#说明

• 测试条件：

  • 测试开发板：J6E。

  • 测试核心数：单核。

  • 性能数据获取频率设置为：5分钟时间内性能参数的平均值。

  • Python版本：Python3.10。

  • 模型来源：OE包内 samples/ucp_tutorial/dnn/ai_benchmark/j6 路径下的模型。

• 缩写说明：

  • C = 计算量，单位为GOPs（十亿次运算/秒）。此数据通过调用 hbm_perf 接口获得。

  • FPS = 每秒帧率。此数据在开发板多线程运行ai_benchmark示例包/script路径下各模型子文件夹的 fps.sh 脚本获取，包含后处理。

  • ITC = 推理耗时，单位为ms（毫秒）。此数据在开发板单线程运行ai_benchmark示例包/script路径下各模型子文件夹的 latency.sh 脚本获取，不含后处理。

  • TCPP = 后处理耗时，单位为ms（毫秒）。此数据在开发板单线程运行ai_benchmark示例包/script路径下各模型子文件夹的 latency.sh 脚本获取。

  • RV = 单次推理读取数据量，单位为mb（兆比特）。此数据通过调用 hbm_perf 接口获得。

  • WV = 单次推理写入数据量，单位为mb（兆比特）。此数据通过调用 hbm_perf 接口获得。

#模型主要性能数据

#模型全部性能数据

#MobileNetv1

• INPUT SIZE: 1x3x224x224
• C(GOPs): 1.14
• FPS: 4753.30
• ITC(ms): 0.480
• TCPP(ms): 0.034
• RV(mb): 4.56
• WV(mb): 0.0041
• Dataset: ImageNet
• ACCURACY: Top1: 0.7374(FLOAT)/0.7298(INT8)

#MobileNetv2

• INPUT SIZE: 1x3x224x224
• C(GOPs): 0.63
• FPS: 4706.60
• ITC(ms): 0.505
• TCPP(ms): 0.035
• RV(mb): 3.95
• WV(mb): 0.0041
• Dataset: ImageNet
• ACCURACY: Top1: 0.7217(FLOAT)/0.7147(INT8)

#ResNet50

• INPUT SIZE: 1x3x224x224
• C(GOPs): 7.72
• FPS: 1131.00
• ITC(ms): 1.182
• TCPP(ms): 0.035
• RV(mb): 26.08
• WV(mb): 0.0041
• Dataset: ImageNet
• ACCURACY: Top1: 0.7703(FLOAT)/0.7678(INT8)

#GoogleNet

• INPUT SIZE: 1x3x224x224
• C(GOPs): 3.00
• FPS: 2799.50
• ITC(ms): 0.689
• TCPP(ms): 0.034
• RV(mb): 6.93
• WV(mb): 0.0041
• Dataset: ImageNet
• ACCURACY: Top1: 0.7018(FLOAT)/0.6993(INT8)
• LINKS: https://github.com/HorizonRobotics-Platform/ModelZoo/tree/master/GoogleNet

#EfficientNet_Lite0

• INPUT SIZE: 1x224x224x3
• C(GOPs): 0.77
• FPS: 4171.00
• ITC(ms): 0.582
• TCPP(ms): 0.034
• RV(mb): 5.21
• WV(mb): 0.0041
• Dataset: ImageNet
• ACCURACY: Top1: 0.7479(FLOAT)/0.7453(INT8)
• LINKS: https://github.com/tensorflow/tpu/tree/master/models/official/efficientnet/lite

#EfficientNet_Lite1

• INPUT SIZE: 1x240x240x3
• C(GOPs): 1.20
• FPS: 3215.30
• ITC(ms): 0.655
• TCPP(ms): 0.034
• RV(mb): 6.11
• WV(mb): 0.0041
• Dataset: ImageNet
• ACCURACY: Top1: 0.7652(FLOAT)/0.7609(INT8)
• LINKS: https://github.com/tensorflow/tpu/tree/master/models/official/efficientnet/lite

#EfficientNet_Lite2

• INPUT SIZE: 1x260x260x3
• C(GOPs): 1.72
• FPS: 2405.10
• ITC(ms): 0.760
• TCPP(ms): 0.034
• RV(mb): 6.95
• WV(mb): 0.0041
• Dataset: ImageNet
• ACCURACY: Top1: 0.7734(FLOAT)/0.7697(INT8)
• LINKS: https://github.com/tensorflow/tpu/tree/master/models/official/efficientnet/lite

