts_bbox_head/transformer/decoder/Sigmoid_2 BPU id(11) HzLut int8/int8 ...ransformer/decoder/Sigmoid_2_output_0_Reshape_0 BPU id(11) Reshape int8/int8 /pts_bbox_head/transformer/decoder/Unsqueeze_3 BPU id(11) Reshape int8/int8 ..._bbox_head/transformer/decoder/layers.3/Reshape BPU id(11) Reshape int8/int8 ...d/transformer/decoder/layers.3/attentions.0/Add BPU id(11) HzSElementwiseAdd int8/int8 ...er/layers.3/attentions.0/Add_output_0_Reshape_0 BPU id(11) Reshape int8/int8 ...ormer/decoder/layers.3/attentions.0/attn/MatMul BPU id(11) HzSQuantizedConv int8/int8 ...ayers.3/attentions.0/attn/MatMul_reshape_output BPU id(11) Reshape int8/int8 ...mer/decoder/layers.3/attentions.0/attn/MatMul_1 BPU id(11) HzSQuantizedConv int8/int8 ...ers.3/attentions.0/attn/MatMul_1_reshape_output BPU id(11) Reshape int8/int8 ...yers.3/attentions.0/attn/MatMul_2_reshape_input BPU id(11) Reshape int8/int8 ...mer/decoder/layers.3/attentions.0/attn/MatMul_2 BPU id(11) HzSQuantizedConv int8/int8 ...ers.3/attentions.0/attn/MatMul_2_reshape_output BPU id(11) Reshape int8/int8 .../decoder/layers.3/attentions.0/attn/Transpose_4 BPU id(11) Transpose int8/int8 ...er/decoder/layers.3/attentions.0/attn/Div_2_mul BPU id(11) HzSElementwiseMul int8/int8 ...rs.3/attentions.0/attn/Div_2_output_0_Transpose BPU id(11) Transpose int8/int8 .../decoder/layers.3/attentions.0/attn/Transpose_5 BPU id(11) Transpose int8/int8 ...mer/decoder/layers.3/attentions.0/attn/MatMul_3 BPU id(11) HzSQuantizedMatmul int8/int8 ...former/decoder/layers.3/attentions.0/attn/Mul_6 BPU id(11) HzSElementwiseMul int8/int16 ...rmer/decoder/layers.3/attentions.0/attn/Softmax CPU -- Softmax float/float ...mer/decoder/layers.3/attentions.0/attn/MatMul_4 BPU id(12) HzSQuantizedMatmul int8/int8 .../decoder/layers.3/attentions.0/attn/Transpose_6 BPU id(12) Transpose int8/int8 ...er/decoder/layers.3/attentions.0/attn/Reshape_3 BPU id(12) Reshape int8/int8 ...sformer/decoder/layers.3/attentions.0/attn/Gemm BPU id(12) HzSQuantizedConv int8/int8 ...transformer/decoder/layers.3/attentions.0/Add_2 BPU id(12) HzSElementwiseAdd int8/int8 ...box_head/transformer/decoder/layers.3/Reshape_4 BPU id(12) Reshape int8/int8 ...transformer/decoder/layers.3/norms.0/ReduceMean BPU id(12) HzSQuantizedReduceMean int8/int8 ...x_head/transformer/decoder/layers.3/norms.0/Sub BPU id(12) HzSElementwiseSub int8/int8 ...x_head/transformer/decoder/layers.3/norms.0/Pow BPU id(12) HzLut int8/int8 ...ansformer/decoder/layers.3/norms.0/ReduceMean_1 BPU id(12) HzSQuantizedReduceMean int8/int8 ...sformer/decoder/layers.3/norms.0/Div_reciprocal BPU id(12) HzLut int8/int8 ...ad/transformer/decoder/layers.3/norms.0/Div_mul BPU id(12) HzSElementwiseMul int8/int8 ...x_head/transformer/decoder/layers.3/norms.0/Mul BPU id(12) HzSElementwiseMul int8/int8 ...head/transformer/decoder/layers.3/norms.0/Add_1 BPU id(12) HzSElementwiseAdd int8/int8 ...box_head/transformer/decoder/layers.3/Transpose BPU id(12) Transpose int8/int8 ...box_head/transformer/decoder/layers.3/Reshape_9 BPU id(12) Reshape int8/int8 ...d/transformer/decoder/layers.3/attentions.1/Add BPU id(12) HzSElementwiseAdd int8/int8 ...layers.3/attentions.1/attn/MatMul_reshape_input BPU id(12) Reshape int8/int8 ...ormer/decoder/layers.3/attentions.1/attn/MatMul BPU id(12) HzSQuantizedConv int8/int8 ...ayers.3/attentions.1/attn/MatMul_reshape_output BPU id(12) Reshape int8/int8 ...mer/decoder/layers.3/attentions.1/attn/MatMul_1 BPU id(12) HzSQuantizedConv int8/int8 ...ers.3/attentions.1/attn/MatMul_1_reshape_output BPU id(12) Reshape int8/int8 ...yers.3/attentions.1/attn/MatMul_2_reshape_input BPU id(12) Reshape int8/int8 ...mer/decoder/layers.3/attentions.1/attn/MatMul_2 BPU id(12) HzSQuantizedConv int8/int8 ...ers.3/attentions.1/attn/MatMul_2_reshape_output BPU id(12) Reshape int8/int8 .../decoder/layers.3/attentions.1/attn/Transpose_4 BPU id(12) Transpose int8/int8 ...er/decoder/layers.3/attentions.1/attn/Div_2_mul BPU id(12) HzSElementwiseMul int8/int8 ...rs.3/attentions.1/attn/Div_2_output_0_Transpose BPU id(12) Transpose int8/int8 .../decoder/layers.3/attentions.1/attn/Transpose_5 BPU id(12) Transpose int8/int8 ...mer/decoder/layers.3/attentions.1/attn/MatMul_3 BPU id(12) HzSQuantizedMatmul int8/int32 ...rmer/decoder/layers.3/attentions.1/attn/Softmax CPU -- Softmax float/float ...mer/decoder/layers.3/attentions.1/attn/MatMul_4 BPU id(13) HzSQuantizedMatmul int8/int8 .../decoder/layers.3/attentions.1/attn/Transpose_6 BPU id(13) Transpose int8/int8 ...er/decoder/layers.3/attentions.1/attn/Reshape_3 BPU id(13) Reshape int8/int8 ...sformer/decoder/layers.3/attentions.1/attn/Gemm BPU id(13) HzSQuantizedConv int8/int8 ...transformer/decoder/layers.3/attentions.1/Add_2 BPU id(13) HzSElementwiseAdd int8/int8 ...ox_head/transformer/decoder/layers.3/Reshape_12 BPU id(13) Reshape int8/int8 ...transformer/decoder/layers.3/norms.1/ReduceMean BPU id(13) HzSQuantizedReduceMean int8/int8 ...x_head/transformer/decoder/layers.3/norms.1/Sub BPU id(13) HzSElementwiseSub int8/int8 ...x_head/transformer/decoder/layers.3/norms.1/Pow BPU id(13) HzLut int8/int8 ...ansformer/decoder/layers.3/norms.1/ReduceMean_1 BPU id(13) HzSQuantizedReduceMean int8/int8 ...sformer/decoder/layers.3/norms.1/Div_reciprocal BPU id(13) HzLut int8/int8 ...ad/transformer/decoder/layers.3/norms.1/Div_mul BPU id(13) HzSElementwiseMul int8/int8 ...x_head/transformer/decoder/layers.3/norms.1/Mul BPU id(13) HzSElementwiseMul int8/int8 ...head/transformer/decoder/layers.3/norms.1/Add_1 BPU id(13) HzSElementwiseAdd int8/int8 ...yers.3/norms.1/Add_1_output_0_reshape_Transpose BPU id(13) Transpose int8/int8 ...coder/layers.3/norms.1/Add_1_output_0_Reshape_0 BPU id(13) Reshape int8/int8 ...d/transformer/decoder/layers.3/attentions.2/Add BPU id(13) HzSQuantizedConv int8/int8 ...er/layers.3/attentions.2/Add_output_0_Reshape_0 BPU id(13) Reshape int8/int8 ...decoder/layers.3/attentions.2/value_proj/MatMul BPU id(3) HzSQuantizedConv int8/int8 ...tentions.2/value_proj/MatMul_output_0_Reshape_0 BPU id(3) Reshape int8/int8 ...r/layers.3/attentions.2/sampling_offsets/MatMul BPU id(13) HzSQuantizedConv int8/int8 ...ntions.2/sampling_offsets/MatMul_reshape_output BPU id(13) Reshape int8/int8 .../layers.3/attentions.2/attention_weights/MatMul BPU id(13) HzSQuantizedConv int8/int32 ...tions.2/attention_weights/MatMul_reshape_output CPU -- Reshape float/float ...ansformer/decoder/layers.3/attentions.2/Softmax CPU -- Softmax float/float ...sformer/decoder/layers.3/attentions.2/Reshape_3 BPU id(15) Reshape int8/int8 ...ansformer/decoder/layers.3/attentions.2/Div_mul BPU id(13) HzSElementwiseMul int8/int16 ...transformer/decoder/layers.3/attentions.2/Add_1 CPU -- Add float/float ...d/transformer/decoder/layers.3/attentions.2/Mul BPU id(14) HzSElementwiseMul int8/int8 ...d/transformer/decoder/layers.3/attentions.2/Sub BPU id(14) HzSElementwiseSub int8/int8 ...er/decoder/layers.3/attentions.2/Gather_reshape BPU id(14) Reshape int8/int8 ...ormer/decoder/layers.3/attentions.2/Transpose_3 BPU id(14) Transpose int8/int8 ...sformer/decoder/layers.3/attentions.2/Reshape_6 BPU id(14) Reshape int8/int8 ...nsformer/decoder/layers.3/attentions.2/Gather_1 BPU id(14) Gather int8/int8 ...nsformer/decoder/layers.3/attentions.2/Gather_2 BPU id(14) Gather int8/int8 ...transformer/decoder/layers.3/attentions.2/Add_2 BPU id(14) HzSElementwiseAdd int8/int8 ...transformer/decoder/layers.3/attentions.2/Mul_1 BPU id(14) HzSElementwiseMul int8/int8 ...transformer/decoder/layers.3/attentions.2/Sub_1 BPU id(14) HzSElementwiseSub int8/int8 ...sformer/decoder/layers.3/attentions.2/Div_1_mul BPU id(14) HzSElementwiseMul int8/int8 .../layers.3/attentions.2/Div_1_output_0_Reshape_0 BPU id(14) Reshape int8/int8 ...transformer/decoder/layers.3/attentions.2/Add_3 BPU id(14) HzSElementwiseAdd int8/int8 ...transformer/decoder/layers.3/attentions.2/Mul_2 BPU id(14) HzSElementwiseMul int8/int8 ...transformer/decoder/layers.3/attentions.2/Sub_2 BPU id(14) HzSElementwiseSub int8/int8 ...sformer/decoder/layers.3/attentions.2/Div_2_mul BPU id(14) HzSElementwiseMul int8/int8 .../layers.3/attentions.2/Div_2_output_0_Reshape_0 BPU id(14) Reshape int8/int8 ...transformer/decoder/layers.3/attentions.2/Floor BPU id(14) HzLut int8/int8 .../transformer/decoder/layers.3/attentions.2/Cast CPU -- Cast float/int8 ...ansformer/decoder/layers.3/attentions.2/Floor_1 BPU id(14) HzLut int8/int8 ...ransformer/decoder/layers.3/attentions.2/Cast_1 CPU -- Cast float/int8 ...transformer/decoder/layers.3/attentions.2/Add_4 CPU -- Add int8/int8 ...transformer/decoder/layers.3/attentions.2/Add_5 CPU -- Add int8/int8 ...ransformer/decoder/layers.3/attentions.2/Cast_2 CPU -- Cast int8/float ...transformer/decoder/layers.3/attentions.2/Sub_3 BPU id(15) HzSElementwiseSub int8/int8 ...ransformer/decoder/layers.3/attentions.2/Cast_3 CPU -- Cast int8/float ...transformer/decoder/layers.3/attentions.2/Sub_4 BPU id(15) HzSElementwiseSub int8/int8 ...transformer/decoder/layers.3/attentions.2/Mul_3 BPU id(15) HzSElementwiseMul int8/int8 ...rmer/decoder/layers.3/attentions.2/Unsqueeze_25 BPU id(15) Reshape int8/int8 ...ransformer/decoder/layers.3/attentions.2/Cast_4 CPU -- Cast int8/float ...transformer/decoder/layers.3/attentions.2/Sub_5 BPU id(15) HzSElementwiseSub int8/int8 ...transformer/decoder/layers.3/attentions.2/Mul_4 BPU id(15) HzSElementwiseMul int8/int8 ...rmer/decoder/layers.3/attentions.2/Unsqueeze_26 BPU id(15) Reshape int8/int8 ...ransformer/decoder/layers.3/attentions.2/Cast_5 CPU -- Cast int8/float ...transformer/decoder/layers.3/attentions.2/Sub_6 BPU id(15) HzSElementwiseSub int8/int8 ...transformer/decoder/layers.3/attentions.2/Mul_5 BPU id(15) HzSElementwiseMul int8/int8 ...rmer/decoder/layers.3/attentions.2/Unsqueeze_27 BPU id(15) Reshape int8/int8 ...transformer/decoder/layers.3/attentions.2/Mul_6 BPU id(15) HzSElementwiseMul int8/int8 ...rmer/decoder/layers.3/attentions.2/Unsqueeze_28 BPU id(15) Reshape int8/int8 ...d/transformer/decoder/layers.3/attentions.2/Pad BPU id(3) HzPad int8/int8 ...transformer/decoder/layers.3/attentions.2/Add_6 CPU -- Add int8/int8 ...transformer/decoder/layers.3/attentions.2/Add_7 CPU -- Add int8/int8 .../transformer/decoder/layers.3/attentions.2/Less CPU -- Less int8/int8 ...former/decoder/layers.3/attentions.2/Where_cast CPU -- Cast int8/int8 ...ormer/decoder/layers.3/attentions.2/Where_mul_x CPU -- Mul int8/int8 ...ormer/decoder/layers.3/attentions.2/Where_equal CPU -- Equal int8/int8 ...r/layers.3/attentions.2/Where_equal_output_cast CPU -- Cast int8/int8 ...ormer/decoder/layers.3/attentions.2/Where_mul_y CPU -- Mul int8/int8 ...sformer/decoder/layers.3/attentions.2/Where_add CPU -- Add int8/int8 ...ansformer/decoder/layers.3/attentions.2/Greater CPU -- Greater int8/int8 ...rmer/decoder/layers.3/attentions.2/Where_1_cast CPU -- Cast int8/int8 ...mer/decoder/layers.3/attentions.2/Where_1_mul_x CPU -- Mul int8/int8 ...mer/decoder/layers.3/attentions.2/Where_1_equal CPU -- Equal int8/int8 ...layers.3/attentions.2/Where_1_equal_output_cast CPU -- Cast int8/int8 ...mer/decoder/layers.3/attentions.2/Where_1_mul_y CPU -- Mul int8/int8 ...ormer/decoder/layers.3/attentions.2/Where_1_add CPU -- Add int8/int8 ...ransformer/decoder/layers.3/attentions.2/Less_1 CPU -- Less int8/int8 ...rmer/decoder/layers.3/attentions.2/Where_2_cast CPU -- Cast int8/int8 ...mer/decoder/layers.3/attentions.2/Where_2_mul_x CPU -- Mul int8/int8 ...mer/decoder/layers.3/attentions.2/Where_2_equal CPU -- Equal int8/int8 ...layers.3/attentions.2/Where_2_equal_output_cast CPU -- Cast int8/int8 ...mer/decoder/layers.3/attentions.2/Where_2_mul_y CPU -- Mul int8/int8 ...ormer/decoder/layers.3/attentions.2/Where_2_add CPU -- Add int8/int8 ...sformer/decoder/layers.3/attentions.2/Greater_1 CPU -- Greater int8/int8 ...rmer/decoder/layers.3/attentions.2/Where_3_cast CPU -- Cast int8/int8 ...mer/decoder/layers.3/attentions.2/Where_3_mul_x CPU -- Mul int8/int8 ...mer/decoder/layers.3/attentions.2/Where_3_equal CPU -- Equal int8/int8 ...layers.3/attentions.2/Where_3_equal_output_cast CPU -- Cast int8/int8 ...mer/decoder/layers.3/attentions.2/Where_3_mul_y CPU -- Mul int8/int8 ...ormer/decoder/layers.3/attentions.2/Where_3_add CPU -- Add int8/int8 ...ransformer/decoder/layers.3/attentions.2/Less_2 CPU -- Less int8/int8 ...rmer/decoder/layers.3/attentions.2/Where_4_cast CPU -- Cast int8/int8 ...mer/decoder/layers.3/attentions.2/Where_4_mul_x CPU -- Mul int8/int8 ...mer/decoder/layers.3/attentions.2/Where_4_equal CPU -- Equal int8/int8 ...layers.3/attentions.2/Where_4_equal_output_cast CPU -- Cast int8/int8 ...mer/decoder/layers.3/attentions.2/Where_4_mul_y CPU -- Mul int8/int8 ...ormer/decoder/layers.3/attentions.2/Where_4_add CPU -- Add int8/int8 ...sformer/decoder/layers.3/attentions.2/Greater_2 CPU -- Greater int8/int8 ...rmer/decoder/layers.3/attentions.2/Where_5_cast CPU -- Cast int8/int8 ...mer/decoder/layers.3/attentions.2/Where_5_mul_x CPU -- Mul int8/int8 ...mer/decoder/layers.3/attentions.2/Where_5_equal CPU -- Equal int8/int8 ...layers.3/attentions.2/Where_5_equal_output_cast CPU -- Cast int8/int8 ...mer/decoder/layers.3/attentions.2/Where_5_mul_y CPU -- Mul int8/int8 ...ormer/decoder/layers.3/attentions.2/Where_5_add CPU -- Add int8/int8 ...ransformer/decoder/layers.3/attentions.2/Less_3 CPU -- Less int8/int8 ...rmer/decoder/layers.3/attentions.2/Where_6_cast CPU -- Cast int8/int8 ...mer/decoder/layers.3/attentions.2/Where_6_mul_x CPU -- Mul int8/int8 ...mer/decoder/layers.3/attentions.2/Where_6_equal CPU -- Equal int8/int8 ...layers.3/attentions.2/Where_6_equal_output_cast CPU -- Cast int8/int8 ...mer/decoder/layers.3/attentions.2/Where_6_mul_y CPU -- Mul int8/int8 ...ormer/decoder/layers.3/attentions.2/Where_6_add CPU -- Add int8/int8 ...sformer/decoder/layers.3/attentions.2/Greater_3 CPU -- Greater int8/int8 ...rmer/decoder/layers.3/attentions.2/Where_7_cast CPU -- Cast int8/int8 ...mer/decoder/layers.3/attentions.2/Where_7_mul_x CPU -- Mul int8/int8 ...mer/decoder/layers.3/attentions.2/Where_7_equal CPU -- Equal int8/int8 ...layers.3/attentions.2/Where_7_equal_output_cast CPU -- Cast int8/int8 ...mer/decoder/layers.3/attentions.2/Where_7_mul_y CPU -- Mul int8/int8 ...ormer/decoder/layers.3/attentions.2/Where_7_add CPU -- Add int8/int8 ...former/decoder/layers.3/attentions.2/Reshape_11 BPU id(3) Reshape int8/int8 ...transformer/decoder/layers.3/attentions.2/Mul_8 CPU -- Mul int8/int8 ...transformer/decoder/layers.3/attentions.2/Add_8 CPU -- Add int8/int8 ...rmer/decoder/layers.3/attentions.2/Unsqueeze_32 CPU -- Reshape int8/int8 ...ransformer/decoder/layers.3/attentions.2/Expand CPU -- Expand int8/int8 ...ransformer/decoder/layers.3/attentions.2/Mul_10 CPU -- Mul int8/int8 ...transformer/decoder/layers.3/attentions.2/Add_9 CPU -- Add int8/int8 ...rmer/decoder/layers.3/attentions.2/Unsqueeze_36 CPU -- Reshape int8/int8 ...nsformer/decoder/layers.3/attentions.2/Expand_1 CPU -- Expand int8/int8 ...ransformer/decoder/layers.3/attentions.2/Add_10 CPU -- Add int8/int8 ...rmer/decoder/layers.3/attentions.2/Unsqueeze_40 CPU -- Reshape int8/int8 ...nsformer/decoder/layers.3/attentions.2/Expand_2 CPU -- Expand int8/int8 ...ransformer/decoder/layers.3/attentions.2/Add_11 CPU -- Add int8/int8 ...rmer/decoder/layers.3/attentions.2/Unsqueeze_44 CPU -- Reshape int8/int8 ...nsformer/decoder/layers.3/attentions.2/Expand_3 CPU -- Expand int8/int8 ...coder/layers.3/attentions.2/GatherElements_Cast CPU -- Cast int8/int32 ...er/decoder/layers.3/attentions.2/GatherElements BPU id(15) GatherElements int8/int8 ...der/layers.3/attentions.2/GatherElements_1_Cast CPU -- Cast int8/int32 .../decoder/layers.3/attentions.2/GatherElements_1 BPU id(15) GatherElements int8/int8 ...der/layers.3/attentions.2/GatherElements_2_Cast CPU -- Cast int8/int32 .../decoder/layers.3/attentions.2/GatherElements_2 BPU id(15) GatherElements int8/int8 ...der/layers.3/attentions.2/GatherElements_3_Cast CPU -- Cast int8/int32 .../decoder/layers.3/attentions.2/GatherElements_3 BPU id(15) GatherElements int8/int8 ...ransformer/decoder/layers.3/attentions.2/Mul_14 BPU id(15) HzSElementwiseMul int8/int8 ...ransformer/decoder/layers.3/attentions.2/Mul_15 BPU id(15) HzSElementwiseMul int8/int8 ...ransformer/decoder/layers.3/attentions.2/Add_12 BPU id(15) HzSElementwiseAdd int8/int8 ...ransformer/decoder/layers.3/attentions.2/Mul_16 BPU id(15) HzSElementwiseMul int8/int8 ...ransformer/decoder/layers.3/attentions.2/Add_13 BPU id(15) HzSElementwiseAdd int8/int8 ...ransformer/decoder/layers.3/attentions.2/Mul_17 BPU id(15) HzSElementwiseMul int8/int8 ...ransformer/decoder/layers.3/attentions.2/Add_14 BPU id(15) HzSElementwiseAdd int8/int8 ...former/decoder/layers.3/attentions.2/Reshape_16 BPU id(15) Reshape int8/int8 ...ormer/decoder/layers.3/attentions.2/Transpose_5 BPU id(15) Transpose int8/int8 ...former/decoder/layers.3/attentions.2/Reshape_17 BPU id(15) Reshape int8/int8 ...ransformer/decoder/layers.3/attentions.2/Mul_18 BPU id(15) HzSElementwiseMul int8/int8 ...sformer/decoder/layers.3/attentions.2/ReduceSum BPU id(15) HzSQuantizedReduceSum int8/int8 ...decoder/layers.3/attentions.2/ReduceSum_reshape BPU id(15) Reshape int8/int8 ...ecoder/layers.3/attentions.2/output_proj/MatMul BPU id(15) HzSQuantizedConv int8/int8 .../attentions.2/output_proj/MatMul_reshape_output BPU id(15) Reshape int8/int8 ...ransformer/decoder/layers.3/attentions.2/Add_15 BPU id(15) HzSElementwiseAdd int8/int8 ...transformer/decoder/layers.3/norms.2/ReduceMean BPU id(15) HzSQuantizedReduceMean int8/int8 ...x_head/transformer/decoder/layers.3/norms.2/Sub BPU id(15) HzSElementwiseSub int8/int8 ...x_head/transformer/decoder/layers.3/norms.2/Pow BPU id(15) HzLut int8/int8 ...ansformer/decoder/layers.3/norms.2/ReduceMean_1 BPU id(15) HzSQuantizedReduceMean int8/int8 ...sformer/decoder/layers.3/norms.2/Div_reciprocal BPU id(15) HzLut int8/int8 ...ad/transformer/decoder/layers.3/norms.2/Div_mul BPU id(15) HzSElementwiseMul int8/int8 ...x_head/transformer/decoder/layers.3/norms.2/Mul BPU id(15) HzSElementwiseMul int8/int8 ...head/transformer/decoder/layers.3/norms.2/Add_1 BPU id(15) HzSElementwiseAdd int8/int8 ...layers/layers.0/layers.0.0/MatMul_reshape_input BPU id(15) Reshape int8/int8 ...yers.3/ffns.0/layers/layers.0/layers.0.0/MatMul BPU id(15) HzSQuantizedConv int8/int8 .../decoder/layers.3/ffns.0/layers/layers.1/MatMul BPU id(15) HzSQuantizedConv int8/int8 ....3/ffns.0/layers/layers.1/MatMul_reshape_output BPU id(15) Reshape int8/int8 ...ox_head/transformer/decoder/layers.3/ffns.0/Add BPU id(15) HzSElementwiseAdd int8/int8 ...transformer/decoder/layers.3/norms.3/ReduceMean BPU id(15) HzSQuantizedReduceMean int8/int8 ...x_head/transformer/decoder/layers.3/norms.3/Sub BPU id(15) HzSElementwiseSub int8/int8 ...x_head/transformer/decoder/layers.3/norms.3/Pow BPU id(15) HzLut int8/int8 ...ansformer/decoder/layers.3/norms.3/ReduceMean_1 BPU id(15) HzSQuantizedReduceMean int8/int8 ...sformer/decoder/layers.3/norms.3/Div_reciprocal BPU id(15) HzLut int8/int8 ...ad/transformer/decoder/layers.3/norms.3/Div_mul BPU id(15) HzSElementwiseMul int8/int8 ...x_head/transformer/decoder/layers.3/norms.3/Mul BPU id(15) HzSElementwiseMul int8/int8 ...head/transformer/decoder/layers.3/norms.3/Add_1 BPU id(15) HzSElementwiseAdd int8/int8 ...ox_head/transformer/decoder/Transpose_3_reshape BPU id(15) Reshape int8/int8 .../decoder/reg_branches.3/reg_branches.3.0/MatMul BPU id(15) HzSQuantizedConv int8/int8 .../decoder/reg_branches.3/reg_branches.3.2/MatMul BPU id(15) HzSQuantizedConv int8/int8 ...put_0_reshape_transpose_calibrated_TO_FUSE_RELU BPU id(11) HzSQuantizedConv int8/int8 /pts_bbox_head/transformer/decoder/Clip_10 BPU id(11) HzLut int8/int8 /pts_bbox_head/transformer/decoder/Sub_3 BPU id(11) HzSElementwiseSub int8/int8 /pts_bbox_head/transformer/decoder/Clip_11 BPU id(11) HzLut int8/int8 ..._bbox_head/transformer/decoder/Div_3_reciprocal BPU id(11) HzLut int8/int8 /pts_bbox_head/transformer/decoder/Div_3_mul BPU id(11) HzSElementwiseMul int8/int8 /pts_bbox_head/transformer/decoder/Log_3 BPU id(11) HzLut int8/int8 .../decoder/reg_branches.3/reg_branches.3.4/MatMul BPU id(15) HzSQuantizedConv int8/int8 /pts_bbox_head/transformer/decoder/Sigmoid_3 BPU id(15) HzLut int8/int8 ...ransformer/decoder/Sigmoid_3_output_0_Reshape_0 BPU id(15) Reshape int8/int8 /pts_bbox_head/transformer/decoder/Unsqueeze_4 BPU id(15) Reshape int8/int8 ..._bbox_head/transformer/decoder/layers.4/Reshape BPU id(15) Reshape int8/int8 ...d/transformer/decoder/layers.4/attentions.0/Add BPU id(15) HzSElementwiseAdd int8/int8 ...er/layers.4/attentions.0/Add_output_0_Reshape_0 BPU id(15) Reshape int8/int8 ...ormer/decoder/layers.4/attentions.0/attn/MatMul BPU id(15) HzSQuantizedConv int8/int8 ...ayers.4/attentions.0/attn/MatMul_reshape_output BPU id(15) Reshape int8/int8 ...mer/decoder/layers.4/attentions.0/attn/MatMul_1 BPU id(15) HzSQuantizedConv int8/int8 ...ers.4/attentions.0/attn/MatMul_1_reshape_output BPU id(15) Reshape int8/int8 ...yers.4/attentions.0/attn/MatMul_2_reshape_input BPU id(15) Reshape int8/int8 ...mer/decoder/layers.4/attentions.0/attn/MatMul_2 BPU id(15) HzSQuantizedConv int8/int8 ...ers.4/attentions.0/attn/MatMul_2_reshape_output BPU id(15) Reshape int8/int8 .../decoder/layers.4/attentions.0/attn/Transpose_4 BPU id(15) Transpose int8/int8 ...er/decoder/layers.4/attentions.0/attn/Div_2_mul BPU id(15) HzSElementwiseMul int8/int8 ...rs.4/attentions.0/attn/Div_2_output_0_Transpose BPU id(15) Transpose int8/int8 .../decoder/layers.4/attentions.0/attn/Transpose_5 BPU id(15) Transpose int8/int8 ...mer/decoder/layers.4/attentions.0/attn/MatMul_3 BPU id(15) HzSQuantizedMatmul int8/int8 ...former/decoder/layers.4/attentions.0/attn/Mul_6 BPU id(15) HzSElementwiseMul int8/int16 ...rmer/decoder/layers.4/attentions.0/attn/Softmax CPU -- Softmax float/float ...mer/decoder/layers.4/attentions.0/attn/MatMul_4 BPU id(16) HzSQuantizedMatmul int8/int8 .../decoder/layers.4/attentions.0/attn/Transpose_6 BPU id(16) Transpose int8/int8 ...er/decoder/layers.4/attentions.0/attn/Reshape_3 BPU id(16) Reshape int8/int8 ...sformer/decoder/layers.4/attentions.0/attn/Gemm BPU id(16) HzSQuantizedConv int8/int8 ...transformer/decoder/layers.4/attentions.0/Add_2 BPU id(16) HzSElementwiseAdd int8/int8 ...box_head/transformer/decoder/layers.4/Reshape_4 BPU id(16) Reshape int8/int8 ...transformer/decoder/layers.4/norms.0/ReduceMean BPU id(16) HzSQuantizedReduceMean int8/int8 ...x_head/transformer/decoder/layers.4/norms.0/Sub BPU id(16) HzSElementwiseSub int8/int8 ...x_head/transformer/decoder/layers.4/norms.0/Pow BPU id(16) HzLut int8/int8 ...ansformer/decoder/layers.4/norms.0/ReduceMean_1 BPU id(16) HzSQuantizedReduceMean int8/int8 ...sformer/decoder/layers.4/norms.0/Div_reciprocal BPU id(16) HzLut int8/int8 ...ad/transformer/decoder/layers.4/norms.0/Div_mul BPU id(16) HzSElementwiseMul int8/int8 ...x_head/transformer/decoder/layers.4/norms.0/Mul BPU id(16) HzSElementwiseMul int8/int8 ...head/transformer/decoder/layers.4/norms.0/Add_1 BPU id(16) HzSElementwiseAdd int8/int8 ...box_head/transformer/decoder/layers.4/Transpose BPU id(16) Transpose int8/int8 ...box_head/transformer/decoder/layers.4/Reshape_9 BPU id(16) Reshape int8/int8 ...d/transformer/decoder/layers.4/attentions.1/Add BPU id(16) HzSElementwiseAdd int8/int8 ...layers.4/attentions.1/attn/MatMul_reshape_input BPU id(16) Reshape int8/int8 ...ormer/decoder/layers.4/attentions.1/attn/MatMul BPU id(16) HzSQuantizedConv int8/int8 ...ayers.4/attentions.1/attn/MatMul_reshape_output BPU id(16) Reshape int8/int8 ...mer/decoder/layers.4/attentions.1/attn/MatMul_1 BPU id(16) HzSQuantizedConv int8/int8 ...ers.4/attentions.1/attn/MatMul_1_reshape_output BPU id(16) Reshape int8/int8 ...yers.4/attentions.1/attn/MatMul_2_reshape_input BPU id(16) Reshape int8/int8 ...mer/decoder/layers.4/attentions.1/attn/MatMul_2 BPU id(16) HzSQuantizedConv int8/int8 ...ers.4/attentions.1/attn/MatMul_2_reshape_output BPU id(16) Reshape int8/int8 .../decoder/layers.4/attentions.1/attn/Transpose_4 BPU id(16) Transpose int8/int8 ...er/decoder/layers.4/attentions.1/attn/Div_2_mul BPU id(16) HzSElementwiseMul int8/int8 ...rs.4/attentions.1/attn/Div_2_output_0_Transpose BPU id(16) Transpose int8/int8 .../decoder/layers.4/attentions.1/attn/Transpose_5 BPU id(16) Transpose int8/int8 ...mer/decoder/layers.4/attentions.1/attn/MatMul_3 BPU id(16) HzSQuantizedMatmul int8/int32 ...rmer/decoder/layers.4/attentions.1/attn/Softmax CPU -- Softmax float/float ...mer/decoder/layers.4/attentions.1/attn/MatMul_4 BPU id(17) HzSQuantizedMatmul int8/int8 .../decoder/layers.4/attentions.1/attn/Transpose_6 BPU id(17) Transpose int8/int8 ...er/decoder/layers.4/attentions.1/attn/Reshape_3 BPU id(17) Reshape int8/int8 ...sformer/decoder/layers.4/attentions.1/attn/Gemm BPU id(17) HzSQuantizedConv int8/int8 ...transformer/decoder/layers.4/attentions.1/Add_2 BPU id(17) HzSElementwiseAdd int8/int8 ...ox_head/transformer/decoder/layers.4/Reshape_12 BPU id(17) Reshape int8/int8 ...transformer/decoder/layers.4/norms.1/ReduceMean BPU id(17) HzSQuantizedReduceMean int8/int8 ...x_head/transformer/decoder/layers.4/norms.1/Sub BPU id(17) HzSElementwiseSub int8/int8 ...x_head/transformer/decoder/layers.4/norms.1/Pow BPU id(17) HzLut int8/int8 ...ansformer/decoder/layers.4/norms.1/ReduceMean_1 BPU id(17) HzSQuantizedReduceMean int8/int8 ...sformer/decoder/layers.4/norms.1/Div_reciprocal BPU id(17) HzLut int8/int8 ...ad/transformer/decoder/layers.4/norms.1/Div_mul BPU id(17) HzSElementwiseMul int8/int8 ...x_head/transformer/decoder/layers.4/norms.1/Mul BPU id(17) HzSElementwiseMul int8/int8 ...head/transformer/decoder/layers.4/norms.1/Add_1 BPU id(17) HzSElementwiseAdd int8/int8 ...yers.4/norms.1/Add_1_output_0_reshape_Transpose BPU id(17) Transpose int8/int8 ...coder/layers.4/norms.1/Add_1_output_0_Reshape_0 BPU id(17) Reshape int8/int8 ...d/transformer/decoder/layers.4/attentions.2/Add BPU id(17) HzSQuantizedConv int8/int8 ...er/layers.4/attentions.2/Add_output_0_Reshape_0 BPU id(17) Reshape int8/int8 ...decoder/layers.4/attentions.2/value_proj/MatMul BPU id(3) HzSQuantizedConv int8/int8 ...tentions.2/value_proj/MatMul_output_0_Reshape_0 BPU id(3) Reshape int8/int8 ...r/layers.4/attentions.2/sampling_offsets/MatMul BPU id(17) HzSQuantizedConv int8/int8 ...ntions.2/sampling_offsets/MatMul_reshape_output BPU id(17) Reshape int8/int8 .../layers.4/attentions.2/attention_weights/MatMul BPU id(17) HzSQuantizedConv int8/int32 ...tions.2/attention_weights/MatMul_reshape_output CPU -- Reshape float/float ...ansformer/decoder/layers.4/attentions.2/Softmax CPU -- Softmax float/float ...sformer/decoder/layers.4/attentions.2/Reshape_3 BPU id(19) Reshape int8/int8 ...ansformer/decoder/layers.4/attentions.2/Div_mul BPU id(17) HzSElementwiseMul int8/int16 ...transformer/decoder/layers.4/attentions.2/Add_1 CPU -- Add float/float ...d/transformer/decoder/layers.4/attentions.2/Mul BPU id(18) HzSElementwiseMul int8/int8 ...d/transformer/decoder/layers.4/attentions.2/Sub BPU id(18) HzSElementwiseSub int8/int8 ...er/decoder/layers.4/attentions.2/Gather_reshape BPU id(18) Reshape int8/int8 ...ormer/decoder/layers.4/attentions.2/Transpose_3 BPU id(18) Transpose int8/int8 ...sformer/decoder/layers.4/attentions.2/Reshape_6 BPU id(18) Reshape int8/int8 ...nsformer/decoder/layers.4/attentions.2/Gather_1 BPU id(18) Gather int8/int8 ...nsformer/decoder/layers.4/attentions.2/Gather_2 BPU id(18) Gather int8/int8 ...transformer/decoder/layers.4/attentions.2/Add_2 BPU id(18) HzSElementwiseAdd int8/int8 ...transformer/decoder/layers.4/attentions.2/Mul_1 BPU id(18) HzSElementwiseMul int8/int8 ...transformer/decoder/layers.4/attentions.2/Sub_1 BPU id(18) HzSElementwiseSub int8/int8 ...sformer/decoder/layers.4/attentions.2/Div_1_mul BPU id(18) HzSElementwiseMul int8/int8 .../layers.4/attentions.2/Div_1_output_0_Reshape_0 BPU id(18) Reshape int8/int8 ...transformer/decoder/layers.4/attentions.2/Add_3 BPU id(18) HzSElementwiseAdd int8/int8 ...transformer/decoder/layers.4/attentions.2/Mul_2 BPU id(18) HzSElementwiseMul int8/int8 ...transformer/decoder/layers.4/attentions.2/Sub_2 BPU id(18) HzSElementwiseSub int8/int8 ...sformer/decoder/layers.4/attentions.2/Div_2_mul BPU id(18) HzSElementwiseMul int8/int8 .../layers.4/attentions.2/Div_2_output_0_Reshape_0 BPU id(18) Reshape int8/int8 ...transformer/decoder/layers.4/attentions.2/Floor BPU id(18) HzLut int8/int8 .../transformer/decoder/layers.4/attentions.2/Cast CPU -- Cast float/int8 ...ansformer/decoder/layers.4/attentions.2/Floor_1 BPU id(18) HzLut int8/int8 ...ransformer/decoder/layers.4/attentions.2/Cast_1 CPU -- Cast float/int8 ...transformer/decoder/layers.4/attentions.2/Add_4 CPU -- Add int8/int8 ...transformer/decoder/layers.4/attentions.2/Add_5 CPU -- Add int8/int8 ...ransformer/decoder/layers.4/attentions.2/Cast_2 CPU -- Cast int8/float ...transformer/decoder/layers.4/attentions.2/Sub_3 BPU id(19) HzSElementwiseSub int8/int8 ...ransformer/decoder/layers.4/attentions.2/Cast_3 CPU -- Cast int8/float ...transformer/decoder/layers.4/attentions.2/Sub_4 BPU id(19) HzSElementwiseSub int8/int8 ...transformer/decoder/layers.4/attentions.2/Mul_3 BPU id(19) HzSElementwiseMul int8/int8 ...rmer/decoder/layers.4/attentions.2/Unsqueeze_25 BPU id(19) Reshape int8/int8 ...ransformer/decoder/layers.4/attentions.2/Cast_4 CPU -- Cast int8/float ...transformer/decoder/layers.4/attentions.2/Sub_5 BPU id(19) HzSElementwiseSub int8/int8 ...transformer/decoder/layers.4/attentions.2/Mul_4 BPU id(19) HzSElementwiseMul int8/int8 ...rmer/decoder/layers.4/attentions.2/Unsqueeze_26 BPU id(19) Reshape int8/int8 ...ransformer/decoder/layers.4/attentions.2/Cast_5 CPU -- Cast int8/float ...transformer/decoder/layers.4/attentions.2/Sub_6 BPU id(19) HzSElementwiseSub int8/int8 ...transformer/decoder/layers.4/attentions.2/Mul_5 BPU id(19) HzSElementwiseMul int8/int8 ...rmer/decoder/layers.4/attentions.2/Unsqueeze_27 BPU id(19) Reshape int8/int8 ...transformer/decoder/layers.4/attentions.2/Mul_6 BPU id(19) HzSElementwiseMul int8/int8 ...rmer/decoder/layers.4/attentions.2/Unsqueeze_28 BPU id(19) Reshape int8/int8 ...d/transformer/decoder/layers.4/attentions.2/Pad BPU id(3) HzPad int8/int8 ...transformer/decoder/layers.4/attentions.2/Add_6 CPU -- Add int8/int8 ...transformer/decoder/layers.4/attentions.2/Add_7 CPU -- Add int8/int8 .../transformer/decoder/layers.4/attentions.2/Less CPU -- Less int8/int8 ...former/decoder/layers.4/attentions.2/Where_cast CPU -- Cast int8/int8 ...ormer/decoder/layers.4/attentions.2/Where_mul_x CPU -- Mul int8/int8 ...ormer/decoder/layers.4/attentions.2/Where_equal CPU -- Equal int8/int8 ...r/layers.4/attentions.2/Where_equal_output_cast CPU -- Cast int8/int8 ...ormer/decoder/layers.4/attentions.2/Where_mul_y CPU -- Mul int8/int8 ...sformer/decoder/layers.4/attentions.2/Where_add CPU -- Add int8/int8 ...ansformer/decoder/layers.4/attentions.2/Greater CPU -- Greater int8/int8 ...rmer/decoder/layers.4/attentions.2/Where_1_cast CPU -- Cast int8/int8 ...mer/decoder/layers.4/attentions.2/Where_1_mul_x CPU -- Mul int8/int8 ...mer/decoder/layers.4/attentions.2/Where_1_equal CPU -- Equal int8/int8 ...layers.4/attentions.2/Where_1_equal_output_cast CPU -- Cast int8/int8 ...mer/decoder/layers.4/attentions.2/Where_1_mul_y CPU -- Mul int8/int8 ...ormer/decoder/layers.4/attentions.2/Where_1_add CPU -- Add int8/int8 ...ransformer/decoder/layers.4/attentions.2/Less_1 CPU -- Less int8/int8 ...rmer/decoder/layers.4/attentions.2/Where_2_cast CPU -- Cast int8/int8 ...mer/decoder/layers.4/attentions.2/Where_2_mul_x CPU -- Mul int8/int8 ...mer/decoder/layers.4/attentions.2/Where_2_equal CPU -- Equal int8/int8 ...layers.4/attentions.2/Where_2_equal_output_cast CPU -- Cast int8/int8 ...mer/decoder/layers.4/attentions.2/Where_2_mul_y CPU -- Mul int8/int8 ...ormer/decoder/layers.4/attentions.2/Where_2_add CPU -- Add int8/int8 ...sformer/decoder/layers.4/attentions.2/Greater_1 CPU -- Greater int8/int8 ...rmer/decoder/layers.4/attentions.2/Where_3_cast CPU -- Cast int8/int8 ...mer/decoder/layers.4/attentions.2/Where_3_mul_x CPU -- Mul int8/int8 ...mer/decoder/layers.4/attentions.2/Where_3_equal CPU -- Equal int8/int8 ...layers.4/attentions.2/Where_3_equal_output_cast CPU -- Cast int8/int8 ...mer/decoder/layers.4/attentions.2/Where_3_mul_y CPU -- Mul int8/int8 ...ormer/decoder/layers.4/attentions.2/Where_3_add CPU -- Add int8/int8 ...ransformer/decoder/layers.4/attentions.2/Less_2 CPU -- Less int8/int8 ...rmer/decoder/layers.4/attentions.2/Where_4_cast CPU -- Cast int8/int8 ...mer/decoder/layers.4/attentions.2/Where_4_mul_x CPU -- Mul int8/int8 ...mer/decoder/layers.4/attentions.2/Where_4_equal CPU -- Equal int8/int8 ...layers.4/attentions.2/Where_4_equal_output_cast CPU -- Cast int8/int8 ...mer/decoder/layers.4/attentions.2/Where_4_mul_y