Upload folder using huggingface_hub

GenAD/with_hlc/20260205_035946.log.json

@@ -0,0 +1,9 @@
+{"env_info": "MMCV: 0.0.1", "config": "point_cloud_range = [-15.0, -30.0, -2.0, 15.0, 30.0, 2.0]\nclass_names = [\n    'car', 'van', 'truck', 'bicycle', 'traffic_sign', 'traffic_cone',\n    'traffic_light', 'pedestrian', 'others'\n]\ndataset_type = 'B2D_VAD_Dataset'\ndata_root = 'data/bench2drive'\ninput_modality = dict(\n    use_lidar=False,\n    use_camera=True,\n    use_radar=False,\n    use_map=False,\n    use_external=True)\nfile_client_args = dict(backend='disk')\ntrain_pipeline = [\n    dict(type='LoadMultiViewImageFromFiles', to_float32=True),\n    dict(type='PhotoMetricDistortionMultiViewImage'),\n    dict(\n        type='LoadAnnotations3D',\n        with_bbox_3d=True,\n        with_label_3d=True,\n        with_attr_label=True),\n    dict(\n        type='VADObjectRangeFilter',\n        point_cloud_range=[-15.0, -30.0, -2.0, 15.0, 30.0, 2.0]),\n    dict(\n        type='VADObjectNameFilter',\n        classes=[\n            'car', 'van', 'truck', 'bicycle', 'traffic_sign', 'traffic_cone',\n            'traffic_light', 'pedestrian', 'others'\n        ]),\n    dict(\n        type='NormalizeMultiviewImage',\n        mean=[123.675, 116.28, 103.53],\n        std=[58.395, 57.12, 57.375],\n        to_rgb=True),\n    dict(type='RandomScaleImageMultiViewImage', scales=[0.8]),\n    dict(type='PadMultiViewImage', size_divisor=32),\n    dict(\n        type='VADFormatBundle3D',\n        class_names=[\n            'car', 'van', 'truck', 'bicycle', 'traffic_sign', 'traffic_cone',\n            'traffic_light', 'pedestrian', 'others'\n        ],\n        with_ego=True),\n    dict(\n        type='CustomCollect3D',\n        keys=[\n            'gt_bboxes_3d', 'gt_labels_3d', 'img', 'ego_his_trajs',\n            'gt_attr_labels', 'ego_fut_trajs', 'ego_fut_masks', 'ego_fut_cmd',\n            'ego_lcf_feat'\n        ])\n]\ntest_pipeline = [\n    dict(type='LoadMultiViewImageFromFiles', to_float32=True),\n    dict(\n        type='LoadAnnotations3D',\n        with_bbox_3d=True,\n        with_label_3d=True,\n        with_attr_label=True),\n    dict(\n        type='VADObjectRangeFilter',\n        point_cloud_range=[-15.0, -30.0, -2.0, 15.0, 30.0, 2.0]),\n    dict(\n        type='VADObjectNameFilter',\n        classes=[\n            'car', 'van', 'truck', 'bicycle', 'traffic_sign', 'traffic_cone',\n            'traffic_light', 'pedestrian', 'others'\n        ]),\n    dict(\n        type='NormalizeMultiviewImage',\n        mean=[123.675, 116.28, 103.53],\n        std=[58.395, 57.12, 57.375],\n        to_rgb=True),\n    dict(\n        type='MultiScaleFlipAug3D',\n        img_scale=(1600, 900),\n        pts_scale_ratio=1,\n        flip=False,\n        transforms=[\n            dict(type='RandomScaleImageMultiViewImage', scales=[0.8]),\n            dict(type='PadMultiViewImage', size_divisor=32),\n            dict(\n                type='VADFormatBundle3D',\n                class_names=[\n                    'car', 'van', 'truck', 'bicycle', 'traffic_sign',\n                    'traffic_cone', 'traffic_light', 'pedestrian', 'others'\n                ],\n                with_label=False,\n                with_ego=True),\n            dict(\n                type='CustomCollect3D',\n                keys=[\n                    'gt_bboxes_3d', 'gt_labels_3d', 'img', 'fut_valid_flag',\n                    'ego_his_trajs', 'ego_fut_trajs', 'ego_fut_masks',\n                    'ego_fut_cmd', 'ego_lcf_feat', 'gt_attr_labels'\n                ])\n        ])\n]\neval_pipeline = [\n    dict(\n        type='LoadPointsFromFile',\n        coord_type='LIDAR',\n        load_dim=5,\n        use_dim=5,\n        file_client_args=dict(backend='disk')),\n    dict(\n        type='LoadPointsFromMultiSweeps',\n        sweeps_num=10,\n        file_client_args=dict(backend='disk')),\n    dict(\n        type='DefaultFormatBundle3D',\n        class_names=[\n            'car', 'truck', 'trailer', 'bus', 'construction_vehicle',\n            'bicycle', 'motorcycle', 'pedestrian', 'traffic_cone', 'barrier'\n        ],\n        with_label=False),\n    dict(type='Collect3D', keys=['points'])\n]\ndata = dict(\n    samples_per_gpu=1,\n    workers_per_gpu=4,\n    train=dict(\n        type='B2D_VAD_Dataset',\n        data_root='data/bench2drive',\n        ann_file='data/infos/b2d_infos_train.pkl',\n        pipeline=[\n            dict(type='LoadMultiViewImageFromFiles', to_float32=True),\n            dict(type='PhotoMetricDistortionMultiViewImage'),\n            dict(\n                type='LoadAnnotations3D',\n                with_bbox_3d=True,\n                with_label_3d=True,\n                with_attr_label=True),\n            dict(\n                type='VADObjectRangeFilter',\n                point_cloud_range=[-15.0, -30.0, -2.0, 15.0, 30.0, 2.0]),\n            dict(\n                type='VADObjectNameFilter',\n                classes=[\n                    'car', 'van', 'truck', 'bicycle', 'traffic_sign',\n                    'traffic_cone', 'traffic_light', 'pedestrian', 'others'\n                ]),\n            dict(\n                type='NormalizeMultiviewImage',\n                mean=[123.675, 116.28, 103.53],\n                std=[58.395, 57.12, 57.375],\n                to_rgb=True),\n            dict(type='RandomScaleImageMultiViewImage', scales=[0.8]),\n            dict(type='PadMultiViewImage', size_divisor=32),\n            dict(\n                type='VADFormatBundle3D',\n                class_names=[\n                    'car', 'van', 'truck', 'bicycle', 'traffic_sign',\n                    'traffic_cone', 'traffic_light', 'pedestrian', 'others'\n                ],\n                with_ego=True),\n            dict(\n                type='CustomCollect3D',\n                keys=[\n                    'gt_bboxes_3d', 'gt_labels_3d', 'img', 'ego_his_trajs',\n                    'gt_attr_labels', 'ego_fut_trajs', 'ego_fut_masks',\n                    'ego_fut_cmd', 'ego_lcf_feat'\n                ])\n        ],\n        classes=[\n            'car', 'van', 'truck', 'bicycle', 'traffic_sign', 'traffic_cone',\n            'traffic_light', 'pedestrian', 'others'\n        ],\n        modality=dict(\n            use_lidar=False,\n            use_camera=True,\n            use_radar=False,\n            use_map=False,\n            use_external=True),\n        test_mode=False,\n        box_type_3d='LiDAR',\n        name_mapping=dict({\n            'vehicle.bh.crossbike':\n            'bicycle',\n            'vehicle.diamondback.century':\n            'bicycle',\n            'vehicle.gazelle.omafiets':\n            'bicycle',\n            'vehicle.chevrolet.impala':\n            'car',\n            'vehicle.dodge.charger_2020':\n            'car',\n            'vehicle.dodge.charger_police':\n            'car',\n            'vehicle.dodge.charger_police_2020':\n            'car',\n            'vehicle.lincoln.mkz_2017':\n            'car',\n            'vehicle.lincoln.mkz_2020':\n            'car',\n            'vehicle.mini.cooper_s_2021':\n            'car',\n            'vehicle.mercedes.coupe_2020':\n            'car',\n            'vehicle.ford.mustang':\n            'car',\n            'vehicle.nissan.patrol_2021':\n            'car',\n            'vehicle.audi.tt':\n            'car',\n            'vehicle.audi.etron':\n            'car',\n            'vehicle.ford.crown':\n            'car',\n            'vehicle.tesla.model3':\n            'car',\n            '/Game/Carla/Static/Car/4Wheeled/ParkedVehicles/FordCrown/SM_FordCrown_parked.SM_FordCrown_parked':\n            'car',\n            '/Game/Carla/Static/Car/4Wheeled/ParkedVehicles/Charger/SM_ChargerParked.SM_ChargerParked':\n            'car',\n            '/Game/Carla/Static/Car/4Wheeled/ParkedVehicles/Lincoln/SM_LincolnParked.SM_LincolnParked':\n            'car',\n            '/Game/Carla/Static/Car/4Wheeled/ParkedVehicles/MercedesCCC/SM_MercedesCCC_Parked.SM_MercedesCCC_Parked':\n            'car',\n            '/Game/Carla/Static/Car/4Wheeled/ParkedVehicles/Mini2021/SM_Mini2021_parked.SM_Mini2021_parked':\n            'car',\n            '/Game/Carla/Static/Car/4Wheeled/ParkedVehicles/NissanPatrol2021/SM_NissanPatrol2021_parked.SM_NissanPatrol2021_parked':\n            'car',\n            '/Game/Carla/Static/Car/4Wheeled/ParkedVehicles/TeslaM3/SM_TeslaM3_parked.SM_TeslaM3_parked':\n            'car',\n            '/Game/Carla/Static/Car/4Wheeled/ParkedVehicles/VolkswagenT2/SM_VolkswagenT2_2021_Parked.SM_VolkswagenT2_2021_Parked':\n            'van',\n            'vehicle.ford.ambulance':\n            'van',\n            'vehicle.carlamotors.firetruck':\n            'truck',\n            'traffic.speed_limit.30':\n            'traffic_sign',\n            'traffic.speed_limit.40':\n            'traffic_sign',\n            'traffic.speed_limit.50':\n            'traffic_sign',\n            'traffic.speed_limit.60':\n            'traffic_sign',\n            'traffic.speed_limit.90':\n            'traffic_sign',\n            'traffic.speed_limit.120':\n            'traffic_sign',\n            'traffic.stop':\n            'traffic_sign',\n            'traffic.yield':\n            'traffic_sign',\n            'traffic.traffic_light':\n            'traffic_light',\n            'static.prop.warningconstruction':\n            'traffic_cone',\n            'static.prop.warningaccident':\n            'traffic_cone',\n            'static.prop.trafficwarning':\n            'traffic_cone',\n            'static.prop.constructioncone':\n            'traffic_cone',\n            'walker.pedestrian.0001':\n            'pedestrian',\n            'walker.pedestrian.0004':\n            'pedestrian',\n            'walker.pedestrian.0005':\n            'pedestrian',\n            'walker.pedestrian.0007':\n            'pedestrian',\n            'walker.pedestrian.0013':\n            'pedestrian',\n            'walker.pedestrian.0014':\n            'pedestrian',\n            'walker.pedestrian.0017':\n            'pedestrian',\n            'walker.pedestrian.0018':\n            'pedestrian',\n            'walker.pedestrian.0019':\n            'pedestrian',\n            'walker.pedestrian.0020':\n            'pedestrian',\n            'walker.pedestrian.0022':\n            'pedestrian',\n            'walker.pedestrian.0025':\n            'pedestrian',\n            'walker.pedestrian.0035':\n            'pedestrian',\n            'walker.pedestrian.0041':\n            'pedestrian',\n            'walker.pedestrian.0046':\n            'pedestrian',\n            'walker.pedestrian.0047':\n            'pedestrian',\n            'static.prop.dirtdebris01':\n            'others',\n            'static.prop.dirtdebris02':\n            'others'\n        }),\n        map_root='data/bench2drive/maps',\n        map_file='data/infos/b2d_map_infos.pkl',\n        bev_size=(100, 100),\n        queue_length=3,\n        past_frames=2,\n        future_frames=6,\n        point_cloud_range=[-15.0, -30.0, -2.0, 15.0, 30.0, 2.0],\n        polyline_points_num=20),\n    val=dict(\n        type='B2D_VAD_Dataset',\n        ann_file='data/infos/b2d_infos_val.pkl',\n        