#EfficientNet_Lite3

• INPUT SIZE: 1x280x280x3
• C(GOPs): 2.77
• FPS: 1849.30
• ITC(ms): 0.893
• TCPP(ms): 0.034
• RV(mb): 9.39
• WV(mb): 0.0041
• Dataset: ImageNet
• ACCURACY: Top1: 0.7917(FLOAT)/0.7887(INT8)
• LINKS: https://github.com/tensorflow/tpu/tree/master/models/official/efficientnet/lite

#EfficientNet_Lite4

• INPUT SIZE: 1x300x300x3
• C(GOPs): 5.11
• FPS: 1272.60
• ITC(ms): 1.127
• TCPP(ms): 0.034
• RV(mb): 14.47
• WV(mb): 0.0041
• Dataset: ImageNet
• ACCURACY: Top1: 0.8063(FLOAT)/0.8043(INT8)
• LINKS: https://github.com/tensorflow/tpu/tree/master/models/official/efficientnet/lite

#Vargconvnet

• INPUT SIZE: 1x3x224x224
• C(GOPs): 9.06
• FPS: 1497.00
• ITC(ms): 1.005
• TCPP(ms): 0.033
• RV(mb): 10.44
• WV(mb): 0.0041
• Dataset: ImageNet
• ACCURACY: Top1: 0.7793(FLOAT)/0.7762(INT8)
• LINKS: https://github.com/HorizonRobotics-Platform/ModelZoo/tree/master/VargConvNet

#Efficientnasnet_m

• INPUT SIZE: 1x3x300x300
• C(GOPs): 4.53
• FPS: 1429.00
• ITC(ms): 1.030
• TCPP(ms): 0.034
• RV(mb): 13.76
• WV(mb): 0.0041
• Dataset: ImageNet
• ACCURACY: Top1: 0.7935(FLOAT)/0.7924(INT8)
• LINKS: https://github.com/HorizonRobotics-Platform/ModelZoo/tree/master/EfficientnasNet

#Efficientnasnet_s

• INPUT SIZE: 1x3x280x280
• C(GOPs): 1.44
• FPS: 3335.50
• ITC(ms): 0.629
• TCPP(ms): 0.034
• RV(mb): 5.45
• WV(mb): 0.0041
• Dataset: ImageNet
• ACCURACY: Top1: 0.7441(FLOAT)/0.7522(INT8)
• LINKS: https://github.com/HorizonRobotics-Platform/ModelZoo/tree/master/EfficientnasNet

#ResNet18

• INPUT SIZE: 1x3x224x224
• C(GOPs): 3.63
• FPS: 2542.60
• ITC(ms): 0.683
• TCPP(ms): 0.034
• RV(mb): 11.87
• WV(mb): 0.0041
• Dataset: ImageNet
• ACCURACY: Top1: 0.7169(FLOAT)/0.7164(INT8)

#YOLOv2_Darknet19

• INPUT SIZE: 1x3x608x608
• C(GOPs): 62.94
• FPS: 225.31
• ITC(ms): 4.753
• TCPP(ms): 0.304
• RV(mb): 52.24
• WV(mb): 1.16
• Dataset: COCO
• ACCURACY: [IoU=0.50:0.95]= 0.2760(FLOAT)/0.2700(INT8)
• LINKS: https://pjreddie.com/darknet/yolo

#YOLOv3_Darknet53

• INPUT SIZE: 1x3x416x416
• C(GOPs): 65.86
• FPS: 209.30
• ITC(ms): 5.161
• TCPP(ms): 1.714
• RV(mb): 67.06
• WV(mb): 8.00
• Dataset: COCO
• ACCURACY: [IoU=0.50:0.95]= 0.3370(FLOAT)/0.3360(INT8)
• LINKS: https://github.com/ChenYingpeng/caffe-yolov3/

#YOLOv5x_v2.0

• INPUT SIZE: 1x3x672x672
• C(GOPs): 243.85
• FPS: 61.11
• ITC(ms): 16.823
• TCPP(ms): 5.916
• RV(mb): 136.68
• WV(mb): 46.38
• Dataset: COCO
• ACCURACY: [IoU=0.50:0.95]= 0.4810(FLOAT)/0.4670(INT8)
• LINKS: https://github.com/ultralytics/yolov5/releases/tag/v2.0