CPU -- Mul int8/int8 ...ormer/decoder/layers.4/attentions.2/Where_4_add CPU -- Add int8/int8 ...sformer/decoder/layers.4/attentions.2/Greater_2 CPU -- Greater int8/int8 ...rmer/decoder/layers.4/attentions.2/Where_5_cast CPU -- Cast int8/int8 ...mer/decoder/layers.4/attentions.2/Where_5_mul_x CPU -- Mul int8/int8 ...mer/decoder/layers.4/attentions.2/Where_5_equal CPU -- Equal int8/int8 ...layers.4/attentions.2/Where_5_equal_output_cast CPU -- Cast int8/int8 ...mer/decoder/layers.4/attentions.2/Where_5_mul_y CPU -- Mul int8/int8 ...ormer/decoder/layers.4/attentions.2/Where_5_add CPU -- Add int8/int8 ...ransformer/decoder/layers.4/attentions.2/Less_3 CPU -- Less int8/int8 ...rmer/decoder/layers.4/attentions.2/Where_6_cast CPU -- Cast int8/int8 ...mer/decoder/layers.4/attentions.2/Where_6_mul_x CPU -- Mul int8/int8 ...mer/decoder/layers.4/attentions.2/Where_6_equal CPU -- Equal int8/int8 ...layers.4/attentions.2/Where_6_equal_output_cast CPU -- Cast int8/int8 ...mer/decoder/layers.4/attentions.2/Where_6_mul_y CPU -- Mul int8/int8 ...ormer/decoder/layers.4/attentions.2/Where_6_add CPU -- Add int8/int8 ...sformer/decoder/layers.4/attentions.2/Greater_3 CPU -- Greater int8/int8 ...rmer/decoder/layers.4/attentions.2/Where_7_cast CPU -- Cast int8/int8 ...mer/decoder/layers.4/attentions.2/Where_7_mul_x CPU -- Mul int8/int8 ...mer/decoder/layers.4/attentions.2/Where_7_equal CPU -- Equal int8/int8 ...layers.4/attentions.2/Where_7_equal_output_cast CPU -- Cast int8/int8 ...mer/decoder/layers.4/attentions.2/Where_7_mul_y CPU -- Mul int8/int8 ...ormer/decoder/layers.4/attentions.2/Where_7_add CPU -- Add int8/int8 ...former/decoder/layers.4/attentions.2/Reshape_11 BPU id(3) Reshape int8/int8 ...transformer/decoder/layers.4/attentions.2/Mul_8 CPU -- Mul int8/int8 ...transformer/decoder/layers.4/attentions.2/Add_8 CPU -- Add int8/int8 ...rmer/decoder/layers.4/attentions.2/Unsqueeze_32 CPU -- Reshape int8/int8 ...ransformer/decoder/layers.4/attentions.2/Expand CPU -- Expand int8/int8 ...ransformer/decoder/layers.4/attentions.2/Mul_10 CPU -- Mul int8/int8 ...transformer/decoder/layers.4/attentions.2/Add_9 CPU -- Add int8/int8 ...rmer/decoder/layers.4/attentions.2/Unsqueeze_36 CPU -- Reshape int8/int8 ...nsformer/decoder/layers.4/attentions.2/Expand_1 CPU -- Expand int8/int8 ...ransformer/decoder/layers.4/attentions.2/Add_10 CPU -- Add int8/int8 ...rmer/decoder/layers.4/attentions.2/Unsqueeze_40 CPU -- Reshape int8/int8 ...nsformer/decoder/layers.4/attentions.2/Expand_2 CPU -- Expand int8/int8 ...ransformer/decoder/layers.4/attentions.2/Add_11 CPU -- Add int8/int8 ...rmer/decoder/layers.4/attentions.2/Unsqueeze_44 CPU -- Reshape int8/int8 ...nsformer/decoder/layers.4/attentions.2/Expand_3 CPU -- Expand int8/int8 ...coder/layers.4/attentions.2/GatherElements_Cast CPU -- Cast int8/int32 ...er/decoder/layers.4/attentions.2/GatherElements BPU id(19) GatherElements int8/int8 ...der/layers.4/attentions.2/GatherElements_1_Cast CPU -- Cast int8/int32 .../decoder/layers.4/attentions.2/GatherElements_1 BPU id(19) GatherElements int8/int8 ...der/layers.4/attentions.2/GatherElements_2_Cast CPU -- Cast int8/int32 .../decoder/layers.4/attentions.2/GatherElements_2 BPU id(19) GatherElements int8/int8 ...der/layers.4/attentions.2/GatherElements_3_Cast CPU -- Cast int8/int32 .../decoder/layers.4/attentions.2/GatherElements_3 BPU id(19) GatherElements int8/int8 ...ransformer/decoder/layers.4/attentions.2/Mul_14 BPU id(19) HzSElementwiseMul int8/int8 ...ransformer/decoder/layers.4/attentions.2/Mul_15 BPU id(19) HzSElementwiseMul int8/int8 ...ransformer/decoder/layers.4/attentions.2/Add_12 BPU id(19) HzSElementwiseAdd int8/int8 ...ransformer/decoder/layers.4/attentions.2/Mul_16 BPU id(19) HzSElementwiseMul int8/int8 ...ransformer/decoder/layers.4/attentions.2/Add_13 BPU id(19) HzSElementwiseAdd int8/int8 ...ransformer/decoder/layers.4/attentions.2/Mul_17 BPU id(19) HzSElementwiseMul int8/int8 ...ransformer/decoder/layers.4/attentions.2/Add_14 BPU id(19) HzSElementwiseAdd int8/int8 ...former/decoder/layers.4/attentions.2/Reshape_16 BPU id(19) Reshape int8/int8 ...ormer/decoder/layers.4/attentions.2/Transpose_5 BPU id(19) Transpose int8/int8 ...former/decoder/layers.4/attentions.2/Reshape_17 BPU id(19) Reshape int8/int8 ...ransformer/decoder/layers.4/attentions.2/Mul_18 BPU id(19) HzSElementwiseMul int8/int8 ...sformer/decoder/layers.4/attentions.2/ReduceSum BPU id(19) HzSQuantizedReduceSum int8/int8 ...decoder/layers.4/attentions.2/ReduceSum_reshape BPU id(19) Reshape int8/int8 ...ecoder/layers.4/attentions.2/output_proj/MatMul BPU id(19) HzSQuantizedConv int8/int8 .../attentions.2/output_proj/MatMul_reshape_output BPU id(19) Reshape int8/int8 ...ransformer/decoder/layers.4/attentions.2/Add_15 BPU id(19) HzSElementwiseAdd int8/int8 ...transformer/decoder/layers.4/norms.2/ReduceMean BPU id(19) HzSQuantizedReduceMean int8/int8 ...x_head/transformer/decoder/layers.4/norms.2/Sub BPU id(19) HzSElementwiseSub int8/int8 ...x_head/transformer/decoder/layers.4/norms.2/Pow BPU id(19) HzLut int8/int8 ...ansformer/decoder/layers.4/norms.2/ReduceMean_1 BPU id(19) HzSQuantizedReduceMean int8/int8 ...sformer/decoder/layers.4/norms.2/Div_reciprocal BPU id(19) HzLut int8/int8 ...ad/transformer/decoder/layers.4/norms.2/Div_mul BPU id(19) HzSElementwiseMul int8/int8 ...x_head/transformer/decoder/layers.4/norms.2/Mul BPU id(19) HzSElementwiseMul int8/int8 ...head/transformer/decoder/layers.4/norms.2/Add_1 BPU id(19) HzSElementwiseAdd int8/int8 ...layers/layers.0/layers.0.0/MatMul_reshape_input BPU id(19) Reshape int8/int8 ...yers.4/ffns.0/layers/layers.0/layers.0.0/MatMul BPU id(19) HzSQuantizedConv int8/int8 .../decoder/layers.4/ffns.0/layers/layers.1/MatMul BPU id(19) HzSQuantizedConv int8/int8 ....4/ffns.0/layers/layers.1/MatMul_reshape_output BPU id(19) Reshape int8/int8 ...ox_head/transformer/decoder/layers.4/ffns.0/Add BPU id(19) HzSElementwiseAdd int8/int8 ...transformer/decoder/layers.4/norms.3/ReduceMean BPU id(19) HzSQuantizedReduceMean int8/int8 ...x_head/transformer/decoder/layers.4/norms.3/Sub BPU id(19) HzSElementwiseSub int8/int8 ...x_head/transformer/decoder/layers.4/norms.3/Pow BPU id(19) HzLut int8/int8 ...ansformer/decoder/layers.4/norms.3/ReduceMean_1 BPU id(19) HzSQuantizedReduceMean int8/int8 ...sformer/decoder/layers.4/norms.3/Div_reciprocal BPU id(19) HzLut int8/int8 ...ad/transformer/decoder/layers.4/norms.3/Div_mul BPU id(19) HzSElementwiseMul int8/int8 ...x_head/transformer/decoder/layers.4/norms.3/Mul BPU id(19) HzSElementwiseMul int8/int8 ...head/transformer/decoder/layers.4/norms.3/Add_1 BPU id(19) HzSElementwiseAdd int8/int8 ...ox_head/transformer/decoder/Transpose_4_reshape BPU id(19) Reshape int8/int8 .../decoder/reg_branches.4/reg_branches.4.0/MatMul BPU id(19) HzSQuantizedConv int8/int8 .../decoder/reg_branches.4/reg_branches.4.2/MatMul BPU id(19) HzSQuantizedConv int8/int8 ...put_0_reshape_transpose_calibrated_TO_FUSE_RELU BPU id(15) HzSQuantizedConv int8/int8 /pts_bbox_head/transformer/decoder/Clip_13 BPU id(15) HzLut int8/int8 /pts_bbox_head/transformer/decoder/Sub_4 BPU id(15) HzSElementwiseSub int8/int8 /pts_bbox_head/transformer/decoder/Clip_14 BPU id(15) HzLut int8/int8 ..._bbox_head/transformer/decoder/Div_4_reciprocal BPU id(15) HzLut int8/int8 /pts_bbox_head/transformer/decoder/Div_4_mul BPU id(15) HzSElementwiseMul int8/int8 /pts_bbox_head/transformer/decoder/Log_4 BPU id(15) HzLut int8/int8 .../decoder/reg_branches.4/reg_branches.4.4/MatMul BPU id(19) HzSQuantizedConv int8/int8 /pts_bbox_head/transformer/decoder/Sigmoid_4 BPU id(19) HzLut int8/int8 ...ransformer/decoder/Sigmoid_4_output_0_Reshape_0 BPU id(19) Reshape int8/int8 /pts_bbox_head/transformer/decoder/Unsqueeze_5 BPU id(19) Reshape int8/int8 ..._bbox_head/transformer/decoder/layers.5/Reshape BPU id(19) Reshape int8/int8 ...d/transformer/decoder/layers.5/attentions.0/Add BPU id(19) HzSElementwiseAdd int8/int8 ...er/layers.5/attentions.0/Add_output_0_Reshape_0 BPU id(19) Reshape int8/int8 ...ormer/decoder/layers.5/attentions.0/attn/MatMul BPU id(19) HzSQuantizedConv int8/int8 ...ayers.5/attentions.0/attn/MatMul_reshape_output BPU id(19) Reshape int8/int8 ...mer/decoder/layers.5/attentions.0/attn/MatMul_1 BPU id(19) HzSQuantizedConv int8/int8 ...ers.5/attentions.0/attn/MatMul_1_reshape_output BPU id(19) Reshape int8/int8 ...yers.5/attentions.0/attn/MatMul_2_reshape_input BPU id(19) Reshape int8/int8 ...mer/decoder/layers.5/attentions.0/attn/MatMul_2 BPU id(19) HzSQuantizedConv int8/int8 ...ers.5/attentions.0/attn/MatMul_2_reshape_output BPU id(19) Reshape int8/int8 .../decoder/layers.5/attentions.0/attn/Transpose_4 BPU id(19) Transpose int8/int8 ...er/decoder/layers.5/attentions.0/attn/Div_2_mul BPU id(19) HzSElementwiseMul int8/int8 ...rs.5/attentions.0/attn/Div_2_output_0_Transpose BPU id(19) Transpose int8/int8 .../decoder/layers.5/attentions.0/attn/Transpose_5 BPU id(19) Transpose int8/int8 ...mer/decoder/layers.5/attentions.0/attn/MatMul_3 BPU id(19) HzSQuantizedMatmul int8/int8 ...former/decoder/layers.5/attentions.0/attn/Mul_6 BPU id(19) HzSElementwiseMul int8/int16 ...rmer/decoder/layers.5/attentions.0/attn/Softmax CPU -- Softmax float/float ...mer/decoder/layers.5/attentions.0/attn/MatMul_4 BPU id(20) HzSQuantizedMatmul int8/int8 .../decoder/layers.5/attentions.0/attn/Transpose_6 BPU id(20) Transpose int8/int8 ...er/decoder/layers.5/attentions.0/attn/Reshape_3 BPU id(20) Reshape int8/int8 ...sformer/decoder/layers.5/attentions.0/attn/Gemm BPU id(20) HzSQuantizedConv int8/int8 ...transformer/decoder/layers.5/attentions.0/Add_2 BPU id(20) HzSElementwiseAdd int8/int8 ...box_head/transformer/decoder/layers.5/Reshape_4 BPU id(20) Reshape int8/int8 ...transformer/decoder/layers.5/norms.0/ReduceMean BPU id(20) HzSQuantizedReduceMean int8/int8 ...x_head/transformer/decoder/layers.5/norms.0/Sub BPU id(20) HzSElementwiseSub int8/int8 ...x_head/transformer/decoder/layers.5/norms.0/Pow BPU id(20) HzLut int8/int8 ...ansformer/decoder/layers.5/norms.0/ReduceMean_1 BPU id(20) HzSQuantizedReduceMean int8/int8 ...sformer/decoder/layers.5/norms.0/Div_reciprocal BPU id(20) HzLut int8/int8 ...ad/transformer/decoder/layers.5/norms.0/Div_mul BPU id(20) HzSElementwiseMul int8/int8 ...x_head/transformer/decoder/layers.5/norms.0/Mul BPU id(20) HzSElementwiseMul int8/int8 ...head/transformer/decoder/layers.5/norms.0/Add_1 BPU id(20) HzSElementwiseAdd int8/int8 ...box_head/transformer/decoder/layers.5/Transpose BPU id(20) Transpose int8/int8 ...box_head/transformer/decoder/layers.5/Reshape_9 BPU id(20) Reshape int8/int8 ...d/transformer/decoder/layers.5/attentions.1/Add BPU id(20) HzSElementwiseAdd int8/int8 ...layers.5/attentions.1/attn/MatMul_reshape_input BPU id(20) Reshape int8/int8 ...ormer/decoder/layers.5/attentions.1/attn/MatMul BPU id(20) HzSQuantizedConv int8/int8 ...ayers.5/attentions.1/attn/MatMul_reshape_output BPU id(20) Reshape