pipeline=[\n            dict(type='LoadMultiViewImageFromFiles', to_float32=True),\n            dict(\n                type='LoadAnnotations3D',\n                with_bbox_3d=True,\n                with_label_3d=True,\n                with_attr_label=True),\n            dict(\n                type='VADObjectRangeFilter',\n                point_cloud_range=[-15.0, -30.0, -2.0, 15.0, 30.0, 2.0]),\n            dict(\n                type='VADObjectNameFilter',\n                classes=[\n                    'car', 'van', 'truck', 'bicycle', 'traffic_sign',\n                    'traffic_cone', 'traffic_light', 'pedestrian', 'others'\n                ]),\n            dict(\n                type='NormalizeMultiviewImage',\n                mean=[123.675, 116.28, 103.53],\n                std=[58.395, 57.12, 57.375],\n                to_rgb=True),\n            dict(\n                type='MultiScaleFlipAug3D',\n                img_scale=(1600, 900),\n                pts_scale_ratio=1,\n                flip=False,\n                transforms=[\n                    dict(type='RandomScaleImageMultiViewImage', scales=[0.8]),\n                    dict(type='PadMultiViewImage', size_divisor=32),\n                    dict(\n                        type='VADFormatBundle3D',\n                        class_names=[\n                            'car', 'van', 'truck', 'bicycle', 'traffic_sign',\n                            'traffic_cone', 'traffic_light', 'pedestrian',\n                            'others'\n                        ],\n                        with_label=False,\n                        with_ego=True),\n                    dict(\n                        type='CustomCollect3D',\n                        keys=[\n                            'gt_bboxes_3d', 'gt_labels_3d', 'img',\n                            'fut_valid_flag', 'ego_his_trajs', 'ego_fut_trajs',\n                            'ego_fut_masks', 'ego_fut_cmd', 'ego_lcf_feat',\n                            'gt_attr_labels'\n                        ])\n                ])\n        ],\n        classes=[\n            'car', 'van', 'truck', 'bicycle', 'traffic_sign', 'traffic_cone',\n            'traffic_light', 'pedestrian', 'others'\n        ],\n        modality=dict(\n            use_lidar=False,\n            use_camera=True,\n            use_radar=False,\n            use_map=False,\n            use_external=True),\n        test_mode=True,\n        box_type_3d='LiDAR',\n        data_root='data/bench2drive',\n        name_mapping=dict({\n            'vehicle.bh.crossbike':\n            'bicycle',\n            'vehicle.diamondback.century':\n            'bicycle',\n            'vehicle.gazelle.omafiets':\n            'bicycle',\n            'vehicle.chevrolet.impala':\n            'car',\n            'vehicle.dodge.charger_2020':\n            'car',\n            'vehicle.dodge.charger_police':\n            'car',\n            'vehicle.dodge.charger_police_2020':\n            'car',\n            'vehicle.lincoln.mkz_2017':\n            'car',\n            'vehicle.lincoln.mkz_2020':\n            'car',\n            'vehicle.mini.cooper_s_2021':\n            'car',\n            'vehicle.mercedes.coupe_2020':\n            'car',\n            'vehicle.ford.mustang':\n            'car',\n            'vehicle.nissan.patrol_2021':\n            'car',\n            'vehicle.audi.tt':\n            'car',\n            'vehicle.audi.etron':\n            'car',\n            'vehicle.ford.crown':\n            'car',\n            'vehicle.tesla.model3':\n            'car',\n            '/Game/Carla/Static/Car/4Wheeled/ParkedVehicles/FordCrown/SM_FordCrown_parked.SM_FordCrown_parked':\n            'car',\n            '/Game/Carla/Static/Car/4Wheeled/ParkedVehicles/Charger/SM_ChargerParked.SM_ChargerParked':\n            'car',\n            '/Game/Carla/Static/Car/4Wheeled/ParkedVehicles/Lincoln/SM_LincolnParked.SM_LincolnParked':\n            'car',\n            '/Game/Carla/Static/Car/4Wheeled/ParkedVehicles/MercedesCCC/SM_MercedesCCC_Parked.SM_MercedesCCC_Parked':\n            'car',\n            '/Game/Carla/Static/Car/4Wheeled/ParkedVehicles/Mini2021/SM_Mini2021_parked.SM_Mini2021_parked':\n            'car',\n            '/Game/Carla/Static/Car/4Wheeled/ParkedVehicles/NissanPatrol2021/SM_NissanPatrol2021_parked.SM_NissanPatrol2021_parked':\n            'car',\n            '/Game/Carla/Static/Car/4Wheeled/ParkedVehicles/TeslaM3/SM_TeslaM3_parked.SM_TeslaM3_parked':\n            'car',\n            '/Game/Carla/Static/Car/4Wheeled/ParkedVehicles/VolkswagenT2/SM_VolkswagenT2_2021_Parked.SM_VolkswagenT2_2021_Parked':\n            'van',\n            'vehicle.ford.ambulance':\n            'van',\n            'vehicle.carlamotors.firetruck':\n            'truck',\n            'traffic.speed_limit.30':\n            'traffic_sign',\n            'traffic.speed_limit.40':\n            'traffic_sign',\n            'traffic.speed_limit.50':\n            'traffic_sign',\n            'traffic.speed_limit.60':\n            'traffic_sign',\n            'traffic.speed_limit.90':\n            'traffic_sign',\n            'traffic.speed_limit.120':\n            'traffic_sign',\n            'traffic.stop':\n            'traffic_sign',\n            'traffic.yield':\n            'traffic_sign',\n            'traffic.traffic_light':\n            'traffic_light',\n            'static.prop.warningconstruction':\n            'traffic_cone',\n            'static.prop.warningaccident':\n            'traffic_cone',\n            'static.prop.trafficwarning':\n            'traffic_cone',\n            'static.prop.constructioncone':\n            'traffic_cone',\n            'walker.pedestrian.0001':\n            'pedestrian',\n            'walker.pedestrian.0004':\n            'pedestrian',\n            'walker.pedestrian.0005':\n            'pedestrian',\n            'walker.pedestrian.0007':\n            'pedestrian',\n            'walker.pedestrian.0013':\n            'pedestrian',\n            'walker.pedestrian.0014':\n            'pedestrian',\n            'walker.pedestrian.0017':\n            'pedestrian',\n            'walker.pedestrian.0018':\n            'pedestrian',\n            'walker.pedestrian.0019':\n            'pedestrian',\n            'walker.pedestrian.0020':\n            'pedestrian',\n            'walker.pedestrian.0022':\n            'pedestrian',\n            'walker.pedestrian.0025':\n            'pedestrian',\n            'walker.pedestrian.0035':\n            'pedestrian',\n            'walker.pedestrian.0041':\n            'pedestrian',\n            'walker.pedestrian.0046':\n            'pedestrian',\n            'walker.pedestrian.0047':\n            'pedestrian',\n            'static.prop.dirtdebris01':\n            'others',\n            'static.prop.dirtdebris02':\n            'others'\n        }),\n        map_root='data/bench2drive/maps',\n        map_file='data/infos/b2d_map_infos.pkl',\n        bev_size=(100, 100),\n        queue_length=3,\n        past_frames=2,\n        future_frames=6,\n        point_cloud_range=[-15.0, -30.0, -2.0, 15.0, 30.0, 2.0],\n        polyline_points_num=20,\n        eval_cfg=dict(\n            dist_ths=[0.5, 1.0, 2.0, 4.0],\n            dist_th_tp=2.0,\n            min_recall=0.1,\n            min_precision=0.1,\n            mean_ap_weight=5,\n            class_names=[\n                'car', 'van', 'truck', 'bicycle', 'traffic_sign',\n                'traffic_cone', 'traffic_light', 'pedestrian'\n            ],\n            tp_metrics=['trans_err', 'scale_err', 'orient_err', 'vel_err'],\n            err_name_maping=dict(\n                trans_err='mATE',\n                scale_err='mASE',\n                orient_err='mAOE',\n                vel_err='mAVE',\n                attr_err='mAAE'),\n            class_range=dict(\n                car=(50, 50),\n                van=(50, 50),\n                truck=(50, 50),\n                bicycle=(40, 40),\n                traffic_sign=(30, 30),\n                traffic_cone=(30, 30),\n                traffic_light=(30, 30),\n                pedestrian=(40, 40)))),\n    test=dict(\n        type='B2D_VAD_Dataset',\n        data_root='data/bench2drive',\n        ann_file='data/infos/b2d_infos_val.pkl',\n        pipeline=[\n            dict(type='LoadMultiViewImageFromFiles', to_float32=True),\n            dict(\n                type='LoadAnnotations3D',\n                with_bbox_3d=True,\n                with_label_3d=True,\n                with_attr_label=True),\n            dict(\n                type='VADObjectRangeFilter',\n                point_cloud_range=[-15.0, -30.0, -2.0, 15.0, 30.0, 2.0]),\n            dict(\n                type='VADObjectNameFilter',\n                classes=[\n                    'car', 'van', 'truck', 'bicycle', 'traffic_sign',\n                    'traffic_cone', 'traffic_light', 'pedestrian', 'others'\n                ]),\n            dict(\n                type='NormalizeMultiviewImage',\n                mean=[123.675, 116.28, 103.53],\n                std=[58.395, 57.12, 57.375],\n                to_rgb=True),\n            dict(\n                type='MultiScaleFlipAug3D',\n                img_scale=(1600, 900),\n                pts_scale_ratio=1,\n                flip=False,\n                transforms=[\n                    dict(type='RandomScaleImageMultiViewImage', scales=[0.8]),\n                    dict(type='PadMultiViewImage', size_divisor=32),\n                    dict(\n                        type='VADFormatBundle3D',\n                        class_names=[\n                            'car', 'van', 'truck', 'bicycle', 'traffic_sign',\n                            'traffic_cone', 'traffic_light', 'pedestrian',\n                            'others'\n                        ],\n                        with_label=False,\n                        with_ego=True),\n                    dict(\n                        type='CustomCollect3D',\n                        keys=[\n                            'gt_bboxes_3d', 'gt_labels_3d', 'img',\n                            'fut_valid_flag', 'ego_his_trajs', 'ego_fut_trajs',\n                            'ego_fut_masks', 'ego_fut_cmd', 'ego_lcf_feat',\n                            'gt_attr_labels'\n                        ])\n                ])\n        ],\n        classes=[\n            'car', 'van', 'truck', 'bicycle', 'traffic_sign', 'traffic_cone',\n            'traffic_light', 'pedestrian', 'others'\n        ],\n        modality=dict(\n            use_lidar=False,\n            use_camera=True,\n            use_radar=False,\n            use_map=False,\n            use_external=True),\n        test_mode=True,\n        box_type_3d='LiDAR',\n        name_mapping=dict({\n            'vehicle.bh.crossbike':\n            'bicycle',\n            'vehicle.diamondback.century':\n            'bicycle',\n            'vehicle.gazelle.omafiets':\n            'bicycle',\n            'vehicle.chevrolet.impala':\n            'car',\n            'vehicle.dodge.charger_2020':\n            'car',\n            'vehicle.dodge.charger_police':\n            'car',\n            'vehicle.dodge.charger_police_2020':\n            