#SSD_MobileNetv1

• INPUT SIZE: 1x3x300x300
• C(GOPs): 2.30
• FPS: 2975.60
• ITC(ms): 0.709
• TCPP(ms): 0.197
• RV(mb): 6.28
• WV(mb): 0.25
• Dataset: VOC
• ACCURACY: mAP: 0.7345(FLOAT)/0.7269(INT8)
• LINKS: https://github.com/chuanqi305/MobileNet-SSD

#Centernet_resnet101

• INPUT SIZE: 1x3x512x512
• C(GOPs): 90.53
• FPS: 120.52
• ITC(ms): 8.663
• TCPP(ms): 0.993
• RV(mb): 75.16
• WV(mb): 25.43
• Dataset: COCO
• ACCURACY: [IoU=0.50:0.95]= 0.3420(FLOAT)/0.3240(INT8)
• LINKS: https://github.com/HorizonRobotics-Platform/ModelZoo/tree/master/Centernet

#YOLOv3_VargDarknet

• INPUT SIZE: 1x3x416x416
• C(GOPs): 42.82
• FPS: 307.32
• ITC(ms): 3.646
• TCPP(ms): 1.647
• RV(mb): 46.53
• WV(mb): 4.91
• Dataset: COCO
• ACCURACY: [IoU=0.50:0.95]= 0.3280(FLOAT)/0.3270(INT8)
• LINKS: https://github.com/HorizonRobotics-Platform/ModelZoo/tree/master/Yolov3_VargDarknet

#Deeplabv3plus_efficientnetb0

• INPUT SIZE: 1x3x1024x2048
• C(GOPs): 30.77
• FPS: 148.77
• ITC(ms): 7.090
• TCPP(ms): 0.317
• RV(mb): 18.86
• WV(mb): 10.49
• Dataset: Cityscapes
• ACCURACY: mIoU: 0.7630(FLOAT)/0.7570(INT8)
• LINKS: https://github.com/HorizonRobotics-Platform/ModelZoo/tree/master/DeeplabV3Plus

#Fastscnn_efficientnetb0

• INPUT SIZE: 1x3x1024x2048
• C(GOPs): 12.48
• FPS: 255.00
• ITC(ms): 4.282
• TCPP(ms): 0.317
• RV(mb): 12.03
• WV(mb): 9.57
• Dataset: Cityscapes
• ACCURACY: mIoU: 0.6997(FLOAT)/0.6910(INT8)
• LINKS: https://github.com/HorizonRobotics-Platform/ModelZoo/tree/master/FastSCNN

#Deeplabv3plus_efficientnetm1

• INPUT SIZE: 1x3x1024x2048
• C(GOPs): 77.04
• FPS: 90.33
• ITC(ms): 11.464
• TCPP(ms): 0.310
• RV(mb): 78.05
• WV(mb): 57.33
• Dataset: Cityscapes
• ACCURACY: mIoU: 0.7794(FLOAT)/0.7754(INT8)
• LINKS: https://github.com/HorizonRobotics-Platform/ModelZoo/tree/master/DeeplabV3Plus

#Deeplabv3plus_efficientnetm2

• INPUT SIZE: 1x3x1024x2048
• C(GOPs): 124.15
• FPS: 64.33
• ITC(ms): 15.913
• TCPP(ms): 0.316
• RV(mb): 138.57
• WV(mb): 80.40
• Dataset: Cityscapes
• ACCURACY: mIoU: 0.7882(FLOAT)/0.7853(INT8)
• LINKS: https://github.com/HorizonRobotics-Platform/ModelZoo/tree/master/DeeplabV3Plus

#Bev_gkt_mixvargenet_multitask

• INPUT SIZE: image: 6x3x512x960 points(0-8): 6x64x64x2
• C(GOPs): 207.16
• FPS: 63.33
• ITC(ms): 16.989
• TCPP(ms): 5.490
• RV(mb): 123.38
• WV(mb): 109.48
• Dataset: Nuscenes
• ACCURACY: NDS: 0.2810(FLOAT)/0.2787(INT8) MeanIOU: 0.4852(FLOAT)/0.4835(INT8) mAP: 0.1991(FLOAT)/0.1992(INT8)

#Bev_ipm_4d_efficientnetb0_multitask

• INPUT SIZE: image: 6x3x512x960 points: 6x128x128x2 prev_feat: 1x164x28x128 prev_point: 1x128x128x2
• C(GOPs): 53.58
• FPS: 108.49
• ITC(ms): 10.699
• TCPP(ms): 5.494
• RV(mb): 69.83
• WV(mb): 56.30
• Dataset: Nuscenes
• ACCURACY: NDS: 0.3721(FLOAT)/0.3735(INT8) MeanIOU: 0.5287(FLOAT)/0.5387(INT8) mAP: 0.2200(FLOAT)/0.2217(INT8)