int8/int8 ...mer/decoder/layers.5/attentions.1/attn/MatMul_1 BPU id(20) HzSQuantizedConv int8/int8 ...ers.5/attentions.1/attn/MatMul_1_reshape_output BPU id(20) Reshape int8/int8 ...yers.5/attentions.1/attn/MatMul_2_reshape_input BPU id(20) Reshape int8/int8 ...mer/decoder/layers.5/attentions.1/attn/MatMul_2 BPU id(20) HzSQuantizedConv int8/int8 ...ers.5/attentions.1/attn/MatMul_2_reshape_output BPU id(20) Reshape int8/int8 .../decoder/layers.5/attentions.1/attn/Transpose_4 BPU id(20) Transpose int8/int8 ...er/decoder/layers.5/attentions.1/attn/Div_2_mul BPU id(20) HzSElementwiseMul int8/int8 ...rs.5/attentions.1/attn/Div_2_output_0_Transpose BPU id(20) Transpose int8/int8 .../decoder/layers.5/attentions.1/attn/Transpose_5 BPU id(20) Transpose int8/int8 ...mer/decoder/layers.5/attentions.1/attn/MatMul_3 BPU id(20) HzSQuantizedMatmul int8/int32 ...rmer/decoder/layers.5/attentions.1/attn/Softmax CPU -- Softmax float/float ...mer/decoder/layers.5/attentions.1/attn/MatMul_4 BPU id(21) HzSQuantizedMatmul int8/int8 .../decoder/layers.5/attentions.1/attn/Transpose_6 BPU id(21) Transpose int8/int8 ...er/decoder/layers.5/attentions.1/attn/Reshape_3 BPU id(21) Reshape int8/int8 ...sformer/decoder/layers.5/attentions.1/attn/Gemm BPU id(21) HzSQuantizedConv int8/int8 ...transformer/decoder/layers.5/attentions.1/Add_2 BPU id(21) HzSElementwiseAdd int8/int8 ...ox_head/transformer/decoder/layers.5/Reshape_12 BPU id(21) Reshape int8/int8 ...transformer/decoder/layers.5/norms.1/ReduceMean BPU id(21) HzSQuantizedReduceMean int8/int8 ...x_head/transformer/decoder/layers.5/norms.1/Sub BPU id(21) HzSElementwiseSub int8/int8 ...x_head/transformer/decoder/layers.5/norms.1/Pow BPU id(21) HzLut int8/int8 ...ansformer/decoder/layers.5/norms.1/ReduceMean_1 BPU id(21) HzSQuantizedReduceMean int8/int8 ...sformer/decoder/layers.5/norms.1/Div_reciprocal BPU id(21) HzLut int8/int8 ...ad/transformer/decoder/layers.5/norms.1/Div_mul BPU id(21) HzSElementwiseMul int8/int8 ...x_head/transformer/decoder/layers.5/norms.1/Mul BPU id(21) HzSElementwiseMul int8/int8 ...head/transformer/decoder/layers.5/norms.1/Add_1 BPU id(21) HzSElementwiseAdd int8/int8 ...yers.5/norms.1/Add_1_output_0_reshape_Transpose BPU id(21) Transpose int8/int8 ...coder/layers.5/norms.1/Add_1_output_0_Reshape_0 BPU id(21) Reshape int8/int8 ...d/transformer/decoder/layers.5/attentions.2/Add BPU id(21) HzSQuantizedConv int8/int8 ...er/layers.5/attentions.2/Add_output_0_Reshape_0 BPU id(21) Reshape int8/int8 ...decoder/layers.5/attentions.2/value_proj/MatMul BPU id(3) HzSQuantizedConv int8/int8 ...tentions.2/value_proj/MatMul_output_0_Reshape_0 BPU id(3) Reshape int8/int8 ...r/layers.5/attentions.2/sampling_offsets/MatMul BPU id(21) HzSQuantizedConv int8/int8 ...ntions.2/sampling_offsets/MatMul_reshape_output BPU id(21) Reshape int8/int8 .../layers.5/attentions.2/attention_weights/MatMul BPU id(21) HzSQuantizedConv int8/int32 ...tions.2/attention_weights/MatMul_reshape_output CPU -- Reshape float/float ...ansformer/decoder/layers.5/attentions.2/Softmax CPU -- Softmax float/float ...sformer/decoder/layers.5/attentions.2/Reshape_3 BPU id(23) Reshape int8/int8 ...ansformer/decoder/layers.5/attentions.2/Div_mul BPU id(21) HzSElementwiseMul int8/int16 ...transformer/decoder/layers.5/attentions.2/Add_1 CPU -- Add float/float ...d/transformer/decoder/layers.5/attentions.2/Mul BPU id(22) HzSElementwiseMul int8/int8 ...d/transformer/decoder/layers.5/attentions.2/Sub BPU id(22) HzSElementwiseSub int8/int8 ...er/decoder/layers.5/attentions.2/Gather_reshape BPU id(22) Reshape int8/int8 ...ormer/decoder/layers.5/attentions.2/Transpose_3 BPU id(22) Transpose int8/int8 ...sformer/decoder/layers.5/attentions.2/Reshape_6 BPU id(22) Reshape int8/int8 ...nsformer/decoder/layers.5/attentions.2/Gather_1 BPU id(22) Gather int8/int8 ...nsformer/decoder/layers.5/attentions.2/Gather_2 BPU id(22) Gather int8/int8 ...transformer/decoder/layers.5/attentions.2/Add_2 BPU id(22) HzSElementwiseAdd int8/int8 ...transformer/decoder/layers.5/attentions.2/Mul_1 BPU id(22) HzSElementwiseMul int8/int8 ...transformer/decoder/layers.5/attentions.2/Sub_1 BPU id(22) HzSElementwiseSub int8/int8 ...sformer/decoder/layers.5/attentions.2/Div_1_mul BPU id(22) HzSElementwiseMul int8/int8 .../layers.5/attentions.2/Div_1_output_0_Reshape_0 BPU id(22) Reshape int8/int8 ...transformer/decoder/layers.5/attentions.2/Add_3 BPU id(22) HzSElementwiseAdd int8/int8 ...transformer/decoder/layers.5/attentions.2/Mul_2 BPU id(22) HzSElementwiseMul int8/int8 ...transformer/decoder/layers.5/attentions.2/Sub_2 BPU id(22) HzSElementwiseSub int8/int8 ...sformer/decoder/layers.5/attentions.2/Div_2_mul BPU id(22) HzSElementwiseMul int8/int8 .../layers.5/attentions.2/Div_2_output_0_Reshape_0 BPU id(22) Reshape int8/int8 ...transformer/decoder/layers.5/attentions.2/Floor BPU id(22) HzLut int8/int8 .../transformer/decoder/layers.5/attentions.2/Cast CPU -- Cast float/int8 ...ansformer/decoder/layers.5/attentions.2/Floor_1 BPU id(22) HzLut int8/int8 ...ransformer/decoder/layers.5/attentions.2/Cast_1 CPU -- Cast float/int8 ...transformer/decoder/layers.5/attentions.2/Add_4 CPU -- Add int8/int8 ...transformer/decoder/layers.5/attentions.2/Add_5 CPU -- Add int8/int8 ...ransformer/decoder/layers.5/attentions.2/Cast_2 CPU -- Cast int8/float ...transformer/decoder/layers.5/attentions.2/Sub_3 BPU id(23) HzSElementwiseSub int8/int8 ...ransformer/decoder/layers.5/attentions.2/Cast_3 CPU -- Cast int8/float ...transformer/decoder/layers.5/attentions.2/Sub_4 BPU id(23) HzSElementwiseSub int8/int8 ...transformer/decoder/layers.5/attentions.2/Mul_3 BPU id(23) HzSElementwiseMul int8/int8 ...rmer/decoder/layers.5/attentions.2/Unsqueeze_25 BPU id(23) Reshape int8/int8 ...ransformer/decoder/layers.5/attentions.2/Cast_4 CPU -- Cast int8/float ...transformer/decoder/layers.5/attentions.2/Sub_5 BPU id(23) HzSElementwiseSub int8/int8 ...transformer/decoder/layers.5/attentions.2/Mul_4 BPU id(23) HzSElementwiseMul int8/int8 ...rmer/decoder/layers.5/attentions.2/Unsqueeze_26 BPU id(23) Reshape int8/int8 ...ransformer/decoder/layers.5/attentions.2/Cast_5 CPU -- Cast int8/float ...transformer/decoder/layers.5/attentions.2/Sub_6 BPU id(23) HzSElementwiseSub int8/int8 ...transformer/decoder/layers.5/attentions.2/Mul_5 BPU id(23) HzSElementwiseMul int8/int8 ...rmer/decoder/layers.5/attentions.2/Unsqueeze_27 BPU id(23) Reshape int8/int8 ...transformer/decoder/layers.5/attentions.2/Mul_6 BPU id(23) HzSElementwiseMul int8/int8 ...rmer/decoder/layers.5/attentions.2/Unsqueeze_28 BPU id(23) Reshape int8/int8 ...d/transformer/decoder/layers.5/attentions.2/Pad BPU id(3) HzPad int8/int8 ...transformer/decoder/layers.5/attentions.2/Add_6 CPU -- Add int8/int8 ...transformer/decoder/layers.5/attentions.2/Add_7 CPU -- Add int8/int8 .../transformer/decoder/layers.5/attentions.2/Less CPU -- Less int8/int8 ...former/decoder/layers.5/attentions.2/Where_cast CPU -- Cast int8/int8 ...ormer/decoder/layers.5/attentions.2/Where_mul_x CPU -- Mul int8/int8 ...ormer/decoder/layers.5/attentions.2/Where_equal CPU -- Equal int8/int8 ...r/layers.5/attentions.2/Where_equal_output_cast CPU -- Cast int8/int8 ...ormer/decoder/layers.5/attentions.2/Where_mul_y CPU -- Mul int8/int8 ...sformer/decoder/layers.5/attentions.2/Where_add CPU -- Add int8/int8 ...ansformer/decoder/layers.5/attentions.2/Greater CPU -- Greater int8/int8 ...rmer/decoder/layers.5/attentions.2/Where_1_cast CPU -- Cast int8/int8 ...mer/decoder/layers.5/attentions.2/Where_1_mul_x CPU -- Mul int8/int8 ...mer/decoder/layers.5/attentions.2/Where_1_equal CPU -- Equal int8/int8 ...layers.5/attentions.2/Where_1_equal_output_cast CPU -- Cast int8/int8 ...mer/decoder/layers.5/attentions.2/Where_1_mul_y CPU -- Mul int8/int8 ...ormer/decoder/layers.5/attentions.2/Where_1_add CPU -- Add int8/int8 ...ransformer/decoder/layers.5/attentions.2/Less_1 CPU -- Less int8/int8 ...rmer/decoder/layers.5/attentions.2/Where_2_cast CPU -- Cast int8/int8 ...mer/decoder/layers.5/attentions.2/Where_2_mul_x CPU -- Mul int8/int8 ...mer/decoder/layers.5/attentions.2/Where_2_equal CPU -- Equal int8/int8 ...layers.5/attentions.2/Where_2_equal_output_cast CPU -- Cast int8/int8 ...mer/decoder/layers.5/attentions.2/Where_2_mul_y CPU -- Mul int8/int8 ...ormer/decoder/layers.5/attentions.2/Where_2_add CPU -- Add int8/int8 ...sformer/decoder/layers.5/attentions.2/Greater_1 CPU -- Greater int8/int8 ...rmer/decoder/layers.5/attentions.2/Where_3_cast CPU -- Cast int8/int8 ...mer/decoder/layers.5/attentions.2/Where_3_mul_x CPU -- Mul int8/int8 ...mer/decoder/layers.5/attentions.2/Where_3_equal CPU -- Equal int8/int8 ...layers.5/attentions.2/Where_3_equal_output_cast CPU -- Cast int8/int8 ...mer/decoder/layers.5/attentions.2/Where_3_mul_y CPU -- Mul int8/int8 ...ormer/decoder/layers.5/attentions.2/Where_3_add CPU -- Add int8/int8 ...ransformer/decoder/layers.5/attentions.2/Less_2 CPU -- Less int8/int8 ...rmer/decoder/layers.5/attentions.2/Where_4_cast CPU -- Cast int8/int8 ...mer/decoder/layers.5/attentions.2/Where_4_mul_x CPU -- Mul int8/int8 ...mer/decoder/layers.5/attentions.2/Where_4_equal CPU -- Equal int8/int8 ...layers.5/attentions.2/Where_4_equal_output_cast CPU -- Cast int8/int8 ...mer/decoder/layers.5/attentions.2/Where_4_mul_y CPU -- Mul int8/int8 ...ormer/decoder/layers.5/attentions.2/Where_4_add CPU -- Add int8/int8 ...sformer/decoder/layers.5/attentions.2/Greater_2 CPU -- Greater int8/int8 ...rmer/decoder/layers.5/attentions.2/Where_5_cast CPU -- Cast int8/int8 ...mer/decoder/layers.5/attentions.2/Where_5_mul_x CPU -- Mul int8/int8 ...mer/decoder/layers.5/attentions.2/Where_5_equal CPU -- Equal int8/int8 ...layers.5/attentions.2/Where_5_equal_output_cast CPU -- Cast int8/int8 ...mer/decoder/layers.5/attentions.2/Where_5_mul_y CPU -- Mul int8/int8 ...ormer/decoder/layers.5/attentions.2/Where_5_add CPU -- Add int8/int8 ...ransformer/decoder/layers.5/attentions.2/Less_3 CPU -- Less int8/int8 ...rmer/decoder/layers.5/attentions.2/Where_6_cast CPU -- Cast int8/int8 ...mer/decoder/layers.5/attentions.2/Where_6_mul_x CPU -- Mul int8/int8 ...mer/decoder/layers.5/attentions.2/Where_6_equal CPU -- Equal int8/int8 ...layers.5/attentions.2/Where_6_equal_output_cast CPU -- Cast int8/int8 ...mer/decoder/layers.5/attentions.2/Where_6_mul_y CPU -- Mul int8/int8 ...ormer/decoder/layers.5/attentions.2/Where_6_add CPU -- Add int8/int8 ...sformer/decoder/layers.5/attentions.2/Greater_3 CPU -- Greater int8/int8 ...rmer/decoder/layers.5/attentions.2/Where_7_cast CPU -- Cast int8/int8 ...mer/decoder/layers.5/attentions.2/Where_7_mul_x CPU -- Mul int8/int8 ...mer/decoder/layers.5/attentions.2/Where_7_equal CPU -- Equal int8/int8 ...layers.5/attentions.2/Where_7_equal_output_cast CPU -- Cast int8/int8 ...mer/decoder/layers.5/attentions.2/Where_7_mul_y CPU -- Mul int8/int8 ...ormer/decoder/layers.5/attentions.2/Where_7_add CPU -- Add int8/int8 ...former/decoder/layers.5/attentions.2/Reshape_11 