'car',\n            'vehicle.lincoln.mkz_2017':\n            'car',\n            'vehicle.lincoln.mkz_2020':\n            'car',\n            'vehicle.mini.cooper_s_2021':\n            'car',\n            'vehicle.mercedes.coupe_2020':\n            'car',\n            'vehicle.ford.mustang':\n            'car',\n            'vehicle.nissan.patrol_2021':\n            'car',\n            'vehicle.audi.tt':\n            'car',\n            'vehicle.audi.etron':\n            'car',\n            'vehicle.ford.crown':\n            'car',\n            'vehicle.tesla.model3':\n            'car',\n            '/Game/Carla/Static/Car/4Wheeled/ParkedVehicles/FordCrown/SM_FordCrown_parked.SM_FordCrown_parked':\n            'car',\n            '/Game/Carla/Static/Car/4Wheeled/ParkedVehicles/Charger/SM_ChargerParked.SM_ChargerParked':\n            'car',\n            '/Game/Carla/Static/Car/4Wheeled/ParkedVehicles/Lincoln/SM_LincolnParked.SM_LincolnParked':\n            'car',\n            '/Game/Carla/Static/Car/4Wheeled/ParkedVehicles/MercedesCCC/SM_MercedesCCC_Parked.SM_MercedesCCC_Parked':\n            'car',\n            '/Game/Carla/Static/Car/4Wheeled/ParkedVehicles/Mini2021/SM_Mini2021_parked.SM_Mini2021_parked':\n            'car',\n            '/Game/Carla/Static/Car/4Wheeled/ParkedVehicles/NissanPatrol2021/SM_NissanPatrol2021_parked.SM_NissanPatrol2021_parked':\n            'car',\n            '/Game/Carla/Static/Car/4Wheeled/ParkedVehicles/TeslaM3/SM_TeslaM3_parked.SM_TeslaM3_parked':\n            'car',\n            '/Game/Carla/Static/Car/4Wheeled/ParkedVehicles/VolkswagenT2/SM_VolkswagenT2_2021_Parked.SM_VolkswagenT2_2021_Parked':\n            'van',\n            'vehicle.ford.ambulance':\n            'van',\n            'vehicle.carlamotors.firetruck':\n            'truck',\n            'traffic.speed_limit.30':\n            'traffic_sign',\n            'traffic.speed_limit.40':\n            'traffic_sign',\n            'traffic.speed_limit.50':\n            'traffic_sign',\n            'traffic.speed_limit.60':\n            'traffic_sign',\n            'traffic.speed_limit.90':\n            'traffic_sign',\n            'traffic.speed_limit.120':\n            'traffic_sign',\n            'traffic.stop':\n            'traffic_sign',\n            'traffic.yield':\n            'traffic_sign',\n            'traffic.traffic_light':\n            'traffic_light',\n            'static.prop.warningconstruction':\n            'traffic_cone',\n            'static.prop.warningaccident':\n            'traffic_cone',\n            'static.prop.trafficwarning':\n            'traffic_cone',\n            'static.prop.constructioncone':\n            'traffic_cone',\n            'walker.pedestrian.0001':\n            'pedestrian',\n            'walker.pedestrian.0004':\n            'pedestrian',\n            'walker.pedestrian.0005':\n            'pedestrian',\n            'walker.pedestrian.0007':\n            'pedestrian',\n            'walker.pedestrian.0013':\n            'pedestrian',\n            'walker.pedestrian.0014':\n            'pedestrian',\n            'walker.pedestrian.0017':\n            'pedestrian',\n            'walker.pedestrian.0018':\n            'pedestrian',\n            'walker.pedestrian.0019':\n            'pedestrian',\n            'walker.pedestrian.0020':\n            'pedestrian',\n            'walker.pedestrian.0022':\n            'pedestrian',\n            'walker.pedestrian.0025':\n            'pedestrian',\n            'walker.pedestrian.0035':\n            'pedestrian',\n            'walker.pedestrian.0041':\n            'pedestrian',\n            'walker.pedestrian.0046':\n            'pedestrian',\n            'walker.pedestrian.0047':\n            'pedestrian',\n            'static.prop.dirtdebris01':\n            'others',\n            'static.prop.dirtdebris02':\n            'others'\n        }),\n        map_root='data/bench2drive/maps',\n        map_file='data/infos/b2d_map_infos.pkl',\n        bev_size=(100, 100),\n        queue_length=3,\n        past_frames=2,\n        future_frames=6,\n        point_cloud_range=[-15.0, -30.0, -2.0, 15.0, 30.0, 2.0],\n        polyline_points_num=20,\n        eval_cfg=dict(\n            dist_ths=[0.5, 1.0, 2.0, 4.0],\n            dist_th_tp=2.0,\n            min_recall=0.1,\n            min_precision=0.1,\n            mean_ap_weight=5,\n            class_names=[\n                'car', 'van', 'truck', 'bicycle', 'traffic_sign',\n                'traffic_cone', 'traffic_light', 'pedestrian'\n            ],\n            tp_metrics=['trans_err', 'scale_err', 'orient_err', 'vel_err'],\n            err_name_maping=dict(\n                trans_err='mATE',\n                scale_err='mASE',\n                orient_err='mAOE',\n                vel_err='mAVE',\n                attr_err='mAAE'),\n            class_range=dict(\n                car=(50, 50),\n                van=(50, 50),\n                truck=(50, 50),\n                bicycle=(40, 40),\n                traffic_sign=(30, 30),\n                traffic_cone=(30, 30),\n                traffic_light=(30, 30),\n                pedestrian=(40, 40)))),\n    shuffler_sampler=dict(type='DistributedGroupSampler'),\n    nonshuffler_sampler=dict(type='DistributedSampler'))\nevaluation = dict(\n    interval=6,\n    pipeline=[\n        dict(type='LoadMultiViewImageFromFiles', to_float32=True),\n        dict(\n            type='LoadAnnotations3D',\n            with_bbox_3d=True,\n            with_label_3d=True,\n            with_attr_label=True),\n        dict(\n            type='VADObjectRangeFilter',\n            point_cloud_range=[-15.0, -30.0, -2.0, 15.0, 30.0, 2.0]),\n        dict(\n            type='VADObjectNameFilter',\n            classes=[\n                'car', 'van', 'truck', 'bicycle', 'traffic_sign',\n                'traffic_cone', 'traffic_light', 'pedestrian', 'others'\n            ]),\n        dict(\n            type='NormalizeMultiviewImage',\n            mean=[123.675, 116.28, 103.53],\n            std=[58.395, 57.12, 57.375],\n            to_rgb=True),\n        dict(\n            type='MultiScaleFlipAug3D',\n            img_scale=(1600, 900),\n            pts_scale_ratio=1,\n            flip=False,\n            transforms=[\n                dict(type='RandomScaleImageMultiViewImage', scales=[0.8]),\n                dict(type='PadMultiViewImage', size_divisor=32),\n                dict(\n                    type='VADFormatBundle3D',\n                    class_names=[\n                        'car', 'van', 'truck', 'bicycle', 'traffic_sign',\n                        'traffic_cone', 'traffic_light', 'pedestrian', 'others'\n                    ],\n                    with_label=False,\n                    with_ego=True),\n                dict(\n                    type='CustomCollect3D',\n                    keys=[\n                        'gt_bboxes_3d', 'gt_labels_3d', 'img',\n                        'fut_valid_flag', 'ego_his_trajs', 'ego_fut_trajs',\n                        'ego_fut_masks', 'ego_fut_cmd', 'ego_lcf_feat',\n                        'gt_attr_labels'\n                    ])\n            ])\n    ],\n    metric='bbox',\n    map_metric='chamfer')\ncheckpoint_config = dict(interval=1, max_keep_ckpts=6)\nlog_config = dict(\n    interval=50,\n    hooks=[dict(type='TextLoggerHook'),\n           dict(type='TensorboardLoggerHook')])\ndist_params = dict(backend='nccl')\nlog_level = 'INFO'\nwork_dir = './work_dirs/GenAD_config_b2d'\nload_from = None\nresume_from = None\nworkflow = [('train', 1)]\nvoxel_size = [0.15, 0.15, 4]\nimg_norm_cfg = dict(\n    mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)\nNameMapping = dict({\n    'vehicle.bh.crossbike':\n    'bicycle',\n    'vehicle.diamondback.century':\n    'bicycle',\n    'vehicle.gazelle.omafiets':\n    'bicycle',\n    'vehicle.chevrolet.impala':\n    'car',\n    'vehicle.dodge.charger_2020':\n    'car',\n    'vehicle.dodge.charger_police':\n    'car',\n    'vehicle.dodge.charger_police_2020':\n    'car',\n    'vehicle.lincoln.mkz_2017':\n    'car',\n    'vehicle.lincoln.mkz_2020':\n    'car',\n    'vehicle.mini.cooper_s_2021':\n    'car',\n    'vehicle.mercedes.coupe_2020':\n    'car',\n    'vehicle.ford.mustang':\n    'car',\n    'vehicle.nissan.patrol_2021':\n    'car',\n    'vehicle.audi.tt':\n    'car',\n    'vehicle.audi.etron':\n    'car',\n    'vehicle.ford.crown':\n    'car',\n    'vehicle.tesla.model3':\n    'car',\n    '/Game/Carla/Static/Car/4Wheeled/ParkedVehicles/FordCrown/SM_FordCrown_parked.SM_FordCrown_parked':\n    'car',\n    '/Game/Carla/Static/Car/4Wheeled/ParkedVehicles/Charger/SM_ChargerParked.SM_ChargerParked':\n    'car',\n    '/Game/Carla/Static/Car/4Wheeled/ParkedVehicles/Lincoln/SM_LincolnParked.SM_LincolnParked':\n    'car',\n    '/Game/Carla/Static/Car/4Wheeled/ParkedVehicles/MercedesCCC/SM_MercedesCCC_Parked.SM_MercedesCCC_Parked':\n    'car',\n    '/Game/Carla/Static/Car/4Wheeled/ParkedVehicles/Mini2021/SM_Mini2021_parked.SM_Mini2021_parked':\n    'car',\n    '/Game/Carla/Static/Car/4Wheeled/ParkedVehicles/NissanPatrol2021/SM_NissanPatrol2021_parked.SM_NissanPatrol2021_parked':\n    'car',\n    '/Game/Carla/Static/Car/4Wheeled/ParkedVehicles/TeslaM3/SM_TeslaM3_parked.SM_TeslaM3_parked':\n    'car',\n    '/Game/Carla/Static/Car/4Wheeled/ParkedVehicles/VolkswagenT2/SM_VolkswagenT2_2021_Parked.SM_VolkswagenT2_2021_Parked':\n    'van',\n    'vehicle.ford.ambulance':\n    'van',\n    'vehicle.carlamotors.firetruck':\n    'truck',\n    'traffic.speed_limit.30':\n    'traffic_sign',\n    'traffic.speed_limit.40':\n    'traffic_sign',\n    'traffic.speed_limit.50':\n    'traffic_sign',\n    'traffic.speed_limit.60':\n    'traffic_sign',\n    'traffic.speed_limit.90':\n    'traffic_sign',\n    'traffic.speed_limit.120':\n    'traffic_sign',\n    'traffic.stop':\n    'traffic_sign',\n    'traffic.yield':\n    'traffic_sign',\n    'traffic.traffic_light':\n    'traffic_light',\n    'static.prop.warningconstruction':\n    'traffic_cone',\n    'static.prop.warningaccident':\n    'traffic_cone',\n    'static.prop.trafficwarning':\n    'traffic_cone',\n    'static.prop.constructioncone':\n    'traffic_cone',\n    'walker.pedestrian.0001':\n    'pedestrian',\n    'walker.pedestrian.0004':\n    'pedestrian',\n    'walker.pedestrian.0005':\n    'pedestrian',\n    'walker.pedestrian.0007':\n    'pedestrian',\n    'walker.pedestrian.0013':\n    'pedestrian',\n    'walker.pedestrian.0014':\n    'pedestrian',\n    'walker.pedestrian.0017':\n    'pedestrian',\n    'walker.pedestrian.0018':\n    'pedestrian',\n    'walker.pedestrian.0019':\n    'pedestrian',\n    'walker.pedestrian.0020':\n    'pedestrian',\n    'walker.pedestrian.0022':\n    'pedestrian',\n    'walker.pedestrian.0025':\n    'pedestrian',\n    'walker.pedestrian.0035':\n    'pedestrian',\n    'walker.pedestrian.0041':\n    'pedestrian',\n    'walker.pedestrian.0046':\n    'pedestrian',\n    'walker.pedestrian.0047':\n    'pedestrian',\n    'static.prop.dirtdebris01':\n    'others',\n    'static.prop.dirtdebris02':\n    'others'\n})\neval_cfg = dict(\n    dist_ths=[0.5, 1.0, 2.0, 4.0],\n    dist_th_tp=2.0,\n    min_recall=0.1,\n    min_precision=0.1,\n    