#Bev_ipm_efficientnetb0_multitask

• INPUT SIZE: image: 6x3x512x960 points: 6x128x128x2
• C(GOPs): 52.97
• FPS: 112.13
• ITC(ms): 9.996
• TCPP(ms): 5.468
• RV(mb): 66.52
• WV(mb): 54.21
• Dataset: Nuscenes
• ACCURACY: NDS: 0.3056(FLOAT)/0.3029(INT8) MeanIOU: 0.5145(FLOAT)/0.5098(INT8) mAP: 0.2170(FLOAT)/0.2163(INT8)

#Bev_lss_efficientnetb0_multitask

• INPUT SIZE: image: 6x3x256x704 points(0&1): 10x128x128x2
• C(GOPs): 24.06
• FPS: 178.75
• ITC(ms): 6.670
• TCPP(ms): 5.438
• RV(mb): 26.98
• WV(mb): 20.48
• Dataset: Nuscenes
• ACCURACY: NDS: 0.3007(FLOAT)/0.3017(INT8) MeanIOU: 0.5180(FLOAT)/0.5147(INT8) mAP: 0.2062(FLOAT)/0.2050(INT8)

#Detr3d_efficientnetb3

• INPUT SIZE: coords(0-3): 6x4x256x2 image: 6x3x512x1408 masks: 1x4x256x24
• C(GOPs): 227.71
• FPS: 29.45
• ITC(ms): 34.624
• TCPP(ms): 1.117
• RV(mb): 376.75
• WV(mb): 228.92
• Dataset: Nuscenes
• ACCURACY: NDS: 0.3304(FLOAT)/0.3279(INT8) mAP: 0.2752(FLOAT)/0.2703(INT8)

#Petr_efficientnetb3

• INPUT SIZE: image: 6x3x512x1408 pos_embed: 1x96x44x256
• C(GOPs): 219.17
• FPS: 19.02
• ITC(ms): 53.187
• TCPP(ms): 1.130
• RV(mb): 260.48
• WV(mb): 149.67
• Dataset: Nuscenes
• ACCURACY: NDS: 0.3765(FLOAT)/0.3741(INT8) mAP: 0.3038(FLOAT)/0.2934(INT8)

#Bevformer_tiny_resnet50_detection

• INPUT SIZE: img: 6x3x480x800 prev_bev: 1x2500x256 prev_bev_ref: 1x50x50x2 queries_rebatch_grid: 6x20x32x2 restore_bev_grid: 1x100x50x2 reference_points_rebatch: 6x640x4x2 bev_pillar_counts: 1x2500x1
• C(GOPs): 387.29
• FPS: 28.84
• ITC(ms): 44.573
• TCPP(ms): 1.405
• RV(mb): 291.43
• WV(mb): 195.32
• Dataset: Nuscenes
• ACCURACY: NDS: 0.3713(FLOAT)/0.3679(INT8) mAP: 0.2673(FLOAT)/0.2614(INT8)

#Flashocc_henet_lss_occ3d_nuscenes

• INPUT SIZE: img: 6x3x512x960 points: 10x128x128x2 points_depth: 10x128x128x2
• C(GOPs): 126.75
• FPS: 89.10
• ITC(ms): 12.299
• TCPP(ms): 40.297
• RV(mb): 83.91
• WV(mb): 68.93
• Dataset: Nuscenes
• ACCURACY: mIoU: 0.3674(FLOAT)/0.3642(INT8)

#Horizon_swin_transformer

• INPUT SIZE: 1x3x224x224
• C(GOPs): 8.98
• FPS: 291.91
• ITC(ms): 3.730
• TCPP(ms): 0.034
• RV(mb): 47.17
• WV(mb): 6.31
• Dataset: ImageNet
• ACCURACY: Top1: 0.8024(FLOAT)/0.7947(INT8)

#Mixvargenet

• INPUT SIZE: 1x3x224x224
• C(GOPs): 2.07
• FPS: 4935.20
• ITC(ms): 0.488
• TCPP(ms): 0.034
• RV(mb): 2.51
• WV(mb): 0.0041
• Dataset: ImageNet
• ACCURACY: Top1: 0.7075(FLOAT)/0.7063(INT8)