BPU id(3) Reshape int8/int8 ...transformer/decoder/layers.5/attentions.2/Mul_8 CPU -- Mul int8/int8 ...transformer/decoder/layers.5/attentions.2/Add_8 CPU -- Add int8/int8 ...rmer/decoder/layers.5/attentions.2/Unsqueeze_32 CPU -- Reshape int8/int8 ...ransformer/decoder/layers.5/attentions.2/Expand CPU -- Expand int8/int8 ...ransformer/decoder/layers.5/attentions.2/Mul_10 CPU -- Mul int8/int8 ...transformer/decoder/layers.5/attentions.2/Add_9 CPU -- Add int8/int8 ...rmer/decoder/layers.5/attentions.2/Unsqueeze_36 CPU -- Reshape int8/int8 ...nsformer/decoder/layers.5/attentions.2/Expand_1 CPU -- Expand int8/int8 ...ransformer/decoder/layers.5/attentions.2/Add_10 CPU -- Add int8/int8 ...rmer/decoder/layers.5/attentions.2/Unsqueeze_40 CPU -- Reshape int8/int8 ...nsformer/decoder/layers.5/attentions.2/Expand_2 CPU -- Expand int8/int8 ...ransformer/decoder/layers.5/attentions.2/Add_11 CPU -- Add int8/int8 ...rmer/decoder/layers.5/attentions.2/Unsqueeze_44 CPU -- Reshape int8/int8 ...nsformer/decoder/layers.5/attentions.2/Expand_3 CPU -- Expand int8/int8 ...coder/layers.5/attentions.2/GatherElements_Cast CPU -- Cast int8/int32 ...er/decoder/layers.5/attentions.2/GatherElements BPU id(23) GatherElements int8/int8 ...der/layers.5/attentions.2/GatherElements_1_Cast CPU -- Cast int8/int32 .../decoder/layers.5/attentions.2/GatherElements_1 BPU id(23) GatherElements int8/int8 ...der/layers.5/attentions.2/GatherElements_2_Cast CPU -- Cast int8/int32 .../decoder/layers.5/attentions.2/GatherElements_2 BPU id(23) GatherElements int8/int8 ...der/layers.5/attentions.2/GatherElements_3_Cast CPU -- Cast int8/int32 .../decoder/layers.5/attentions.2/GatherElements_3 BPU id(23) GatherElements int8/int8 ...ransformer/decoder/layers.5/attentions.2/Mul_14 BPU id(23) HzSElementwiseMul int8/int8 ...ransformer/decoder/layers.5/attentions.2/Mul_15 BPU id(23) HzSElementwiseMul int8/int8 ...ransformer/decoder/layers.5/attentions.2/Add_12 BPU id(23) HzSElementwiseAdd int8/int8 ...ransformer/decoder/layers.5/attentions.2/Mul_16 BPU id(23) HzSElementwiseMul int8/int8 ...ransformer/decoder/layers.5/attentions.2/Add_13 BPU id(23) HzSElementwiseAdd int8/int8 ...ransformer/decoder/layers.5/attentions.2/Mul_17 BPU id(23) HzSElementwiseMul int8/int8 ...ransformer/decoder/layers.5/attentions.2/Add_14 BPU id(23) HzSElementwiseAdd int8/int8 ...former/decoder/layers.5/attentions.2/Reshape_16 BPU id(23) Reshape int8/int8 ...ormer/decoder/layers.5/attentions.2/Transpose_5 BPU id(23) Transpose int8/int8 ...former/decoder/layers.5/attentions.2/Reshape_17 BPU id(23) Reshape int8/int8 ...ransformer/decoder/layers.5/attentions.2/Mul_18 BPU id(23) HzSElementwiseMul int8/int8 ...sformer/decoder/layers.5/attentions.2/ReduceSum BPU id(23) HzSQuantizedReduceSum int8/int8 ...decoder/layers.5/attentions.2/ReduceSum_reshape BPU id(23) Reshape int8/int8 ...ecoder/layers.5/attentions.2/output_proj/MatMul BPU id(23) HzSQuantizedConv int8/int8 .../attentions.2/output_proj/MatMul_reshape_output BPU id(23) Reshape int8/int8 ...ransformer/decoder/layers.5/attentions.2/Add_15 BPU id(23) HzSElementwiseAdd int8/int8 ...transformer/decoder/layers.5/norms.2/ReduceMean BPU id(23) HzSQuantizedReduceMean int8/int8 ...x_head/transformer/decoder/layers.5/norms.2/Sub BPU id(23) HzSElementwiseSub int8/int8 ...x_head/transformer/decoder/layers.5/norms.2/Pow BPU id(23) HzLut int8/int8 ...ansformer/decoder/layers.5/norms.2/ReduceMean_1 BPU id(23) HzSQuantizedReduceMean int8/int8 ...sformer/decoder/layers.5/norms.2/Div_reciprocal BPU id(23) HzLut int8/int8 ...ad/transformer/decoder/layers.5/norms.2/Div_mul BPU id(23) HzSElementwiseMul int8/int8 ...x_head/transformer/decoder/layers.5/norms.2/Mul BPU id(23) HzSElementwiseMul int8/int8 ...head/transformer/decoder/layers.5/norms.2/Add_1 BPU id(23) HzSElementwiseAdd int8/int8 ...layers/layers.0/layers.0.0/MatMul_reshape_input BPU id(23) Reshape int8/int8 ...yers.5/ffns.0/layers/layers.0/layers.0.0/MatMul BPU id(23) HzSQuantizedConv int8/int8 .../decoder/layers.5/ffns.0/layers/layers.1/MatMul BPU id(23) HzSQuantizedConv int8/int8 ....5/ffns.0/layers/layers.1/MatMul_reshape_output BPU id(23) Reshape int8/int8 ...ox_head/transformer/decoder/layers.5/ffns.0/Add BPU id(23) HzSElementwiseAdd int8/int8 ...transformer/decoder/layers.5/norms.3/ReduceMean BPU id(23) HzSQuantizedReduceMean int8/int8 ...x_head/transformer/decoder/layers.5/norms.3/Sub BPU id(23) HzSElementwiseSub int8/int8 ...x_head/transformer/decoder/layers.5/norms.3/Pow BPU id(23) HzLut int8/int8 ...ansformer/decoder/layers.5/norms.3/ReduceMean_1 BPU id(23) HzSQuantizedReduceMean int8/int8 ...sformer/decoder/layers.5/norms.3/Div_reciprocal BPU id(23) HzLut int8/int8 ...ad/transformer/decoder/layers.5/norms.3/Div_mul BPU id(23) HzSElementwiseMul int8/int8 ...x_head/transformer/decoder/layers.5/norms.3/Mul BPU id(23) HzSElementwiseMul int8/int8 ...head/transformer/decoder/layers.5/norms.3/Add_1 BPU id(23) HzSElementwiseAdd int8/int8 ...ox_head/transformer/decoder/Transpose_5_reshape BPU id(23) Reshape int8/int8 .../decoder/reg_branches.5/reg_branches.5.0/MatMul BPU id(23) HzSQuantizedConv int8/int8 .../decoder/reg_branches.5/reg_branches.5.2/MatMul BPU id(23) HzSQuantizedConv int8/int8 ...put_0_reshape_transpose_calibrated_TO_FUSE_RELU BPU id(19) HzSQuantizedConv int8/int8 /pts_bbox_head/transformer/decoder/Clip_16 BPU id(19) HzLut int8/int8 /pts_bbox_head/transformer/decoder/Sub_5 BPU id(19) HzSElementwiseSub int8/int8 /pts_bbox_head/transformer/decoder/Clip_17 BPU id(19) HzLut int8/int8 ..._bbox_head/transformer/decoder/Div_5_reciprocal BPU id(19) HzLut int8/int8 /pts_bbox_head/transformer/decoder/Div_5_mul BPU id(19) HzSElementwiseMul int8/int8 /pts_bbox_head/transformer/decoder/Log_5 BPU id(19) HzLut int8/int8 .../decoder/reg_branches.5/reg_branches.5.4/MatMul BPU id(23) HzSQuantizedConv int8/int8 /pts_bbox_head/transformer/decoder/Sigmoid_5 BPU id(23) HzLut int8/int8 /pts_bbox_head/transformer/decoder/Unsqueeze_6 BPU id(3) Reshape int8/int8 /pts_bbox_head/transformer/decoder/Unsqueeze_7 BPU id(7) Reshape int8/int8 /pts_bbox_head/transformer/decoder/Unsqueeze_8 BPU id(11) Reshape int8/int8 /pts_bbox_head/transformer/decoder/Unsqueeze_9 BPU id(15) Reshape int8/int8 /pts_bbox_head/transformer/decoder/Unsqueeze_10 BPU id(19) Reshape int8/int8 /pts_bbox_head/transformer/decoder/Unsqueeze_11 BPU id(23) Reshape int8/int8 /pts_bbox_head/transformer/decoder/Concat BPU id(23) Concat int8/int8 /pts_bbox_head/transformer/decoder/Unsqueeze_12 BPU id(3) Reshape int8/int8 /pts_bbox_head/transformer/decoder/Unsqueeze_13 BPU id(7) Reshape int8/int8 /pts_bbox_head/transformer/decoder/Unsqueeze_14 BPU id(11) Reshape int8/int8 /pts_bbox_head/transformer/decoder/Unsqueeze_15 BPU id(15) Reshape int8/int8 /pts_bbox_head/transformer/decoder/Unsqueeze_16 BPU id(19) Reshape int8/int8 ...igmoid_5_output_0_reshape_calibrated_Requantize BPU id(23) HzRequantize int8/int8 /pts_bbox_head/transformer/decoder/Concat_1 BPU id(23) Concat int8/int8 /pts_bbox_head/Gather BPU id(23) Gather int8/int8 /pts_bbox_head/Reshape BPU id(23) Reshape int8/int8 /pts_bbox_head/ReduceMean BPU id(23) HzSQuantizedReduceMean int8/int8 ...box_head/cls_branches.0/cls_branches.0.0/MatMul BPU id(23) HzSQuantizedConv int8/int8 ...head/cls_branches.0/cls_branches.0.1/ReduceMean BPU id(23) HzSQuantizedReduceMean int8/int8 /pts_bbox_head/cls_branches.0/cls_branches.0.1/Sub BPU id(23) HzSElementwiseSub int8/int8 /pts_bbox_head/cls_branches.0/cls_branches.0.1/Pow BPU id(23) HzLut int8/int8 ...ad/cls_branches.0/cls_branches.0.1/ReduceMean_1 BPU id(23) HzSQuantizedReduceMean int8/int8 .../cls_branches.0/cls_branches.0.1/Div_reciprocal BPU id(23) HzLut int8/int8 ...ox_head/cls_branches.0/cls_branches.0.1/Div_mul BPU id(23) HzSElementwiseMul int8/int8 /pts_bbox_head/cls_branches.0/cls_branches.0.1/Mul BPU id(23) HzSQuantizedConv int8/int8 ...box_head/cls_branches.0/cls_branches.0.3/MatMul BPU id(23) HzSQuantizedConv int8/int8 ...head/cls_branches.0/cls_branches.0.4/ReduceMean BPU id(23) HzSQuantizedReduceMean int8/int8 /pts_bbox_head/cls_branches.0/cls_branches.0.4/Sub BPU id(23) HzSElementwiseSub int8/int8 /pts_bbox_head/cls_branches.0/cls_branches.0.4/Pow BPU id(23) HzLut int8/int8 ...ad/cls_branches.0/cls_branches.0.4/ReduceMean_1 BPU id(23) HzSQuantizedReduceMean int8/int8 .../cls_branches.0/cls_branches.0.4/Div_reciprocal BPU id(23) HzLut int8/int8 ...ox_head/cls_branches.0/cls_branches.0.4/Div_mul BPU id(23) HzSElementwiseMul int8/int8 /pts_bbox_head/cls_branches.0/cls_branches.0.4/Mul BPU id(23) HzSQuantizedConv int8/int8 ...box_head/cls_branches.0/cls_branches.0.6/MatMul BPU id(23) HzSQuantizedConv int8/int8 ...ranches.0/reg_branches.0.0/MatMul_reshape_input BPU id(23) Reshape int8/int8 ...box_head/reg_branches.0/reg_branches.0.0/MatMul BPU id(23) HzSQuantizedConv int8/int8 ...box_head/reg_branches.0/reg_branches.0.2/MatMul BPU id(23) HzSQuantizedConv int8/int8 ...box_head/reg_branches.0/reg_branches.0.4/MatMul BPU id(23) HzSQuantizedConv int8/int8 /pts_bbox_head/Sigmoid BPU id(23) HzLut int8/int8 ...igmoid/pts_bbox_head/Sigmoid_output_0_Reshape_0 BPU id(23) Reshape int8/int8 /pts_bbox_head/Gather_1 BPU id(23) Gather int8/int8 /pts_bbox_head/Gather_1_reshape BPU id(23) Reshape int8/int8 /pts_bbox_head/Gather_2 BPU id(23) Gather int8/int8 /pts_bbox_head/Gather_2_reshape BPU id(23) Reshape int8/int8 /pts_bbox_head/ReduceMin CPU -- ReduceMin float/float /pts_bbox_head/ReduceMax BPU id(23) HzQuantizedReduceMax int8/int8 /pts_bbox_head/ReduceMin_1 CPU -- ReduceMin float/float /pts_bbox_head/ReduceMax_1 BPU id(23) HzQuantizedReduceMax int8/int8 ...x_head/ReduceMax_output_0_calibrated_Requantize BPU id(23) HzRequantize int8/int8 ...head/ReduceMax_1_output_0_calibrated_Requantize BPU id(23) HzRequantize int8/int8 /pts_bbox_head/Concat_7 BPU id(24) Concat int8/int8 /pts_bbox_head/Split BPU id(24) Split int8/int8 /pts_bbox_head/Add_2 BPU id(24) HzSElementwiseAdd int8/int8 /pts_bbox_head/Div_1_mul BPU id(24) HzSElementwiseMul int8/int8 /pts_bbox_head/Add_3 BPU id(24) HzSElementwiseAdd int8/int8 /pts_bbox_head/Div_2_mul BPU id(24) HzSElementwiseMul int8/int8 /pts_bbox_head/Sub_1 BPU id(24) HzSElementwiseSub int8/int8 /pts_bbox_head/Sub_2 BPU id(24) HzSElementwiseSub int8/int8 /pts_bbox_head/Concat_8 BPU id(24) Concat int8/int8 /pts_bbox_head/Gather_3 BPU id(23) Gather int8/int8 ..._head/Gather_3_output_0_calibrated_TO_FUSE_RELU BPU id(23) HzSQuantizedConv int8/int8 /pts_bbox_head/Clip_4 BPU id(23) HzLut int8/int8 /pts_bbox_head/Sub_3 BPU id(23) HzSElementwiseSub int8/int8 /pts_bbox_head/Clip_5 BPU id(23) HzLut int8/int8 /pts_bbox_head/Div_3_reciprocal BPU id(23) HzLut int8/int8 /pts_bbox_head/Div_3_mul BPU id(23) HzSElementwiseMul int8/int8 /pts_bbox_head/Log_1 BPU id(23) HzLut int8/int8 /pts_bbox_head/Gather_4 BPU id(23) Gather int8/int8 /pts_bbox_head/Reshape_2 BPU id(23) Reshape int8/int8 /pts_bbox_head/ReduceMean_1 BPU