mean_ap_weight=5,\n    class_names=[\n        'car', 'van', 'truck', 'bicycle', 'traffic_sign', 'traffic_cone',\n        'traffic_light', 'pedestrian'\n    ],\n    tp_metrics=['trans_err', 'scale_err', 'orient_err', 'vel_err'],\n    err_name_maping=dict(\n        trans_err='mATE',\n        scale_err='mASE',\n        orient_err='mAOE',\n        vel_err='mAVE',\n        attr_err='mAAE'),\n    class_range=dict(\n        car=(50, 50),\n        van=(50, 50),\n        truck=(50, 50),\n        bicycle=(40, 40),\n        traffic_sign=(30, 30),\n        traffic_cone=(30, 30),\n        traffic_light=(30, 30),\n        pedestrian=(40, 40)))\nnum_classes = 9\nmap_classes = [\n    'Broken', 'Solid', 'SolidSolid', 'Center', 'TrafficLight', 'StopSign'\n]\nmap_num_vec = 100\nmap_fixed_ptsnum_per_gt_line = 20\nmap_fixed_ptsnum_per_pred_line = 20\nmap_eval_use_same_gt_sample_num_flag = True\nmap_num_classes = 6\npast_frames = 2\nfuture_frames = 6\n_dim_ = 256\n_pos_dim_ = 128\n_ffn_dim_ = 512\n_num_levels_ = 1\nbev_h_ = 100\nbev_w_ = 100\nqueue_length = 3\ntotal_epochs = 6\nmodel = dict(\n    type='GenAD',\n    use_grid_mask=True,\n    video_test_mode=True,\n    pretrained=dict(img='ckpts/resnet50-19c8e357.pth'),\n    img_backbone=dict(\n        type='ResNet',\n        depth=50,\n        num_stages=4,\n        out_indices=(3, ),\n        frozen_stages=1,\n        norm_cfg=dict(type='BN', requires_grad=False),\n        norm_eval=True,\n        style='pytorch'),\n    img_neck=dict(\n        type='FPN',\n        in_channels=[2048],\n        out_channels=256,\n        start_level=0,\n        add_extra_convs='on_output',\n        num_outs=1,\n        relu_before_extra_convs=True),\n    pts_bbox_head=dict(\n        type='GenADHead',\n        map_thresh=0.5,\n        dis_thresh=0.2,\n        pe_normalization=True,\n        tot_epoch=6,\n        use_traj_lr_warmup=False,\n        query_thresh=0.0,\n        query_use_fix_pad=False,\n        ego_his_encoder=None,\n        ego_lcf_feat_idx=None,\n        valid_fut_ts=6,\n        ego_fut_mode=6,\n        agent_dim=300,\n        ego_agent_decoder=dict(\n            type='CustomTransformerDecoder',\n            num_layers=1,\n            return_intermediate=False,\n            transformerlayers=dict(\n                type='BaseTransformerLayer',\n                attn_cfgs=[\n                    dict(\n                        type='MultiheadAttention',\n                        embed_dims=256,\n                        num_heads=8,\n                        dropout=0.0)\n                ],\n                feedforward_channels=512,\n                ffn_dropout=0.0,\n                operation_order=('cross_attn', 'norm', 'ffn', 'norm'))),\n        ego_map_decoder=dict(\n            type='CustomTransformerDecoder',\n            num_layers=1,\n            return_intermediate=False,\n            transformerlayers=dict(\n                type='BaseTransformerLayer',\n                attn_cfgs=[\n                    dict(\n                        type='MultiheadAttention',\n                        embed_dims=256,\n                        num_heads=8,\n                        dropout=0.0)\n                ],\n                feedforward_channels=512,\n                ffn_dropout=0.0,\n                operation_order=('cross_attn', 'norm', 'ffn', 'norm'))),\n        motion_decoder=dict(\n            type='CustomTransformerDecoder',\n            num_layers=1,\n            return_intermediate=False,\n            transformerlayers=dict(\n                type='BaseTransformerLayer',\n                attn_cfgs=[\n                    dict(\n                        type='MultiheadAttention',\n                        embed_dims=256,\n                        num_heads=8,\n                        dropout=0.0)\n                ],\n                feedforward_channels=512,\n                ffn_dropout=0.0,\n                operation_order=('cross_attn', 'norm', 'ffn', 'norm'))),\n        motion_map_decoder=dict(\n            type='CustomTransformerDecoder',\n            num_layers=1,\n            return_intermediate=False,\n            transformerlayers=dict(\n                type='BaseTransformerLayer',\n                attn_cfgs=[\n                    dict(\n                        type='MultiheadAttention',\n                        embed_dims=256,\n                        num_heads=8,\n                        dropout=0.0)\n                ],\n                feedforward_channels=512,\n                ffn_dropout=0.0,\n                operation_order=('cross_attn', 'norm', 'ffn', 'norm'))),\n        use_pe=True,\n        bev_h=100,\n        bev_w=100,\n        num_query=300,\n        num_classes=9,\n        in_channels=256,\n        sync_cls_avg_factor=True,\n        with_box_refine=True,\n        as_two_stage=False,\n        map_num_vec=100,\n        map_num_classes=6,\n        map_num_pts_per_vec=20,\n        map_num_pts_per_gt_vec=20,\n        map_query_embed_type='instance_pts',\n        map_transform_method='minmax',\n        map_gt_shift_pts_pattern='v2',\n        map_dir_interval=1,\n        map_code_size=2,\n        map_code_weights=[1.0, 1.0, 1.0, 1.0],\n        transformer=dict(\n            type='VADPerceptionTransformer',\n            map_num_vec=100,\n            map_num_pts_per_vec=20,\n            rotate_prev_bev=True,\n            use_shift=True,\n            use_can_bus=True,\n            embed_dims=256,\n            encoder=dict(\n                type='BEVFormerEncoder',\n                num_layers=3,\n                pc_range=[-15.0, -30.0, -2.0, 15.0, 30.0, 2.0],\n                num_points_in_pillar=4,\n                return_intermediate=False,\n                transformerlayers=dict(\n                    type='BEVFormerLayer',\n                    attn_cfgs=[\n                        dict(\n                            type='TemporalSelfAttention',\n                            embed_dims=256,\n                            num_levels=1),\n                        dict(\n                            type='SpatialCrossAttention',\n                            pc_range=[-15.0, -30.0, -2.0, 15.0, 30.0, 2.0],\n                            deformable_attention=dict(\n                                type='MSDeformableAttention3D',\n                                embed_dims=256,\n                                num_points=8,\n                                num_levels=1),\n                            embed_dims=256)\n                    ],\n                    feedforward_channels=512,\n                    ffn_dropout=0.0,\n                    operation_order=('self_attn', 'norm', 'cross_attn', 'norm',\n                                     'ffn', 'norm'))),\n            decoder=dict(\n                type='DetectionTransformerDecoder',\n                num_layers=3,\n                return_intermediate=True,\n                transformerlayers=dict(\n                    type='DetrTransformerDecoderLayer',\n                    attn_cfgs=[\n                        dict(\n                            type='MultiheadAttention',\n                            embed_dims=256,\n                            num_heads=8,\n                            dropout=0.0),\n                        dict(\n                            type='CustomMSDeformableAttention',\n                            embed_dims=256,\n                            num_levels=1)\n                    ],\n                    feedforward_channels=512,\n                    ffn_dropout=0.0,\n                    operation_order=('self_attn', 'norm', 'cross_attn', 'norm',\n                                     'ffn', 'norm'))),\n            map_decoder=dict(\n                type='MapDetectionTransformerDecoder',\n                num_layers=3,\n                return_intermediate=True,\n                transformerlayers=dict(\n                    type='DetrTransformerDecoderLayer',\n                    attn_cfgs=[\n                        dict(\n                            type='MultiheadAttention',\n                            embed_dims=256,\n                            num_heads=8,\n                            dropout=0.0),\n                        dict(\n                            type='CustomMSDeformableAttention',\n                            embed_dims=256,\n                            num_levels=1)\n                    ],\n                    feedforward_channels=512,\n                    ffn_dropout=0.0,\n                    operation_order=('self_attn', 'norm', 'cross_attn', 'norm',\n                                     'ffn', 'norm')))),\n        bbox_coder=dict(\n            type='CustomNMSFreeCoder',\n            post_center_range=[-20, -35, -10.0, 20, 35, 10.0],\n            pc_range=[-15.0, -30.0, -2.0, 15.0, 30.0, 2.0],\n            max_num=100,\n            voxel_size=[0.15, 0.15, 4],\n            num_classes=9),\n        map_bbox_coder=dict(\n            type='MapNMSFreeCoder',\n            post_center_range=[-20, -35, -20, -35, 20, 35, 20, 35],\n            pc_range=[-15.0, -30.0, -2.0, 15.0, 30.0, 2.0],\n            max_num=50,\n            voxel_size=[0.15, 0.15, 4],\n            num_classes=6),\n        positional_encoding=dict(\n            type='LearnedPositionalEncoding',\n            num_feats=128,\n            row_num_embed=100,\n            col_num_embed=100),\n        loss_cls=dict(\n            type='FocalLoss',\n            use_sigmoid=True,\n            gamma=2.0,\n            alpha=0.25,\n            loss_weight=2.0),\n        loss_bbox=dict(type='L1Loss', loss_weight=0.25),\n        loss_traj=dict(type='L1Loss', loss_weight=0.2),\n        loss_traj_cls=dict(\n            type='FocalLoss',\n            use_sigmoid=True,\n            gamma=2.0,\n            alpha=0.25,\n            loss_weight=0.2),\n        loss_iou=dict(type='GIoULoss', loss_weight=0.0),\n        loss_map_cls=dict(\n            type='FocalLoss',\n            use_sigmoid=True,\n            gamma=2.0,\n            alpha=0.25,\n            loss_weight=2.0),\n        loss_map_bbox=dict(type='L1Loss', loss_weight=0.0),\n        loss_map_iou=dict(type='GIoULoss', loss_weight=0.0),\n        loss_map_pts=dict(type='PtsL1Loss', loss_weight=1.0),\n        loss_map_dir=dict(type='PtsDirCosLoss', loss_weight=0.005),\n        loss_plan_reg=dict(type='L1Loss', loss_weight=1.0),\n        loss_plan_bound=dict(\n            type='PlanMapBoundLoss', loss_weight=1.0, dis_thresh=1.0),\n        loss_plan_col=dict(type='PlanCollisionLoss', loss_weight=1.0),\n        loss_plan_dir=dict(type='PlanMapDirectionLoss', loss_weight=0.5),\n        loss_vae_gen=dict(type='ProbabilisticLoss', loss_weight=1.0)),\n    train_cfg=dict(\n        pts=dict(\n            grid_size=[512, 512, 1],\n            voxel_size=[0.15, 0.15, 4],\n            point_cloud_range=[-15.0, -30.0, -2.0, 15.0, 30.0, 2.0],\n            out_size_factor=4,\n            assigner=dict(\n                type='HungarianAssigner3D',\n                cls_cost=dict(type='FocalLossCost', weight=2.0),\n                reg_cost=dict(type='BBox3DL1Cost', weight=0.25),\n                iou_cost=dict(type='IoUCost', weight=0.0),\n                pc_range=[-15.0, -30.0, -2.0, 15.0, 30.0, 2.0]),\n            map_assigner=dict(\n                