#Vargnetv2

• INPUT SIZE: 1x3x224x224
• C(GOPs): 0.72
• FPS: 4254.40
• ITC(ms): 0.535
• TCPP(ms): 0.035
• RV(mb): 4.68
• WV(mb): 0.0041
• Dataset: ImageNet
• ACCURACY: Top1: 0.7342(FLOAT)/0.7317(INT8)

#Vit_small

• INPUT SIZE: 1x3x224x224
• C(GOPs): 9.20
• FPS: 533.30
• ITC(ms): 2.174
• TCPP(ms): 0.034
• RV(mb): 26.12
• WV(mb): 0.0041
• Dataset: ImageNet
• ACCURACY: Top1: 0.7950(FLOAT)/0.7927(INT8)

#Centerpoint_pointpillar

• INPUT SIZE: points: 300000x5 voxel_feature: 1x5x20x40000 coors: 40000x4
• C(GOPs): 127.73
• FPS: 89.97
• ITC(ms): 60.183
• TCPP(ms): 13.971
• RV(mb): 49.11
• WV(mb): 24.78
• Dataset: Nuscenes
• ACCURACY: NDS: 0.5832(FLOAT)/0.5819(INT8) mAP: 0.4804(FLOAT)/0.4780(INT8)

#Detr_efficientnetb3

• INPUT SIZE: 1x3x800x1333
• C(GOPs): 67.39
• FPS: 53.59
• ITC(ms): 19.033
• TCPP(ms): 0.345
• RV(mb): 278.29
• WV(mb): 153.76
• Dataset: MS COCO
• ACCURACY: [IoU=0.50:0.95]= 0.3721(FLOAT)/0.3605(INT8)

#Detr_resnet50

• INPUT SIZE: 1x3x800x1333
• C(GOPs): 203.07
• FPS: 39.75
• ITC(ms): 25.549
• TCPP(ms): 0.345
• RV(mb): 409.13
• WV(mb): 270.43
• Dataset: MS COCO
• ACCURACY: [IoU=0.50:0.95]= 0.3569(FLOAT)/0.3142(INT8)

#FCOS3D_efficientnetb0

• INPUT SIZE: 1x3x512x896
• C(GOPs): 19.94
• FPS: 426.75
• ITC(ms): 3.405
• TCPP(ms): 2.740
• RV(mb): 11.50
• WV(mb): 4.14
• Dataset: nuscenes
• ACCURACY: NDS: 0.3061(FLOAT)/0.3029(INT8) mAP: 0.2133(FLOAT)/0.2079(INT8)

#Fcos_efficientnetb0

• INPUT SIZE: 1x3x512x512
• C(GOPs): 5.02
• FPS: 1043.50
• ITC(ms): 1.931
• TCPP(ms): 0.136
• RV(mb): 6.51
• WV(mb): 2.78
• Dataset: MS COCO
• ACCURACY: [IoU=0.50:0.95]= 0.3626(FLOAT)/0.3565(INT8)

#Ganet_mixvargenet

• INPUT SIZE: 1x3x320x800
• C(GOPs): 10.74
• FPS: 1478.00
• ITC(ms): 1.013
• TCPP(ms): 0.211
• RV(mb): 2.18
• WV(mb): 0.53
• Dataset: CuLane
• ACCURACY: F1Score: 0.7949(FLOAT)/0.7881(INT8)

#Keypoint_efficientnetb0

• INPUT SIZE: 1x3x128x128
• C(GOPs): 0.45
• FPS: 4697.00
• ITC(ms): 0.491
• TCPP(ms): 0.071
• RV(mb): 4.62
• WV(mb): 0.01
• Dataset: Carfusion
• ACCURACY: PCK(alpha=0.1): 0.9433(FLOAT)/0.9432(INT8)

#Pointpillars_kitti_car

• INPUT SIZE: 150000x4
• C(GOPs): 66.82
• FPS: 23.73
• ITC(ms): 229.510
• TCPP(ms): 0.537
• RV(mb): 49.90
• WV(mb): 30.04
• Dataset: Kitti3d
• ACCURACY: APDet= 0.7733(FLOAT)/0.7676(INT8)

#Deformable_detr_resnet50

• INPUT SIZE: 1x3x800x1333
• C(GOPs): 408.94
• FPS: 4.76
• ITC(ms): 210.470
• TCPP(ms): 15.570
• RV(mb): 3830.91
• WV(mb): 2918.69
• Dataset: MS COCO
• ACCURACY: [IoU=0.50:0.95]= 0.4414(FLOAT)/0.4204(INT8)