id(23) HzSQuantizedReduceMean int8/int8 ...box_head/cls_branches.1/cls_branches.1.0/MatMul BPU id(23) HzSQuantizedConv int8/int8 ...head/cls_branches.1/cls_branches.1.1/ReduceMean BPU id(23) HzSQuantizedReduceMean int8/int8 /pts_bbox_head/cls_branches.1/cls_branches.1.1/Sub BPU id(23) HzSElementwiseSub int8/int8 /pts_bbox_head/cls_branches.1/cls_branches.1.1/Pow BPU id(23) HzLut int8/int8 ...ad/cls_branches.1/cls_branches.1.1/ReduceMean_1 BPU id(23) HzSQuantizedReduceMean int8/int8 .../cls_branches.1/cls_branches.1.1/Div_reciprocal BPU id(23) HzLut int8/int8 ...ox_head/cls_branches.1/cls_branches.1.1/Div_mul BPU id(23) HzSElementwiseMul int8/int8 /pts_bbox_head/cls_branches.1/cls_branches.1.1/Mul BPU id(23) HzSQuantizedConv int8/int8 ...box_head/cls_branches.1/cls_branches.1.3/MatMul BPU id(23) HzSQuantizedConv int8/int8 ...head/cls_branches.1/cls_branches.1.4/ReduceMean BPU id(23) HzSQuantizedReduceMean int8/int8 /pts_bbox_head/cls_branches.1/cls_branches.1.4/Sub BPU id(23) HzSElementwiseSub int8/int8 /pts_bbox_head/cls_branches.1/cls_branches.1.4/Pow BPU id(23) HzLut int8/int8 ...ad/cls_branches.1/cls_branches.1.4/ReduceMean_1 BPU id(23) HzSQuantizedReduceMean int8/int8 .../cls_branches.1/cls_branches.1.4/Div_reciprocal BPU id(23) HzLut int8/int8 ...ox_head/cls_branches.1/cls_branches.1.4/Div_mul BPU id(23) HzSElementwiseMul int8/int8 /pts_bbox_head/cls_branches.1/cls_branches.1.4/Mul BPU id(23) HzSQuantizedConv int8/int8 ...box_head/cls_branches.1/cls_branches.1.6/MatMul BPU id(23) HzSQuantizedConv int8/int8 ...ranches.1/reg_branches.1.0/MatMul_reshape_input BPU id(23) Reshape int8/int8 ...box_head/reg_branches.1/reg_branches.1.0/MatMul BPU id(23) HzSQuantizedConv int8/int8 ...box_head/reg_branches.1/reg_branches.1.2/MatMul BPU id(23) HzSQuantizedConv int8/int8 ...box_head/reg_branches.1/reg_branches.1.4/MatMul BPU id(23) HzSQuantizedConv int8/int8 UNIT_CONV_FOR_/pts_bbox_head/Add_4 BPU id(23) HzSQuantizedConv int8/int8 /pts_bbox_head/Sigmoid_1 BPU id(23) HzLut int8/int8 /pts_bbox_head/Gather_5 BPU id(23) Gather int8/int8 /pts_bbox_head/Gather_5_reshape BPU id(23) Reshape int8/int8 /pts_bbox_head/Gather_6 BPU id(23) Gather int8/int8 /pts_bbox_head/Gather_6_reshape BPU id(23) Reshape int8/int8 /pts_bbox_head/ReduceMin_2 CPU -- ReduceMin float/float /pts_bbox_head/ReduceMax_2 BPU id(23) HzQuantizedReduceMax int8/int8 /pts_bbox_head/ReduceMin_3 CPU -- ReduceMin float/float /pts_bbox_head/ReduceMax_3 BPU id(23) HzQuantizedReduceMax int8/int8 ...head/ReduceMax_2_output_0_calibrated_Requantize BPU id(23) HzRequantize int8/int8 ...head/ReduceMax_3_output_0_calibrated_Requantize BPU id(23) HzRequantize int8/int8 /pts_bbox_head/Concat_10 BPU id(24) Concat int8/int8 /pts_bbox_head/Split_1 BPU id(24) Split int8/int8 /pts_bbox_head/Add_5 BPU id(24) HzSElementwiseAdd int8/int8 /pts_bbox_head/Div_4_mul BPU id(24) HzSElementwiseMul int8/int8 /pts_bbox_head/Add_6 BPU id(24) HzSElementwiseAdd int8/int8 /pts_bbox_head/Div_5_mul BPU id(24) HzSElementwiseMul int8/int8 /pts_bbox_head/Sub_4 BPU id(24) HzSElementwiseSub int8/int8 /pts_bbox_head/Sub_5 BPU id(24) HzSElementwiseSub int8/int8 /pts_bbox_head/Concat_11 BPU id(24) Concat int8/int8 /pts_bbox_head/Gather_7 BPU id(23) Gather int8/int8 ..._head/Gather_7_output_0_calibrated_TO_FUSE_RELU BPU id(23) HzSQuantizedConv int8/int8 /pts_bbox_head/Clip_7 BPU id(23) HzLut int8/int8 /pts_bbox_head/Sub_6 BPU id(23) HzSElementwiseSub int8/int8 /pts_bbox_head/Clip_8 BPU id(23) HzLut int8/int8 /pts_bbox_head/Div_6_reciprocal BPU id(23) HzLut int8/int8 /pts_bbox_head/Div_6_mul BPU id(23) HzSElementwiseMul int8/int8 /pts_bbox_head/Log_2 BPU id(23) HzLut int8/int8 /pts_bbox_head/Gather_8 BPU id(23) Gather int8/int8 /pts_bbox_head/Reshape_4 BPU id(23) Reshape int8/int8 /pts_bbox_head/ReduceMean_2 BPU id(23) HzSQuantizedReduceMean int8/int8 ...box_head/cls_branches.2/cls_branches.2.0/MatMul BPU id(23) HzSQuantizedConv int8/int8 ...head/cls_branches.2/cls_branches.2.1/ReduceMean BPU id(23) HzSQuantizedReduceMean int8/int8 /pts_bbox_head/cls_branches.2/cls_branches.2.1/Sub BPU id(23) HzSElementwiseSub int8/int8 /pts_bbox_head/cls_branches.2/cls_branches.2.1/Pow BPU id(23) HzLut int8/int8 ...ad/cls_branches.2/cls_branches.2.1/ReduceMean_1 BPU id(23) HzSQuantizedReduceMean int8/int8 .../cls_branches.2/cls_branches.2.1/Div_reciprocal BPU id(23) HzLut int8/int8 ...ox_head/cls_branches.2/cls_branches.2.1/Div_mul BPU id(23) HzSElementwiseMul int8/int8 /pts_bbox_head/cls_branches.2/cls_branches.2.1/Mul BPU id(23) HzSQuantizedConv int8/int8 ...box_head/cls_branches.2/cls_branches.2.3/MatMul BPU id(23) HzSQuantizedConv int8/int8 ...head/cls_branches.2/cls_branches.2.4/ReduceMean BPU id(23) HzSQuantizedReduceMean int8/int8 /pts_bbox_head/cls_branches.2/cls_branches.2.4/Sub BPU id(23) HzSElementwiseSub int8/int8 /pts_bbox_head/cls_branches.2/cls_branches.2.4/Pow BPU id(23) HzLut int8/int8 ...ad/cls_branches.2/cls_branches.2.4/ReduceMean_1 BPU id(23) HzSQuantizedReduceMean int8/int8 .../cls_branches.2/cls_branches.2.4/Div_reciprocal BPU id(23) HzLut int8/int8 ...ox_head/cls_branches.2/cls_branches.2.4/Div_mul BPU id(23) HzSElementwiseMul int8/int8 /pts_bbox_head/cls_branches.2/cls_branches.2.4/Mul BPU id(23) HzSQuantizedConv int8/int8 ...box_head/cls_branches.2/cls_branches.2.6/MatMul BPU id(23) HzSQuantizedConv int8/int8 ...ranches.2/reg_branches.2.0/MatMul_reshape_input BPU id(23) Reshape int8/int8 ...box_head/reg_branches.2/reg_branches.2.0/MatMul BPU id(23) HzSQuantizedConv int8/int8 ...box_head/reg_branches.2/reg_branches.2.2/MatMul BPU id(23) HzSQuantizedConv int8/int8 ...box_head/reg_branches.2/reg_branches.2.4/MatMul BPU id(23) HzSQuantizedConv int8/int8 UNIT_CONV_FOR_/pts_bbox_head/Add_7 BPU id(23) HzSQuantizedConv int8/int8 /pts_bbox_head/Sigmoid_2 BPU id(23) HzLut int8/int8 /pts_bbox_head/Gather_9 BPU id(23) Gather int8/int8 /pts_bbox_head/Gather_9_reshape BPU id(23) Reshape int8/int8 /pts_bbox_head/Gather_10 BPU id(23) Gather int8/int8 /pts_bbox_head/Gather_10_reshape BPU id(23) Reshape int8/int8 /pts_bbox_head/ReduceMin_4 CPU -- ReduceMin float/float /pts_bbox_head/ReduceMax_4 BPU id(23) HzQuantizedReduceMax int8/int8 /pts_bbox_head/ReduceMin_5 CPU -- ReduceMin float/float /pts_bbox_head/ReduceMax_5 BPU id(23) HzQuantizedReduceMax int8/int8 ...head/ReduceMax_4_output_0_calibrated_Requantize BPU id(23) HzRequantize int8/int8 ...head/ReduceMax_5_output_0_calibrated_Requantize BPU id(23) HzRequantize int8/int8 /pts_bbox_head/Concat_13 BPU id(24) Concat int8/int8 /pts_bbox_head/Split_2 BPU id(24) Split int8/int8 /pts_bbox_head/Add_8 BPU id(24) HzSElementwiseAdd int8/int8 /pts_bbox_head/Div_7_mul BPU id(24) HzSElementwiseMul int8/int8 /pts_bbox_head/Add_9 BPU id(24) HzSElementwiseAdd int8/int8 /pts_bbox_head/Div_8_mul BPU id(24) HzSElementwiseMul int8/int8 /pts_bbox_head/Sub_7 BPU id(24) HzSElementwiseSub int8/int8 /pts_bbox_head/Sub_8 BPU id(24) HzSElementwiseSub int8/int8 /pts_bbox_head/Concat_14 BPU id(24) Concat int8/int8 /pts_bbox_head/Gather_11 BPU id(23) Gather int8/int8 ...head/Gather_11_output_0_calibrated_TO_FUSE_RELU BPU id(23) HzSQuantizedConv int8/int8 /pts_bbox_head/Clip_10 BPU id(23) HzLut int8/int8 /pts_bbox_head/Sub_9 BPU id(23) HzSElementwiseSub int8/int8 /pts_bbox_head/Clip_11 BPU id(23) HzLut int8/int8 /pts_bbox_head/Div_9_reciprocal BPU id(23) HzLut int8/int8 /pts_bbox_head/Div_9_mul BPU id(23) HzSElementwiseMul int8/int8 /pts_bbox_head/Log_3 BPU id(23) HzLut int8/int8 /pts_bbox_head/Gather_12 BPU id(23) Gather int8/int8 /pts_bbox_head/Reshape_6 BPU id(23) Reshape int8/int8 /pts_bbox_head/ReduceMean_3 BPU id(23) HzSQuantizedReduceMean int8/int8 ...box_head/cls_branches.3/cls_branches.3.0/MatMul BPU id(23) HzSQuantizedConv int8/int8 ...head/cls_branches.3/cls_branches.3.1/ReduceMean BPU id(23) HzSQuantizedReduceMean int8/int8 /pts_bbox_head/cls_branches.3/cls_branches.3.1/Sub BPU id(23) HzSElementwiseSub int8/int8 /pts_bbox_head/cls_branches.3/cls_branches.3.1/Pow BPU id(23) HzLut int8/int8 ...ad/cls_branches.3/cls_branches.3.1/ReduceMean_1 BPU id(23) HzSQuantizedReduceMean int8/int8 .../cls_branches.3/cls_branches.3.1/Div_reciprocal BPU id(23) HzLut int8/int8 ...ox_head/cls_branches.3/cls_branches.3.1/Div_mul BPU id(23) HzSElementwiseMul int8/int8 /pts_bbox_head/cls_branches.3/cls_branches.3.1/Mul BPU id(23) HzSQuantizedConv int8/int8 ...box_head/cls_branches.3/cls_branches.3.3/MatMul BPU id(23) HzSQuantizedConv int8/int8 ...head/cls_branches.3/cls_branches.3.4/ReduceMean BPU id(23) HzSQuantizedReduceMean int8/int8 /pts_bbox_head/cls_branches.3/cls_branches.3.4/Sub BPU id(23) HzSElementwiseSub int8/int8 /pts_bbox_head/cls_branches.3/cls_branches.3.4/Pow BPU id(23) HzLut int8/int8 ...ad/cls_branches.3/cls_branches.3.4/ReduceMean_1 BPU id(23) HzSQuantizedReduceMean int8/int8 .../cls_branches.3/cls_branches.3.4/Div_reciprocal BPU id(23) HzLut int8/int8 ...ox_head/cls_branches.3/cls_branches.3.4/Div_mul BPU id(23) HzSElementwiseMul int8/int8 /pts_bbox_head/cls_branches.3/cls_branches.3.4/Mul BPU id(23) HzSQuantizedConv int8/int8 ...box_head/cls_branches.3/cls_branches.3.6/MatMul BPU id(23) HzSQuantizedConv int8/int8 ...ranches.3/reg_branches.3.0/MatMul_reshape_input BPU id(23) Reshape int8/int8 ...box_head/reg_branches.3/reg_branches.3.0/MatMul BPU id(23) HzSQuantizedConv int8/int8 ...box_head/reg_branches.3/reg_branches.3.2/MatMul BPU id(23) HzSQuantizedConv int8/int8 ...box_head/reg_branches.3/reg_branches.3.4/MatMul BPU id(23) HzSQuantizedConv int8/int8 UNIT_CONV_FOR_/pts_bbox_head/Add_10 BPU id(23) HzSQuantizedConv int8/int8 /pts_bbox_head/Sigmoid_3 BPU id(23) HzLut int8/int8 /pts_bbox_head/Gather_13 BPU id(23) Gather int8/int8 /pts_bbox_head/Gather_13_reshape BPU id(23) Reshape int8/int8 /pts_bbox_head/Gather_14 BPU id(23) Gather int8/int8 /pts_bbox_head/Gather_14_reshape BPU id(23) Reshape int8/int8 /pts_bbox_head/ReduceMin_6 CPU -- ReduceMin float/float /pts_bbox_head/ReduceMax_6 BPU id(23) HzQuantizedReduceMax int8/int8 /pts_bbox_head/ReduceMin_7 CPU -- ReduceMin float/float /pts_bbox_head/ReduceMax_7 BPU id(23) HzQuantizedReduceMax int8/int8 ...head/ReduceMax_6_output_0_calibrated_Requantize BPU id(23) HzRequantize int8/int8 ...head/ReduceMax_7_output_0_calibrated_Requantize BPU id(23) HzRequantize int8/int8 /pts_bbox_head/Concat_16 BPU id(24) Concat int8/int8 /pts_bbox_head/Split_3 BPU id(24) Split int8/int8 /pts_bbox_head/Add_11 BPU id(24) HzSElementwiseAdd int8/int8 /pts_bbox_head/Div_10_mul BPU id(24) HzSElementwiseMul int8/int8 /pts_bbox_head/Add_12 BPU id(24) HzSElementwiseAdd int8/int8 /pts_bbox_head/Div_11_mul BPU id(24) HzSElementwiseMul int8/int8 /pts_bbox_head/Sub_10 BPU id(24) HzSElementwiseSub int8/int8 /pts_bbox_head/Sub_11 BPU id(24) HzSElementwiseSub int8/int8 /pts_bbox_head/Concat_17 BPU id(24) Concat int8/int8 /pts_bbox_head/Gather_15 BPU id(23) Gather int8/int8 ...head/Gather_15_output_0_calibrated_TO_FUSE_RELU BPU id(23) HzSQuantizedConv