type='MapHungarianAssigner3D',\n                cls_cost=dict(type='FocalLossCost', weight=2.0),\n                reg_cost=dict(\n                    type='BBoxL1Cost', weight=0.0, box_format='xywh'),\n                iou_cost=dict(type='IoUCost', iou_mode='giou', weight=0.0),\n                pts_cost=dict(type='OrderedPtsL1Cost', weight=1.0),\n                pc_range=[-15.0, -30.0, -2.0, 15.0, 30.0, 2.0]))))\ninfo_root = 'data/infos'\nmap_root = 'data/bench2drive/maps'\nmap_file = 'data/infos/b2d_map_infos.pkl'\nann_file_train = 'data/infos/b2d_infos_train.pkl'\nann_file_val = 'data/infos/b2d_infos_val.pkl'\nann_file_test = 'data/infos/b2d_infos_val.pkl'\ninference_only_pipeline = [\n    dict(type='LoadMultiViewImageFromFiles', to_float32=True),\n    dict(\n        type='NormalizeMultiviewImage',\n        mean=[123.675, 116.28, 103.53],\n        std=[58.395, 57.12, 57.375],\n        to_rgb=True),\n    dict(type='PadMultiViewImage', size_divisor=32),\n    dict(\n        type='MultiScaleFlipAug3D',\n        img_scale=(1600, 900),\n        pts_scale_ratio=1,\n        flip=False,\n        transforms=[\n            dict(type='RandomScaleImageMultiViewImage', scales=[0.8]),\n            dict(type='PadMultiViewImage', size_divisor=32),\n            dict(\n                type='VADFormatBundle3D',\n                class_names=[\n                    'car', 'van', 'truck', 'bicycle', 'traffic_sign',\n                    'traffic_cone', 'traffic_light', 'pedestrian', 'others'\n                ],\n                with_label=False,\n                with_ego=True),\n            dict(type='CustomCollect3D', keys=['img', 'ego_fut_cmd'])\n        ])\n]\noptimizer = dict(\n    type='AdamW',\n    lr=0.0002,\n    paramwise_cfg=dict(custom_keys=dict(img_backbone=dict(lr_mult=0.1))),\n    weight_decay=0.01)\noptimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))\nlr_config = dict(\n    by_epoch=False,\n    policy='CosineAnnealing',\n    warmup='linear',\n    warmup_iters=500,\n    warmup_ratio=0.3333333333333333,\n    min_lr_ratio=0.001)\nrunner = dict(type='EpochBasedRunner', max_epochs=6)\nfind_unused_parameters = True\ncustom_hooks = [dict(type='CustomSetEpochInfoHook')]\ngpu_ids = range(0, 1)\n", "seed": 0, "exp_name": "GenAD_config_b2d.py"}
+{"mode": "train", "epoch": 1, "iter": 50, "lr": 8e-05, "memory": 8494, "data_time": 0.13884, "loss_cls": 1.41891, "loss_bbox": 1.71781, "loss_traj": 1.74877, "loss_traj_cls": 0.14629, "loss_map_cls": 1.3291, "loss_map_bbox": 0.0, "loss_map_iou": 0.0, "loss_map_pts": 6.35612, "loss_map_dir": 0.09256, "loss_plan_reg": 0.17353, "loss_plan_bound": 0.0, "loss_plan_col": 0.0, "loss_plan_dir": 0.0, "d0.loss_cls": 1.45409, "d0.loss_bbox": 1.72049, "d1.loss_cls": 1.40296, "d1.loss_bbox": 1.72447, "d0.loss_map_cls": 1.37195, "d0.loss_map_bbox": 0.0, "d0.loss_map_iou": 0.0, "d0.loss_map_pts": 6.71753, "d0.loss_map_dir": 0.09488, "d1.loss_map_cls": 1.33623, "d1.loss_map_bbox": 0.0, "d1.loss_map_iou": 0.0, "d1.loss_map_pts": 6.43922, "d1.loss_map_dir": 0.09356, "loss_vae_gen": 0.03039, "loss": 35.36887, "grad_norm": 41.56687, "time": 0.65558}
+{"mode": "train", "epoch": 1, "iter": 100, "lr": 9e-05, "memory": 8494, "data_time": 0.03709, "loss_cls": 1.21195, "loss_bbox": 1.62593, "loss_traj": 1.57033, "loss_traj_cls": 0.13658, "loss_map_cls": 0.75363, "loss_map_bbox": 0.0, "loss_map_iou": 0.0, "loss_map_pts": 4.38475, "loss_map_dir": 0.08034, "loss_plan_reg": 0.14843, "loss_plan_bound": 0.0, "loss_plan_col": 0.0, "loss_plan_dir": 0.0, "d0.loss_cls": 1.17208, "d0.loss_bbox": 1.56816, "d1.loss_cls": 1.19265, "d1.loss_bbox": 1.57286, "d0.loss_map_cls": 0.75884, "d0.loss_map_bbox": 0.0, "d0.loss_map_iou": 0.0, "d0.loss_map_pts": 4.25897, "d0.loss_map_dir": 0.07717, "d1.loss_map_cls": 0.75242, "d1.loss_map_bbox": 0.0, "d1.loss_map_iou": 0.0, "d1.loss_map_pts": 4.26995, "d1.loss_map_dir": 0.07924, "loss_vae_gen": 0.00286, "loss": 25.61713, "grad_norm": 40.91608, "time": 0.51697}
+{"mode": "train", "epoch": 1, "iter": 150, "lr": 0.00011, "memory": 8494, "data_time": 0.03725, "loss_cls": 1.11486, "loss_bbox": 1.64284, "loss_traj": 1.04021, "loss_traj_cls": 0.13315, "loss_map_cls": 0.71101, "loss_map_bbox": 0.0, "loss_map_iou": 0.0, "loss_map_pts": 3.52099, "loss_map_dir": 0.06382, "loss_plan_reg": 0.1131, "loss_plan_bound": 0.0, "loss_plan_col": 0.0, "loss_plan_dir": 0.0, "d0.loss_cls": 1.01967, "d0.loss_bbox": 1.57587, "d1.loss_cls": 1.09074, "d1.loss_bbox": 1.56643, "d0.loss_map_cls": 0.70636, "d0.loss_map_bbox": 0.0, "d0.loss_map_iou": 0.0, "d0.loss_map_pts": 3.42962, "d0.loss_map_dir": 0.06285, "d1.loss_map_cls": 0.70884, "d1.loss_map_bbox": 0.0, "d1.loss_map_iou": 0.0, "d1.loss_map_pts": 3.46119, "d1.loss_map_dir": 0.06338, "loss_vae_gen": 0.00269, "loss": 22.02763, "grad_norm": 40.69017, "time": 0.51409}
+{"mode": "train", "epoch": 1, "iter": 200, "lr": 0.00012, "memory": 8494, "data_time": 0.07033, "loss_cls": 1.09031, "loss_bbox": 1.4055, "loss_traj": 1.02646, "loss_traj_cls": 0.13682, "loss_map_cls": 0.78629, "loss_map_bbox": 0.0, "loss_map_iou": 0.0, "loss_map_pts": 3.31181, "loss_map_dir": 0.05782, "loss_plan_reg": 0.1581, "loss_plan_bound": 0.0, "loss_plan_col": 0.0, "loss_plan_dir": 0.0, "d0.loss_cls": 1.03103, "d0.loss_bbox": 1.37762, "d1.loss_cls": 1.09068, "d1.loss_bbox": 1.39371, "d0.loss_map_cls": 0.77216, "d0.loss_map_bbox": 0.0, "d0.loss_map_iou": 0.0, "d0.loss_map_pts": 3.25041, "d0.loss_map_dir": 0.05623, "d1.loss_map_cls": 0.78438, "d1.loss_map_bbox": 0.0, "d1.loss_map_iou": 0.0, "d1.loss_map_pts": 3.26293, "d1.loss_map_dir": 0.05741, "loss_vae_gen": 0.00288, "loss": 21.05258, "grad_norm": 49.73068, "time": 0.53878}
+{"mode": "train", "epoch": 1, "iter": 250, "lr": 0.00013, "memory": 8494, "data_time": 0.02916, "loss_cls": 1.068, "loss_bbox": 1.35327, "loss_traj": 1.32546, "loss_traj_cls": 0.13112, "loss_map_cls": 0.69527, "loss_map_bbox": 0.0, "loss_map_iou": 0.0, "loss_map_pts": 3.15151, "loss_map_dir": 0.05627, "loss_plan_reg": 0.16075, "loss_plan_bound": 0.0, "loss_plan_col": 0.0, "loss_plan_dir": 0.0, "d0.loss_cls": 0.95736, "d0.loss_bbox": 1.35561, "d1.loss_cls": 1.03383, "d1.loss_bbox": 1.33779, "d0.loss_map_cls": 0.68224, "d0.loss_map_bbox": 0.0, "d0.loss_map_iou": 0.0, "d0.loss_map_pts": 3.14028, "d0.loss_map_dir": 0.05626, "d1.loss_map_cls": 0.69662, "d1.loss_map_bbox": 0.0, "d1.loss_map_iou": 0.0, "d1.loss_map_pts": 3.12826, "d1.loss_map_dir": 0.0558, "loss_vae_gen": 0.00196, "loss": 20.38765, "grad_norm": 46.97533, "time": 0.50611}
+{"mode": "train", "epoch": 1, "iter": 300, "lr": 0.00015, "memory": 8495, "data_time": 0.04951, "loss_cls": 1.0105, "loss_bbox": 1.2045, "loss_traj": 1.10531, "loss_traj_cls": 0.12885, "loss_map_cls": 0.72558, "loss_map_bbox": 0.0, "loss_map_iou": 0.0, "loss_map_pts": 3.2146, "loss_map_dir": 0.0503, "loss_plan_reg": 0.13109, "loss_plan_bound": 0.0, "loss_plan_col": 0.0, "loss_plan_dir": 0.0, "d0.loss_cls": 0.89587, "d0.loss_bbox": 1.24509, "d1.loss_cls": 0.96248, "d1.loss_bbox": 1.21346, "d0.loss_map_cls": 0.69704, "d0.loss_map_bbox": 0.0, "d0.loss_map_iou": 0.0, "d0.loss_map_pts": 3.09375, "d0.loss_map_dir": 0.04884, "d1.loss_map_cls": 0.72199, "d1.loss_map_bbox": 0.0, "d1.loss_map_iou": 0.0, "d1.loss_map_pts": 3.13981, "d1.loss_map_dir": 0.04985, "loss_vae_gen": 0.00191, "loss": 19.64081, "grad_norm": 48.04271, "time": 0.50706}
+{"mode": "train", "epoch": 1, "iter": 350, "lr": 0.00016, "memory": 8495, "data_time": 0.03778, "loss_cls": 1.06324, "loss_bbox": 1.28378, "loss_traj": 0.89384, "loss_traj_cls": 0.13387, "loss_map_cls": 0.69681, "loss_map_bbox": 0.0, "loss_map_iou": 0.0, "loss_map_pts": 2.8678, "loss_map_dir": 0.04711, "loss_plan_reg": 0.16238, "loss_plan_bound": 0.0, "loss_plan_col": 0.0, "loss_plan_dir": 0.0, "d0.loss_cls": 0.97306, "d0.loss_bbox": 1.32794, "d1.loss_cls": 1.04407, "d1.loss_bbox": 1.29131, "d0.loss_map_cls": 0.664, "d0.loss_map_bbox": 0.0, "d0.loss_map_iou": 0.0, "d0.loss_map_pts": 2.82712, "d0.loss_map_dir": 0.04757, "d1.loss_map_cls": 0.69473, "d1.loss_map_bbox": 0.0, "d1.loss_map_iou": 0.0, "d1.loss_map_pts": 2.83693, "d1.loss_map_dir": 0.04698, "loss_vae_gen": 0.00116, "loss": 18.90371, "grad_norm": 51.59715, "time": 0.51373}
+{"mode": "train", "epoch": 1, "iter": 400, "lr": 0.00017, "memory": 8495, "data_time": 0.04639, "loss_cls": 1.05345, "loss_bbox": 1.45453, "loss_traj": 1.07566, "loss_traj_cls": 0.12975, "loss_map_cls": 0.70507, "loss_map_bbox": 0.0, "loss_map_iou": 0.0, "loss_map_pts": 3.17895, "loss_map_dir": 0.04788, "loss_plan_reg": 0.16547, "loss_plan_bound": 0.0, "loss_plan_col": 0.0, "loss_plan_dir": 0.0, "d0.loss_cls": 0.98056, "d0.loss_bbox": 1.34076, "d1.loss_cls": 1.04872, "d1.loss_bbox": 1.33337, "d0.loss_map_cls": 0.66494, "d0.loss_map_bbox": 0.0, "d0.loss_map_iou": 0.0, "d0.loss_map_pts": 3.14847, "d0.loss_map_dir": 0.04874, "d1.loss_map_cls": 0.70193, "d1.loss_map_bbox": 0.0, "d1.loss_map_iou": 0.0, "d1.loss_map_pts": 3.15375, "d1.loss_map_dir": 0.04782, "loss_vae_gen": 0.00141, "loss": 20.28121, "grad_norm": 58.06685, "time": 0.52213}

GenAD/with_hlc/20260213_081015.log.json

@@ -0,0 +1,6 @@
+{"config": "point_cloud_range = [-15.0, -30.0, -2.0, 15.0, 30.0, 2.0]\nclass_names = [\n    'car', 'van', 'truck', 'bicycle', 'traffic_sign', 'traffic_cone',\n    'traffic_light', 'pedestrian', 'others'\n]\ndataset_type = 'B2D_VAD_Dataset'\ndata_root = 'data/bench2drive'\ninput_modality = dict(\n    use_lidar=False,\n    use_camera=True,\n    use_radar=False,\n    use_map=False,\n    use_external=True)\nfile_client_args = dict(backend='disk')\ntrain_pipeline = [\n    dict(type='LoadMultiViewImageFromFiles', to_float32=True),\n    dict(type='PhotoMetricDistortionMultiViewImage'),\n    dict(\n        type='LoadAnnotations3D',\n        with_bbox_3d=True,\n        with_label_3d=True,\n        with_attr_label=True),\n    dict(\n        type='VADObjectRangeFilter',\n        point_cloud_range=[-15.0, -30.0, -2.0, 15.0, 30.0, 2.0]),\n    dict(\n        type='VADObjectNameFilter',\n        classes=[\n            'car', 'van', 'truck', 'bicycle', 'traffic_sign', 'traffic_cone',\n            'traffic_light', 'pedestrian', 'others'\n        ]),\n    dict(\n        type='NormalizeMultiviewImage',\n        mean=[123.675, 116.28, 103.53],\n        std=[58.395, 57.12, 57.375],\n        to_rgb=True),\n    dict(type='RandomScaleImageMultiViewImage', scales=[0.8]),\n    dict(type='PadMultiViewImage', size_divisor=32),\n    dict(\n        type='VADFormatBundle3D',\n        class_names=[\n            'car', 'van', 'truck', 'bicycle', 'traffic_sign', 'traffic_cone',\n            'traffic_light', 'pedestrian', 'others'\n        ],\n        with_ego=True),\n    dict(\n        type='CustomCollect3D',\n        keys=[\n            'gt_bboxes_3d', 'gt_labels_3d', 'img', 'ego_his_trajs',\n            'gt_attr_labels', 'ego_fut_trajs', 'ego_fut_masks', 'ego_fut_cmd',\n            'ego_lcf_feat'\n        ])\n]\ntest_pipeline = [\n    dict(type='LoadMultiViewImageFromFiles', to_float32=True),\n    dict(\n        type='LoadAnnotations3D',\n        with_bbox_3d=True,\n        with_label_3d=True,\n        with_attr_label=True),\n    dict(\n        type='VADObjectRangeFilter',\n        point_cloud_range=[-15.0, -30.0, -2.0, 15.0, 