#Stereonetplus_mixvargenet

• INPUT SIZE: 2x3x544x960
• C(GOPs): 48.57
• FPS: 208.93
• ITC(ms): 5.219
• TCPP(ms): 1.960
• RV(mb): 38.97
• WV(mb): 34.60
• Dataset: SceneFlow
• ACCURACY: EPE: 1.1270(FLOAT)/1.1342(INT8)

#Centerpoint_mixvargnet_multitask

• INPUT SIZE: points: 300000x5 voxel_feature: 1x5x20x40000 coors: 40000x4
• C(GOPs): 51.45
• FPS: 90.59
• ITC(ms): 57.678
• TCPP(ms): 12.097
• RV(mb): 32.41
• WV(mb): 16.79
• Dataset: Nuscenes
• ACCURACY: NDS: 0.5809(FLOAT)/0.5754(INT8) MeanIOU: 0.9128(FLOAT)/0.9121(INT8) mAP: 0.4726(FLOAT)/0.4629(INT8)

#Unet_mobilenetv1

• INPUT SIZE: 1x3x1024x2048
• C(GOPs): 7.36
• FPS: 780.85
• ITC(ms): 1.728
• TCPP(ms): 0.148
• RV(mb): 14.85
• WV(mb): 9.57
• Dataset: Cityscapes
• ACCURACY: mIoU: 0.6802(FLOAT)/0.6757(INT8)

#Motr_efficientnetb3

• INPUT SIZE: image: 1x800x1422x3 track_query: 1x2x128x156 ref_points: 1x2x128x4 mask_query: 1x1x256x1
• C(GOPs): 64.43
• FPS: 71.27
• ITC(ms): 14.111
• TCPP(ms): 5.135
• RV(mb): 120.39
• WV(mb): 40.80
• Dataset: Mot17
• ACCURACY: MOTA: 0.5805(FLOAT)/0.5728(INT8)

#Densetnt_vectornet

• INPUT SIZE: goals_2d: 30x1x2048x2 goals_2d_mask: 30x1x2048x1 instance_mask: 30x1x96x1 lane_feat: 30x9x64x11 traj_feat: 30x19x32x9
• C(GOPs): 12.50
• FPS: 155.06
• ITC(ms): 11.607
• TCPP(ms): 2.302
• RV(mb): 61.68
• WV(mb): 41.81
• Dataset: Argoverse 1
• ACCURACY: minFDA: 1.2975(FLOAT)/1.3058(INT8)

#Maptroe_henet_tinym_bevformer

• INPUT SIZE: img: 6x3x480x800 osm_mask: 1x1x50x100 queries_rebatch_grid: 6x20x100x2 restore_bev_grid: 1x100x100x2 reference_points_rebatch: 6x2000x4x2 bev_pillar_counts: 1x5000x1
• C(GOPs): 134.57
• FPS: 67.78
• ITC(ms): 15.350
• TCPP(ms): 0.254
• RV(mb): 155.95
• WV(mb): 70.67
• Dataset: Nuscenes
• ACCURACY: mAP: 0.6633(FLOAT)/0.6509(INT8)

#Qcnet_oe

• INPUT SIZE: valid_mask: 1x30x10 valid_mask_a2a: 1x10x30x30 agent_type: 1x30x1 x_a_cur: 1x1x30x1,1x1x30x1,1x1x30x1,1x1x30x1 r_pl2a_cur: 1x1x30x80,1x1x30x80,1x1x30x80 r_t_cur: 1x1x30x6,1x1x30x6,1x1x30x6,1x1x30x6 r_a2a_cur: 1x1x30x30,1x1x30x30,1x1x30x30 x_a_mid_emb: 1x30x2x128 x_a: 1x30x6x128 pl_type,is_intersection: 1x80 r_pl2pl: 1x1x80x80,1x1x80x80,1x1x80x80 r_pt2pl: 1x1x80x50,1x1x80x50,1x1x80x50 mask_pl2pl: 1x80x80 magnitude,pt_type,side,mask: 1x80x50 mask_a2m: 1x30x30 mask_dst: 1x30x1 type_pl2pl: 1x80x80
• C(GOPs): 7.85
• FPS: 201.85
• ITC(ms): 12.935
• TCPP(ms): 0.822
• RV(mb): 48.28
• WV(mb): 25.99
• Dataset: Argoverse 2
• ACCURACY: hitrate: 0.8026(FLOAT)/0.7906(INT8)