int8/int8 /pts_bbox_head/Clip_13 BPU id(23) HzLut int8/int8 /pts_bbox_head/Sub_12 BPU id(23) HzSElementwiseSub int8/int8 /pts_bbox_head/Clip_14 BPU id(23) HzLut int8/int8 /pts_bbox_head/Div_12_reciprocal BPU id(23) HzLut int8/int8 /pts_bbox_head/Div_12_mul BPU id(23) HzSElementwiseMul int8/int8 /pts_bbox_head/Log_4 BPU id(23) HzLut int8/int8 /pts_bbox_head/Gather_16 BPU id(23) Gather int8/int8 /pts_bbox_head/Reshape_8 BPU id(23) Reshape int8/int8 /pts_bbox_head/ReduceMean_4 BPU id(23) HzSQuantizedReduceMean int8/int8 ...box_head/cls_branches.4/cls_branches.4.0/MatMul BPU id(23) HzSQuantizedConv int8/int8 ...head/cls_branches.4/cls_branches.4.1/ReduceMean BPU id(23) HzSQuantizedReduceMean int8/int8 /pts_bbox_head/cls_branches.4/cls_branches.4.1/Sub BPU id(23) HzSElementwiseSub int8/int8 /pts_bbox_head/cls_branches.4/cls_branches.4.1/Pow BPU id(23) HzLut int8/int8 ...ad/cls_branches.4/cls_branches.4.1/ReduceMean_1 BPU id(23) HzSQuantizedReduceMean int8/int8 .../cls_branches.4/cls_branches.4.1/Div_reciprocal BPU id(23) HzLut int8/int8 ...ox_head/cls_branches.4/cls_branches.4.1/Div_mul BPU id(23) HzSElementwiseMul int8/int8 /pts_bbox_head/cls_branches.4/cls_branches.4.1/Mul BPU id(23) HzSQuantizedConv int8/int8 ...box_head/cls_branches.4/cls_branches.4.3/MatMul BPU id(23) HzSQuantizedConv int8/int8 ...head/cls_branches.4/cls_branches.4.4/ReduceMean BPU id(23) HzSQuantizedReduceMean int8/int8 /pts_bbox_head/cls_branches.4/cls_branches.4.4/Sub BPU id(23) HzSElementwiseSub int8/int8 /pts_bbox_head/cls_branches.4/cls_branches.4.4/Pow BPU id(23) HzLut int8/int8 ...ad/cls_branches.4/cls_branches.4.4/ReduceMean_1 BPU id(23) HzSQuantizedReduceMean int8/int8 .../cls_branches.4/cls_branches.4.4/Div_reciprocal BPU id(23) HzLut int8/int8 ...ox_head/cls_branches.4/cls_branches.4.4/Div_mul BPU id(23) HzSElementwiseMul int8/int8 /pts_bbox_head/cls_branches.4/cls_branches.4.4/Mul BPU id(23) HzSQuantizedConv int8/int8 ...box_head/cls_branches.4/cls_branches.4.6/MatMul BPU id(23) HzSQuantizedConv int8/int8 ...ranches.4/reg_branches.4.0/MatMul_reshape_input BPU id(23) Reshape int8/int8 ...box_head/reg_branches.4/reg_branches.4.0/MatMul BPU id(23) HzSQuantizedConv int8/int8 ...box_head/reg_branches.4/reg_branches.4.2/MatMul BPU id(23) HzSQuantizedConv int8/int8 ...box_head/reg_branches.4/reg_branches.4.4/MatMul BPU id(23) HzSQuantizedConv int8/int8 UNIT_CONV_FOR_/pts_bbox_head/Add_13 BPU id(23) HzSQuantizedConv int8/int8 /pts_bbox_head/Sigmoid_4 BPU id(23) HzLut int8/int8 /pts_bbox_head/Gather_17 BPU id(23) Gather int8/int8 /pts_bbox_head/Gather_17_reshape BPU id(23) Reshape int8/int8 /pts_bbox_head/Gather_18 BPU id(23) Gather int8/int8 /pts_bbox_head/Gather_18_reshape BPU id(23) Reshape int8/int8 /pts_bbox_head/ReduceMin_8 CPU -- ReduceMin float/float /pts_bbox_head/ReduceMax_8 BPU id(23) HzQuantizedReduceMax int8/int8 /pts_bbox_head/ReduceMin_9 CPU -- ReduceMin float/float /pts_bbox_head/ReduceMax_9 BPU id(23) HzQuantizedReduceMax int8/int8 ...head/ReduceMax_8_output_0_calibrated_Requantize BPU id(23) HzRequantize int8/int8 ...head/ReduceMax_9_output_0_calibrated_Requantize BPU id(23) HzRequantize int8/int8 /pts_bbox_head/Concat_19 BPU id(24) Concat int8/int8 /pts_bbox_head/Split_4 BPU id(24) Split int8/int8 /pts_bbox_head/Add_14 BPU id(24) HzSElementwiseAdd int8/int8 /pts_bbox_head/Div_13_mul BPU id(24) HzSElementwiseMul int8/int8 /pts_bbox_head/Add_15 BPU id(24) HzSElementwiseAdd int8/int8 /pts_bbox_head/Div_14_mul BPU id(24) HzSElementwiseMul int8/int8 /pts_bbox_head/Sub_13 BPU id(24) HzSElementwiseSub int8/int8 /pts_bbox_head/Sub_14 BPU id(24) HzSElementwiseSub int8/int8 /pts_bbox_head/Concat_20 BPU id(24) Concat int8/int8 /pts_bbox_head/Gather_19 BPU id(23) Gather int8/int8 ...head/Gather_19_output_0_calibrated_TO_FUSE_RELU BPU id(23) HzSQuantizedConv int8/int8 /pts_bbox_head/Clip_16 BPU id(23) HzLut int8/int8 /pts_bbox_head/Sub_15 BPU id(23) HzSElementwiseSub int8/int8 /pts_bbox_head/Clip_17 BPU id(23) HzLut int8/int8 /pts_bbox_head/Div_15_reciprocal BPU id(23) HzLut int8/int8 /pts_bbox_head/Div_15_mul BPU id(23) HzSElementwiseMul int8/int8 /pts_bbox_head/Log_5 BPU id(23) HzLut int8/int8 /pts_bbox_head/Gather_20 BPU id(23) Gather int8/int8 /pts_bbox_head/Reshape_10 BPU id(23) Reshape int8/int8 /pts_bbox_head/ReduceMean_5 BPU id(23) HzSQuantizedReduceMean int8/int8 ...box_head/cls_branches.5/cls_branches.5.0/MatMul BPU id(23) HzSQuantizedConv int8/int8 ...head/cls_branches.5/cls_branches.5.1/ReduceMean BPU id(23) HzSQuantizedReduceMean int8/int8 /pts_bbox_head/cls_branches.5/cls_branches.5.1/Sub BPU id(23) HzSElementwiseSub int8/int8 /pts_bbox_head/cls_branches.5/cls_branches.5.1/Pow BPU id(23) HzLut int8/int8 ...ad/cls_branches.5/cls_branches.5.1/ReduceMean_1 BPU id(23) HzSQuantizedReduceMean int8/int8 .../cls_branches.5/cls_branches.5.1/Div_reciprocal BPU id(23) HzLut int8/int8 ...ox_head/cls_branches.5/cls_branches.5.1/Div_mul BPU id(23) HzSElementwiseMul int8/int8 /pts_bbox_head/cls_branches.5/cls_branches.5.1/Mul BPU id(23) HzSQuantizedConv int8/int8 ...box_head/cls_branches.5/cls_branches.5.3/MatMul BPU id(23) HzSQuantizedConv int8/int8 ...head/cls_branches.5/cls_branches.5.4/ReduceMean BPU id(23) HzSQuantizedReduceMean int8/int8 /pts_bbox_head/cls_branches.5/cls_branches.5.4/Sub BPU id(23) HzSElementwiseSub int8/int8 /pts_bbox_head/cls_branches.5/cls_branches.5.4/Pow BPU id(23) HzLut int8/int8 ...ad/cls_branches.5/cls_branches.5.4/ReduceMean_1 BPU id(23) HzSQuantizedReduceMean int8/int8 .../cls_branches.5/cls_branches.5.4/Div_reciprocal BPU id(23) HzLut int8/int8 ...ox_head/cls_branches.5/cls_branches.5.4/Div_mul BPU id(23) HzSElementwiseMul int8/int8 /pts_bbox_head/cls_branches.5/cls_branches.5.4/Mul BPU id(23) HzSQuantizedConv int8/int8 ...box_head/cls_branches.5/cls_branches.5.6/MatMul BPU id(23) HzSQuantizedConv int8/int8 ...ranches.5/reg_branches.5.0/MatMul_reshape_input BPU id(23) Reshape int8/int8 ...box_head/reg_branches.5/reg_branches.5.0/MatMul BPU id(23) HzSQuantizedConv int8/int8 ...box_head/reg_branches.5/reg_branches.5.2/MatMul BPU id(23) HzSQuantizedConv int8/int8 ...box_head/reg_branches.5/reg_branches.5.4/MatMul BPU id(23) HzSQuantizedConv int8/int8 UNIT_CONV_FOR_/pts_bbox_head/Add_16 BPU id(23) HzSQuantizedConv int8/int8 /pts_bbox_head/Sigmoid_5 BPU id(23) HzLut int8/int8 /pts_bbox_head/Gather_21 BPU id(23) Gather int8/int8 /pts_bbox_head/Gather_21_reshape BPU id(23) Reshape int8/int8 /pts_bbox_head/Gather_22 BPU id(23) Gather int8/int8 /pts_bbox_head/Gather_22_reshape BPU id(23) Reshape int8/int8 /pts_bbox_head/ReduceMin_10 CPU -- ReduceMin float/float /pts_bbox_head/ReduceMax_10 BPU id(23) HzQuantizedReduceMax int8/int8 /pts_bbox_head/ReduceMin_11 CPU -- ReduceMin float/float /pts_bbox_head/ReduceMax_11 BPU id(23) HzQuantizedReduceMax int8/int8 ...ead/ReduceMax_10_output_0_calibrated_Requantize BPU id(23) HzRequantize int8/int8 ...ead/ReduceMax_11_output_0_calibrated_Requantize BPU id(23) HzRequantize int8/int8 /pts_bbox_head/Concat_22 BPU id(24) Concat int8/int8 /pts_bbox_head/Split_5 BPU id(24) Split int8/int8 /pts_bbox_head/Add_17 BPU id(24) HzSElementwiseAdd int8/int8 /pts_bbox_head/Div_16_mul BPU id(24) HzSElementwiseMul int8/int8 /pts_bbox_head/Add_18 BPU id(24) HzSElementwiseAdd int8/int8 /pts_bbox_head/Div_17_mul BPU id(24) HzSElementwiseMul int8/int8 /pts_bbox_head/Sub_16 BPU id(24) HzSElementwiseSub int8/int8 /pts_bbox_head/Sub_17 BPU id(24) HzSElementwiseSub int8/int8 /pts_bbox_head/Concat_23 BPU id(24) Concat int8/int8 /pts_bbox_head/Concat_24 BPU id(23) Concat int8/int8 /pts_bbox_head/Unsqueeze_56 BPU id(24) Reshape int8/int8 /pts_bbox_head/Unsqueeze_57 BPU id(24) Reshape int8/int8 /pts_bbox_head/Unsqueeze_58 BPU id(24) Reshape int8/int8 /pts_bbox_head/Unsqueeze_59 BPU id(24) Reshape int8/int8 /pts_bbox_head/Unsqueeze_60 BPU id(24) Reshape int8/int8 /pts_bbox_head/Unsqueeze_61 BPU id(24) Reshape int8/int8 /pts_bbox_head/Concat_25 BPU id(24) Concat int8/int8 /Gather_1 BPU id(23) Gather int8/int8 /Gather_2 BPU id(24) Gather int8/int8 /Sigmoid BPU id(23) HzLut int8/int8 /TopK BPU id(23) HzQuantizedTopK int8/int32 /TopK_output_1_Cast CPU -- Cast int8/int8 /Mod CPU -- Mod int8/int8 /Div CPU -- Div int8/int8 /Gather_5 CPU -- Gather float/float /Split BPU id(25) Split int8/int8 /Mul BPU id(25) HzSElementwiseMul int8/int8 /Sub BPU id(25) HzSElementwiseSub int8/int16 /Mul_1 BPU id(25) HzSElementwiseMul int8/int8 /Sub_1 BPU id(25) HzSElementwiseSub int8/int16 /Mul_2 BPU id(25) HzSElementwiseMul int8/int8 /Add BPU id(25) HzSElementwiseAdd int8/int16 /Mul_3 BPU id(25) HzSElementwiseMul int8/int8 /Add_1 BPU id(25) HzSElementwiseAdd int8/int16 /Concat_2 BPU id(25) Concat int16/int16 /Slice_1 BPU id(25) Slice int16/int16 /Mul_4 BPU id(25) HzSElementwiseMul int16/int8 /Add_2 BPU id(25) HzSElementwiseAdd int8/int16 /Reshape_3 BPU id(25) Reshape int16/int16 /ScatterND CPU -- ScatterND float/float /Slice_5 BPU id(26) Slice int8/int8 /Mul_7 BPU id(26) HzSElementwiseMul int8/int8 /Add_4 BPU id(26) HzSElementwiseAdd int8/int16 /Reshape_4 BPU id(26) Reshape int16/int16 /ScatterND_1 CPU -- ScatterND float/float 2025-03-13 14:11:55,586 file: build.py func: build line No: 39 End to Horizon NN Model Convert. 2025-03-13 14:11:56,156 file: hb_mapper_makertbin.py func: hb_mapper_makertbin line No: 597 start convert to *.bin file.... 2025-03-13 14:11:56,268 file: onnx2horizonrt.py func: onnx2horizonrt line No: 4300 ONNX model output num : 4 2025-03-13 14:12:00,540 file: tool_utils.py func: tool_utils line No: 131 exception in command: makertbin 2025-03-13 14:12:00,541 file: tool_utils.py func: tool_utils line No: 132 Traceback (most recent call last): File "/usr/local/lib/python3.8/dist-packages/horizon_tc_ui/hb_mapper_makertbin.py", line 745, in run build_runtime_model_wrapper(hybrid_model, file_name, File "/usr/local/lib/python3.8/dist-packages/horizon_tc_ui/hbdtort/onnx2horizonrt.py", line 4334, in build_runtime_model_wrapper make_input_type(runtime_graph, input_type, input_layout_rt) File "/usr/local/lib/python3.8/dist-packages/horizon_tc_ui/hbdtort/onnx2horizonrt.py", line 4016, in make_input_type input_type_mapping[input.type.elem_type]) KeyError: 14 During handling of the above exception, another exception occurred: Traceback (most recent call last): File "/usr/local/lib/python3.8/dist-packages/horizon_tc_ui/utils/tool_utils.py", line 129, in __decorator func(*args, **kargs) File "/usr/local/lib/python3.8/dist-packages/horizon_tc_ui/hb_mapper.py", line 154, in makertbin MakertbinRunner(config, model_type).run(__version__) File "/usr/local/lib/python3.8/dist-packages/horizon_tc_ui/hb_mapper_makertbin.py", line 749, in run raise ValueError( ValueError: *** ERROR-OCCUR-DURING {runtime.runtime_model_generation} ***, error message: 14 2025-03-13 14:12:00,541 file: tool_utils.py func: tool_utils line No: 133 *** ERROR-OCCUR-DURING {runtime.runtime_model_generation} ***, error message: 14