30.0, 2.0]),\n    dict(\n        type='VADObjectNameFilter',\n        classes=[\n            'car', 'van', 'truck', 'bicycle', 'traffic_sign', 'traffic_cone',\n            'traffic_light', 'pedestrian', 'others'\n        ]),\n    dict(\n        type='NormalizeMultiviewImage',\n        mean=[123.675, 116.28, 103.53],\n        std=[58.395, 57.12, 57.375],\n        to_rgb=True),\n    dict(\n        type='MultiScaleFlipAug3D',\n        img_scale=(1600, 900),\n        pts_scale_ratio=1,\n        flip=False,\n        transforms=[\n            dict(type='RandomScaleImageMultiViewImage', scales=[0.8]),\n            dict(type='PadMultiViewImage', size_divisor=32),\n            dict(\n                type='VADFormatBundle3D',\n                class_names=[\n                    'car', 'van', 'truck', 'bicycle', 'traffic_sign',\n                    'traffic_cone', 'traffic_light', 'pedestrian', 'others'\n                ],\n                with_label=False,\n                with_ego=True),\n            dict(\n                type='CustomCollect3D',\n                keys=[\n                    'gt_bboxes_3d', 'gt_labels_3d', 'img', 'fut_valid_flag',\n                    'ego_his_trajs', 'ego_fut_trajs', 'ego_fut_masks',\n                    'ego_fut_cmd', 'ego_lcf_feat', 'gt_attr_labels'\n                ])\n        ])\n]\neval_pipeline = [\n    dict(\n        type='LoadPointsFromFile',\n        coord_type='LIDAR',\n        load_dim=5,\n        use_dim=5,\n        file_client_args=dict(backend='disk')),\n    dict(\n        type='LoadPointsFromMultiSweeps',\n        sweeps_num=10,\n        file_client_args=dict(backend='disk')),\n    dict(\n        type='DefaultFormatBundle3D',\n        class_names=[\n            'car', 'truck', 'trailer', 'bus', 'construction_vehicle',\n            'bicycle', 'motorcycle', 'pedestrian', 'traffic_cone', 'barrier'\n        ],\n        with_label=False),\n    dict(type='Collect3D', keys=['points'])\n]\ndata = dict(\n    samples_per_gpu=1,\n    workers_per_gpu=4,\n    train=dict(\n        type='B2D_VAD_Dataset',\n        data_root='data/bench2drive',\n        ann_file='data/infos/b2d_infos_train.pkl',\n        pipeline=[\n            dict(type='LoadMultiViewImageFromFiles', to_float32=True),\n            dict(type='PhotoMetricDistortionMultiViewImage'),\n            dict(\n                type='LoadAnnotations3D',\n                with_bbox_3d=True,\n                with_label_3d=True,\n                with_attr_label=True),\n            dict(\n                type='VADObjectRangeFilter',\n                point_cloud_range=[-15.0, -30.0, -2.0, 15.0, 30.0, 2.0]),\n            dict(\n                type='VADObjectNameFilter',\n                classes=[\n                    'car', 'van', 'truck', 'bicycle', 'traffic_sign',\n                    'traffic_cone', 'traffic_light', 'pedestrian', 'others'\n                ]),\n            dict(\n                type='NormalizeMultiviewImage',\n                mean=[123.675, 116.28, 103.53],\n                std=[58.395, 57.12, 57.375],\n                to_rgb=True),\n            dict(type='RandomScaleImageMultiViewImage', scales=[0.8]),\n            dict(type='PadMultiViewImage', size_divisor=32),\n            dict(\n                type='VADFormatBundle3D',\n                class_names=[\n                    'car', 'van', 'truck', 'bicycle', 'traffic_sign',\n                    'traffic_cone', 'traffic_light', 'pedestrian', 'others'\n                ],\n                with_ego=True),\n            dict(\n                type='CustomCollect3D',\n                keys=[\n                    'gt_bboxes_3d', 'gt_labels_3d', 'img', 'ego_his_trajs',\n                    'gt_attr_labels', 'ego_fut_trajs', 'ego_fut_masks',\n                    'ego_fut_cmd', 'ego_lcf_feat'\n                ])\n        ],\n        classes=[\n            'car', 'van', 'truck', 'bicycle', 'traffic_sign', 'traffic_cone',\n            'traffic_light', 'pedestrian', 'others'\n        ],\n        modality=dict(\n            use_lidar=False,\n            use_camera=True,\n            use_radar=False,\n            use_map=False,\n            use_external=True),\n        test_mode=False,\n        box_type_3d='LiDAR',\n        name_mapping=dict({\n            'vehicle.bh.crossbike':\n            'bicycle',\n            'vehicle.diamondback.century':\n            'bicycle',\n            'vehicle.gazelle.omafiets':\n            'bicycle',\n            'vehicle.chevrolet.impala':\n            'car',\n            'vehicle.dodge.charger_2020':\n            'car',\n            'vehicle.dodge.charger_police':\n            'car',\n            'vehicle.dodge.charger_police_2020':\n            'car',\n            'vehicle.lincoln.mkz_2017':\n            'car',\n            'vehicle.lincoln.mkz_2020':\n            'car',\n            'vehicle.mini.cooper_s_2021':\n            'car',\n            'vehicle.mercedes.coupe_2020':\n            'car',\n            'vehicle.ford.mustang':\n            'car',\n            'vehicle.nissan.patrol_2021':\n            'car',\n            'vehicle.audi.tt':\n            'car',\n            'vehicle.audi.etron':\n            'car',\n            'vehicle.ford.crown':\n            'car',\n            'vehicle.tesla.model3':\n            'car',\n            '/Game/Carla/Static/Car/4Wheeled/ParkedVehicles/FordCrown/SM_FordCrown_parked.SM_FordCrown_parked':\n            'car',\n            '/Game/Carla/Static/Car/4Wheeled/ParkedVehicles/Charger/SM_ChargerParked.SM_ChargerParked':\n            'car',\n            '/Game/Carla/Static/Car/4Wheeled/ParkedVehicles/Lincoln/SM_LincolnParked.SM_LincolnParked':\n            'car',\n            '/Game/Carla/Static/Car/4Wheeled/ParkedVehicles/MercedesCCC/SM_MercedesCCC_Parked.SM_MercedesCCC_Parked':\n            'car',\n            '/Game/Carla/Static/Car/4Wheeled/ParkedVehicles/Mini2021/SM_Mini2021_parked.SM_Mini2021_parked':\n            'car',\n            '/Game/Carla/Static/Car/4Wheeled/ParkedVehicles/NissanPatrol2021/SM_NissanPatrol2021_parked.SM_NissanPatrol2021_parked':\n            'car',\n            '/Game/Carla/Static/Car/4Wheeled/ParkedVehicles/TeslaM3/SM_TeslaM3_parked.SM_TeslaM3_parked':\n            'car',\n            '/Game/Carla/Static/Car/4Wheeled/ParkedVehicles/VolkswagenT2/SM_VolkswagenT2_2021_Parked.SM_VolkswagenT2_2021_Parked':\n            'van',\n            'vehicle.ford.ambulance':\n            'van',\n            'vehicle.carlamotors.firetruck':\n            'truck',\n            'traffic.speed_limit.30':\n            'traffic_sign',\n            'traffic.speed_limit.40':\n            'traffic_sign',\n            'traffic.speed_limit.50':\n            'traffic_sign',\n            'traffic.speed_limit.60':\n            'traffic_sign',\n            'traffic.speed_limit.90':\n            'traffic_sign',\n            'traffic.speed_limit.120':\n            'traffic_sign',\n            'traffic.stop':\n            'traffic_sign',\n            'traffic.yield':\n            'traffic_sign',\n            'traffic.traffic_light':\n            'traffic_light',\n            'static.prop.warningconstruction':\n            'traffic_cone',\n            'static.prop.warningaccident':\n            'traffic_cone',\n            'static.prop.trafficwarning':\n            'traffic_cone',\n            'static.prop.constructioncone':\n            'traffic_cone',\n            'walker.pedestrian.0001':\n            'pedestrian',\n            'walker.pedestrian.0004':\n            'pedestrian',\n            'walker.pedestrian.0005':\n            'pedestrian',\n            'walker.pedestrian.0007':\n            'pedestrian',\n            'walker.pedestrian.0013':\n            'pedestrian',\n            'walker.pedestrian.0014':\n            'pedestrian',\n            'walker.pedestrian.0017':\n            'pedestrian',\n            'walker.pedestrian.0018':\n            'pedestrian',\n            'walker.pedestrian.0019':\n            'pedestrian',\n            'walker.pedestrian.0020':\n            'pedestrian',\n            'walker.pedestrian.0022':\n            'pedestrian',\n            'walker.pedestrian.0025':\n            'pedestrian',\n            'walker.pedestrian.0035':\n            'pedestrian',\n            'walker.pedestrian.0041':\n            'pedestrian',\n            'walker.pedestrian.0046':\n            'pedestrian',\n            'walker.pedestrian.0047':\n            'pedestrian',\n            'static.prop.dirtdebris01':\n            'others',\n            'static.prop.dirtdebris02':\n            'others'\n        }),\n        map_root='data/bench2drive/maps',\n        map_file='data/infos/b2d_map_infos.pkl',\n        bev_size=(100, 100),\n        queue_length=3,\n        past_frames=2,\n        future_frames=6,\n        point_cloud_range=[-15.0, -30.0, -2.0, 15.0, 30.0, 2.0],\n        polyline_points_num=20),\n    val=dict(\n        type='B2D_VAD_Dataset',\n        ann_file='data/infos/b2d_infos_val.pkl',\n        pipeline=[\n            dict(type='LoadMultiViewImageFromFiles', to_float32=True),\n            dict(\n                type='LoadAnnotations3D',\n                with_bbox_3d=True,\n                with_label_3d=True,\n                with_attr_label=True),\n            dict(\n                type='VADObjectRangeFilter',\n                point_cloud_range=[-15.0, -30.0, -2.0, 15.0, 30.0, 2.0]),\n            dict(\n                type='VADObjectNameFilter',\n                classes=[\n                    'car', 'van', 'truck', 'bicycle', 'traffic_sign',\n                    'traffic_cone', 'traffic_light', 'pedestrian', 'others'\n                ]),\n            dict(\n                type='NormalizeMultiviewImage',\n                mean=[123.675, 116.28, 103.53],\n                std=[58.395, 57.12, 57.375],\n                to_rgb=True),\n            dict(\n                type='MultiScaleFlipAug3D',\n                img_scale=(1600, 900),\n                pts_scale_ratio=1,\n                flip=False,\n                transforms=[\n                    dict(type='RandomScaleImageMultiViewImage', scales=[0.8]),\n                    dict(type='PadMultiViewImage', size_divisor=32),\n                    dict(\n                        type='VADFormatBundle3D',\n                        class_names=[\n                            'car', 'van', 'truck', 'bicycle', 'traffic_sign',\n                            'traffic_cone', 'traffic_light', 'pedestrian',\n                            'others'\n                        ],\n                        with_label=False,\n                        with_ego=True),\n                    dict(\n                        type='CustomCollect3D',\n                        keys=[\n                            'gt_bboxes_3d', 'gt_labels_3d', 'img',\n                            'fut_valid_flag', 'ego_his_trajs', 'ego_fut_trajs',\n                            'ego_fut_masks', 'ego_fut_cmd', 'ego_lcf_feat',\n                            'gt_attr_labels'\n                        ])\n                ])\n        ],\n        classes=[\n            'car', 'van', 'truck', 'bicycle', 'traffic_sign', 'traffic_cone',\n            'traffic_light', 'pedestrian', 'others'\n        ],\n        modality=dict(\n            use_lidar=False,\n            use_camera=True,\n            use_radar=False,\n            use_map=False,\n            use_external=True),\n        test_mode=True,\n        box_type_3d='LiDAR',\n        data_root='data/bench2drive',\n        name_mapping=dict({\n            'vehicle.bh.crossbike':\n            'bicycle',\n            'vehicle.diamondback.century':\n            'bicycle',\n            'vehicle.gazelle.omafiets':\n            'bicycle',\n            'vehicle.chevrolet.impala':\n            'car',\n            'vehicle.dodge.charger_2020':\n            'car',\n            'vehicle.dodge.charger_police':\n            'car',\n            'vehicle.dodge.charger_police_2020':\n            'car',\n            'vehicle.lincoln.mkz_2017':\n            'car',\n            'vehicle.lincoln.mkz_2020':\n            'car',\n            'vehicle.mini.cooper_s_2021':\n            'car',\n            'vehicle.mercedes.coupe_2020':\n            'car',\n            'vehicle.ford.mustang':\n            'car',\n            'vehicle.nissan.patrol_2021':\n            'car',\n            'vehicle.audi.tt':\n            'car',\n            'vehicle.audi.etron':\n            'car',\n            'vehicle.ford.crown':\n            'car',\n            'vehicle.tesla.model3':\n            'car',\n            '/Game/Carla/Static/Car/4Wheeled/ParkedVehicles/FordCrown/SM_FordCrown_parked.SM_FordCrown_parked':\n            'car',\n            '/Game/Carla/Static/Car/4Wheeled/ParkedVehicles/Charger/SM_ChargerParked.SM_ChargerParked':\n            'car',\n            '/Game/Carla/Static/Car/4Wheeled/ParkedVehicles/Lincoln/SM_LincolnParked.SM_LincolnParked':\n            'car',\n            '/Game/Carla/Static/Car/4Wheeled/ParkedVehicles/MercedesCCC/SM_MercedesCCC_Parked.SM_MercedesCCC_Parked':\n            'car',\n            '/Game/Carla/Static/Car/4Wheeled/ParkedVehicles/Mini2021/SM_Mini2021_parked.SM_Mini2021_parked':\n            'car',\n            '/Game/Carla/Static/Car/4Wheeled/ParkedVehicles/NissanPatrol2021/SM_NissanPatrol2021_parked.SM_NissanPatrol2021_parked':\n            'car',\n            '/Game/Carla/Static/Car/4Wheeled/ParkedVehicles/TeslaM3/SM_TeslaM3_parked.SM_TeslaM3_parked':\n            'car',\n            '/Game/Carla/Static/Car/4Wheeled/ParkedVehicles/VolkswagenT2/SM_VolkswagenT2_2021_Parked.SM_VolkswagenT2_2021_Parked':\n            'van',\n            'vehicle.ford.ambulance':\n            'van',\n            'vehicle.carlamotors.firetruck':\n            'truck',\n            'traffic.speed_limit.30':\n            'traffic_sign',\n            'traffic.speed_limit.40':\n            'traffic_sign',\n            'traffic.speed_limit.50':\n            'traffic_sign',\n            'traffic.speed_limit.60':\n            'traffic_sign',\n            'traffic.speed_limit.90':\n            'traffic_sign',\n            'traffic.speed_limit.120':\n            'traffic_sign',\n            'traffic.stop':\n            'traffic_sign',\n            'traffic.yield':\n            'traffic_sign',\n            'traffic.traffic_light':\n            'traffic_light',\n            'static.prop.warningconstruction':\n            'traffic_cone',\n            'static.prop.warningaccident':\n            'traffic_cone',\n            'static.prop.trafficwarning':\n            'traffic_cone',\n            'static.prop.constructioncone':\n            'traffic_cone',\n            'walker.pedestrian.0001':\n            'pedestrian',\n            'walker.pedestrian.0004':\n            'pedestrian',\n            'walker.pedestrian.0005':\n            'pedestrian',\n            'walker.pedestrian.0007':\n            'pedestrian',\n            'walker.pedestrian.0013':\n            'pedestrian',\n            'walker.pedestrian.0014':\n            'pedestrian',\n            'walker.pedestrian.0017':\n            'pedestrian',\n            'walker.pedestrian.0018':\n            'pedestrian',\n            'walker.pedestrian.0019':\n            'pedestrian',\n            'walker.pedestrian.0020':\n            'pedestrian',\n            'walker.pedestrian.0022':\n            'pedestrian',\n            'walker.pedestrian.0025':\n            'pedestrian',\n            'walker.pedestrian.0035':\n            'pedestrian',\n            'walker.pedestrian.0041':\n            'pedestrian',\n            'walker.pedestrian.0046':\n            'pedestrian',\n            'walker.pedestrian.0047':\n            'pedestrian',\n            'static.prop.dirtdebris01':\n            'others',\n            'static.prop.dirtdebris02':\n            'others'\n        }),\n        map_root='data/bench2drive/maps',\n        map_file='data/infos/b2d_map_infos.pkl',\n        bev_size=(100, 100),\n        queue_length=3,\n        past_frames=2,\n        future_frames=6,\n        point_cloud_range=[-15.0, -30.0, -2.0, 15.0, 30.0, 2.0],\n        polyline_points_num=20,\n        eval_cfg=dict(\n            dist_ths=[0.5, 1.0, 2.0, 4.0],\n            dist_th_tp=2.0,\n            min_recall=0.1,\n            min_precision=0.1,\n            mean_ap_weight=5,\n            class_names=[\n                'car', 'van', 'truck', 'bicycle', 'traffic_sign',\n                'traffic_cone', 'traffic_light', 'pedestrian'\n            ],\n            tp_metrics=['trans_err', 'scale_err', 'orient_err', 'vel_err'],\n            err_name_maping=dict(\n                trans_err='mATE',\n                scale_err='mASE',\n                orient_err='mAOE',\n                vel_err='mAVE',\n                attr_err='mAAE'),\n            class_range=dict(\n                car=(50, 50),\n                van=(50, 50),\n                truck=(50, 50),\n                bicycle=(40, 40),\n                traffic_sign=(30, 30),\n                traffic_cone=(30, 30),\n                traffic_light=(30, 30),\n                pedestrian=(40, 40)))),\n    test=dict(\n        type='B2D_VAD_Dataset',\n        data_root='data/bench2drive',\n        ann_file='data/infos/b2d_infos_val.pkl',\n        pipeline=[\n            dict(type='LoadMultiViewImageFromFiles', to_float32=True),\n            dict(\n                type='LoadAnnotations3D',\n                with_bbox_3d=True,\n                with_label_3d=True,\n                with_attr_label=True),\n            dict(\n                type='VADObjectRangeFilter',\n                point_cloud_range=[-15.0, -30.0, -2.0, 15.0, 30.0, 2.0]),\n            dict(\n                type='VADObjectNameFilter',\n                classes=[\n                    'car', 'van', 'truck', 'bicycle', 'traffic_sign',\n                    'traffic_cone', 'traffic_light', 'pedestrian', 'others'\n                ]),\n            dict(\n                type='NormalizeMultiviewImage',\n                mean=[123.675, 116.28, 103.53],\n                std=[58.395, 57.12, 57.375],\n                to_rgb=True),\n            dict(\n                type='MultiScaleFlipAug3D',\n                img_scale=(1600, 900),\n                pts_scale_ratio=1,\n                flip=False,\n                transforms=[\n                    dict(type='RandomScaleImageMultiViewImage', scales=[0.8]),\n                    dict(type='PadMultiViewImage', size_divisor=32),\n                    dict(\n                        type='VADFormatBundle3D',\n                        class_names=[\n                            'car', 'van', 'truck', 'bicycle', 'traffic_sign',\n                            'traffic_cone', 'traffic_light', 'pedestrian',\n                            'others'\n                        ],\n                        with_label=False,\n                        with_ego=True),\n                    dict(\n                        type='CustomCollect3D',\n                        keys=[\n                            'gt_bboxes_3d', 'gt_labels_3d', 'img',\n                            'fut_valid_flag', 'ego_his_trajs', 'ego_fut_trajs',\n                            'ego_fut_masks', 'ego_fut_cmd', 'ego_lcf_feat',\n                            'gt_attr_labels'\n                        ])\n                ])\n        ],\n        classes=[\n            'car', 'van', 'truck', 'bicycle', 'traffic_sign', 'traffic_cone',\n            'traffic_light', 'pedestrian', 'others'\n        ],\n        modality=dict(\n            use_lidar=False,\n            use_camera=True,\n            use_radar=False,\n            use_map=False,\n            use_external=True),\n        test_mode=True,\n        box_type_3d='LiDAR',\n        name_mapping=dict({\n            'vehicle.bh.crossbike':\n            'bicycle',\n            'vehicle.diamondback.century':\n            'bicycle',\n            'vehicle.gazelle.omafiets':\n            'bicycle',\n            'vehicle.chevrolet.impala':\n            'car',\n            'vehicle.dodge.charger_2020':\n            'car',\n            'vehicle.dodge.charger_police':\n            'car',\n            'vehicle.dodge.charger_police_2020':\n            'car',\n            'vehicle.lincoln.mkz_2017':\n            'car',\n            'vehicle.lincoln.mkz_2020':\n            'car',\n            'vehicle.mini.cooper_s_2021':\n            'car',\n            'vehicle.mercedes.coupe_2020':\n            'car',\n            'vehicle.ford.mustang':\n            'car',\n            'vehicle.nissan.patrol_2021':\n            'car',\n            'vehicle.audi.tt':\n            'car',\n            'vehicle.audi.etron':\n            'car',\n            'vehicle.ford.crown':\n            'car',\n            'vehicle.tesla.model3':\n            'car',\n            '/Game/Carla/Static/Car/4Wheeled/ParkedVehicles/FordCrown/SM_FordCrown_parked.SM_FordCrown_parked':\n            'car',\n            '/Game/Carla/Static/Car/4Wheeled/ParkedVehicles/Charger/SM_ChargerParked.SM_ChargerParked':\n            'car',\n            '/Game/Carla/Static/Car/4Wheeled/ParkedVehicles/Lincoln/SM_LincolnParked.SM_LincolnParked':\n            'car',\n            '/Game/Carla/Static/Car/4Wheeled/ParkedVehicles/MercedesCCC/SM_MercedesCCC_Parked.SM_MercedesCCC_Parked':\n            'car',\n            '/Game/Carla/Static/Car/4Wheeled/ParkedVehicles/Mini2021/SM_Mini2021_parked.SM_Mini2021_parked':\n            'car',\n            '/Game/Carla/Static/Car/4Wheeled/ParkedVehicles/NissanPatrol2021/SM_NissanPatrol2021_parked.SM_NissanPatrol2021_parked':\n            'car',\n            '/Game/Carla/Static/Car/4Wheeled/ParkedVehicles/TeslaM3/SM_TeslaM3_parked.SM_TeslaM3_parked':\n            'car',\n            '/Game/Carla/Static/Car/4Wheeled/ParkedVehicles/VolkswagenT2/SM_VolkswagenT2_2021_Parked.SM_VolkswagenT2_2021_Parked':\n            'van',\n            'vehicle.ford.ambulance':\n            'van',\n            'vehicle.carlamotors.firetruck':\n            'truck',\n            'traffic.speed_limit.30':\n            'traffic_sign',\n            'traffic.speed_limit.40':\n            'traffic_sign',\n            'traffic.speed_limit.50':\n            'traffic_sign',\n            'traffic.speed_limit.60':\n            'traffic_sign',\n            'traffic.speed_limit.90':\n            'traffic_sign',\n            'traffic.speed_limit.120':\n            'traffic_sign',\n            'traffic.stop':\n            'traffic_sign',\n            'traffic.yield':\n            'traffic_sign',\n            'traffic.traffic_light':\n            'traffic_light',\n            'static.prop.warningconstruction':\n            'traffic_cone',\n            'static.prop.warningaccident':\n            'traffic_cone',\n            'static.prop.trafficwarning':\n            'traffic_cone',\n            'static.prop.constructioncone':\n            'traffic_cone',\n            'walker.pedestrian.0001':\n            'pedestrian',\n            'walker.pedestrian.0004':\n            'pedestrian',\n            'walker.pedestrian.0005':\n            'pedestrian',\n            'walker.pedestrian.0007':\n            'pedestrian',\n            'walker.pedestrian.0013':\n            'pedestrian',\n            'walker.pedestrian.0014':\n            'pedestrian',\n            'walker.pedestrian.0017':\n            'pedestrian',\n            'walker.pedestrian.0018':\n            'pedestrian',\n            'walker.pedestrian.0019':\n            'pedestrian',\n            'walker.pedestrian.0020':\n            'pedestrian',\n            'walker.pedestrian.0022':\n            'pedestrian',\n            'walker.pedestrian.0025':\n            'pedestrian',\n            'walker.pedestrian.0035':\n            'pedestrian',\n            'walker.pedestrian.0041':\n            'pedestrian',\n            'walker.pedestrian.0046':\n            'pedestrian',\n            'walker.pedestrian.0047':\n            'pedestrian',\n            'static.prop.dirtdebris01':\n            'others',\n            'static.prop.dirtdebris02':\n            'others'\n        }),\n        map_root='data/bench2drive/maps',\n        map_file='data/infos/b2d_map_infos.pkl',\n        bev_size=(100, 100),\n        queue_length=3,\n        past_frames=2,\n        future_frames=6,\n        point_cloud_range=[-15.0, -30.0, -2.0, 15.0, 30.0, 2.0],\n        polyline_points_num=20,\n        eval_cfg=dict(\n            dist_ths=[0.5, 1.0, 2.0, 4.0],\n            dist_th_tp=2.0,\n            min_recall=0.1,\n            min_precision=0.1,\n            mean_ap_weight=5,\n            class_names=[\n                'car', 'van', 'truck', 'bicycle', 'traffic_sign',\n                'traffic_cone', 'traffic_light', 'pedestrian'\n            ],\n            tp_metrics=['trans_err', 'scale_err', 'orient_err', 'vel_err'],\n            err_name_maping=dict(\n                trans_err='mATE',\n                scale_err='mASE',\n                orient_err='mAOE',\n                vel_err='mAVE',\n                attr_err='mAAE'),\n            class_range=dict(\n                car=(50, 50),\n                van=(50, 50),\n                truck=(50, 50),\n                bicycle=(40, 40),\n                traffic_sign=(30, 30),\n                traffic_cone=(30, 30),\n                traffic_light=(30, 30),\n                pedestrian=(40, 40)))),\n    shuffler_sampler=dict(type='DistributedGroupSampler'),\n    nonshuffler_sampler=dict(type='DistributedSampler'))\nevaluation = dict(\n    interval=6,\n    pipeline=[\n        dict(type='LoadMultiViewImageFromFiles', to_float32=True),\n        dict(\n            type='LoadAnnotations3D',\n            with_bbox_3d=True,\n            with_label_3d=True,\n            with_attr_label=True),\n        dict(\n            type='VADObjectRangeFilter',\n            point_cloud_range=[-15.0, -30.0, -2.0, 15.0, 30.0, 2.0]),\n        dict(\n            type='VADObjectNameFilter',\n            classes=[\n                'car', 'van', 'truck', 'bicycle', 'traffic_sign',\n                'traffic_cone', 'traffic_light', 'pedestrian', 'others'\n            ]),\n        dict(\n            type='NormalizeMultiviewImage',\n            mean=[123.675, 116.28, 103.53],\n            std=[58.395, 57.12, 57.375],\n            to_rgb=True),\n        dict(\n            type='MultiScaleFlipAug3D',\n            img_scale=(1600, 900),\n            pts_scale_ratio=1,\n            flip=False,\n            transforms=[\n                dict(type='RandomScaleImageMultiViewImage', scales=[0.8]),\n                dict(type='PadMultiViewImage', size_divisor=32),\n                dict(\n                    type='VADFormatBundle3D',\n                    class_names=[\n                        'car', 'van', 'truck', 'bicycle', 'traffic_sign',\n                        'traffic_cone', 'traffic_light', 'pedestrian', 'others'\n                    ],\n                    with_label=False,\n                    with_ego=True),\n                dict(\n                    type='CustomCollect3D',\n                    keys=[\n                        'gt_bboxes_3d', 'gt_labels_3d', 'img',\n                        'fut_valid_flag', 'ego_his_trajs', 'ego_fut_trajs',\n                        'ego_fut_masks', 'ego_fut_cmd', 'ego_lcf_feat',\n                        'gt_attr_labels'\n                    ])\n            ])\n    ],\n    metric='bbox',\n    map_metric='chamfer')\ncheckpoint_config = dict(interval=1, max_keep_ckpts=6)\nlog_config = dict(\n    interval=50,\n    hooks=[dict(type='TextLoggerHook'),\n           dict(type='TensorboardLoggerHook')])\ndist_params = dict(backend='nccl')\nlog_level = 'INFO'\nwork_dir = './work_dirs/GenAD_config_b2d'\nload_from = None\nresume_from = None\nworkflow = [('train', 1)]\nvoxel_size = [0.15, 0.15, 4]\nimg_norm_cfg = dict(\n    mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)\nNameMapping = dict({\n    'vehicle.bh.crossbike':\n    'bicycle',\n    'vehicle.diamondback.century':\n    'bicycle',\n    'vehicle.gazelle.omafiets':\n    'bicycle',\n    'vehicle.chevrolet.impala':\n    'car',\n    'vehicle.dodge.charger_2020':\n    'car',\n    'vehicle.dodge.charger_police':\n    'car',\n    'vehicle.dodge.charger_police_2020':\n    'car',\n    'vehicle.lincoln.mkz_2017':\n    'car',\n    'vehicle.lincoln.mkz_2020':\n    'car',\n    'vehicle.mini.cooper_s_2021':\n    'car',\n    'vehicle.mercedes.coupe_2020':\n    'car',\n    'vehicle.ford.mustang':\n    'car',\n    'vehicle.nissan.patrol_2021':\n    'car',\n    'vehicle.audi.tt':\n    'car',\n    'vehicle.audi.etron':\n    'car',\n    'vehicle.ford.crown':\n    'car',\n    'vehicle.tesla.model3':\n    'car',\n    '/Game/Carla/Static/Car/4Wheeled/ParkedVehicles/FordCrown/SM_FordCrown_parked.SM_FordCrown_parked':\n    'car',\n    '/Game/Carla/Static/Car/4Wheeled/ParkedVehicles/Charger/SM_ChargerParked.SM_ChargerParked':\n    'car',\n    '/Game/Carla/Static/Car/4Wheeled/ParkedVehicles/Lincoln/SM_LincolnParked.SM_LincolnParked':\n    'car',\n    '/Game/Carla/Static/Car/4Wheeled/ParkedVehicles/MercedesCCC/SM_MercedesCCC_Parked.SM_MercedesCCC_Parked':\n    'car',\n    '/Game/Carla/Static/Car/4Wheeled/ParkedVehicles/Mini2021/SM_Mini2021_parked.SM_Mini2021_parked':\n    'car',\n    '/Game/Carla/Static/Car/4Wheeled/ParkedVehicles/NissanPatrol2021/SM_NissanPatrol2021_parked.SM_NissanPatrol2021_parked':\n    'car',\n    '/Game/Carla/Static/Car/4Wheeled/ParkedVehicles/TeslaM3/SM_TeslaM3_parked.SM_TeslaM3_parked':\n    'car',\n    '/Game/Carla/Static/Car/4Wheeled/ParkedVehicles/VolkswagenT2/SM_VolkswagenT2_2021_Parked.SM_VolkswagenT2_2021_Parked':\n    'van',\n    'vehicle.ford.ambulance':\n    'van',\n    'vehicle.carlamotors.firetruck':\n    'truck',\n    'traffic.speed_limit.30':\n    'traffic_sign',\n    'traffic.speed_limit.40':\n    'traffic_sign',\n    'traffic.speed_limit.50':\n    'traffic_sign',\n    'traffic.speed_limit.60':\n    'traffic_sign',\n    'traffic.speed_limit.90':\n    'traffic_sign',\n    'traffic.speed_limit.120':\n    'traffic_sign',\n    'traffic.stop':\n    'traffic_sign',\n    'traffic.yield':\n    'traffic_sign',\n    'traffic.traffic_light':\n    'traffic_light',\n    'static.prop.warningconstruction':\n    'traffic_cone',\n    'static.prop.warningaccident':\n    'traffic_cone',\n    'static.prop.trafficwarning':\n    'traffic_cone',\n    'static.prop.constructioncone':\n    'traffic_cone',\n    'walker.pedestrian.0001':\n    'pedestrian',\n    'walker.pedestrian.0004':\n    'pedestrian',\n    'walker.pedestrian.0005':\n    'pedestrian',\n    'walker.pedestrian.0007':\n    'pedestrian',\n    'walker.pedestrian.0013':\n    'pedestrian',\n    'walker.pedestrian.0014':\n    'pedestrian',\n    'walker.pedestrian.0017':\n    'pedestrian',\n    'walker.pedestrian.0018':\n    'pedestrian',\n    'walker.pedestrian.0019':\n    'pedestrian',\n    'walker.pedestrian.0020':\n    'pedestrian',\n    'walker.pedestrian.0022':\n    'pedestrian',\n    'walker.pedestrian.0025':\n    'pedestrian',\n    'walker.pedestrian.0035':\n    'pedestrian',\n    'walker.pedestrian.0041':\n    'pedestrian',\n    'walker.pedestrian.0046':\n    'pedestrian',\n    'walker.pedestrian.0047':\n    'pedestrian',\n    'static.prop.dirtdebris01':\n    'others',\n    'static.prop.dirtdebris02':\n    'others'\n})\neval_cfg = dict(\n    dist_ths=[0.5, 1.0, 2.0, 4.0],\n    dist_th_tp=2.0,\n    min_recall=0.1,\n    min_precision=0.1,\n    mean_ap_weight=5,\n    class_names=[\n        'car', 'van', 'truck', 'bicycle', 'traffic_sign', 'traffic_cone',\n        'traffic_light', 'pedestrian'\n    ],\n    tp_metrics=['trans_err', 'scale_err', 'orient_err', 'vel_err'],\n    err_name_maping=dict(\n        trans_err='mATE',\n        scale_err='mASE',\n        orient_err='mAOE',\n        vel_err='mAVE',\n        attr_err='mAAE'),\n    class_range=dict(\n        car=(50, 50),\n        van=(50, 50),\n        truck=(50, 50),\n        bicycle=(40, 40),\n        traffic_sign=(30, 30),\n        traffic_cone=(30, 30),\n        traffic_light=(30, 30),\n        pedestrian=(40, 40)))\nnum_classes = 9\nmap_classes = [\n    'Broken', 'Solid', 'SolidSolid', 'Center', 'TrafficLight', 'StopSign'\n]\nmap_num_vec = 100\nmap_fixed_ptsnum_per_gt_line = 20\nmap_fixed_ptsnum_per_pred_line = 20\nmap_eval_use_same_gt_sample_num_flag = True\nmap_num_classes = 6\npast_frames = 2\nfuture_frames = 6\n_dim_ = 256\n_pos_dim_ = 128\n_ffn_dim_ = 512\n_num_levels_ = 1\nbev_h_ = 100\nbev_w_ = 100\nqueue_length = 3\ntotal_epochs = 6\nmodel = dict(\n    type='GenAD',\n    use_grid_mask=True,\n    video_test_mode=True,\n    pretrained=dict(img='ckpts/resnet50-19c8e357.pth'),\n    img_backbone=dict(\n        type='ResNet',\n        depth=50,\n        num_stages=4,\n        out_indices=(3, ),\n        frozen_stages=1,\n        norm_cfg=dict(type='BN', requires_grad=False),\n        norm_eval=True,\n        style='pytorch'),\n    img_neck=dict(\n        type='FPN',\n        in_channels=[2048],\n        out_channels=256,\n        start_level=0,\n        add_extra_convs='on_output',\n        num_outs=1,\n        relu_before_extra_convs=True),\n    pts_bbox_head=dict(\n        type='GenADHead',\n        map_thresh=0.5,\n        dis_thresh=0.2,\n        pe_normalization=True,\n        tot_epoch=6,\n        use_traj_lr_warmup=False,\n        query_thresh=0.0,\n        query_use_fix_pad=False,\n        ego_his_encoder=None,\n        ego_lcf_feat_idx=None,\n        valid_fut_ts=6,\n        ego_fut_mode=6,\n        agent_dim=300,\n        ego_agent_decoder=dict(\n            type='CustomTransformerDecoder',\n            num_layers=1,\n            return_intermediate=False,\n            transformerlayers=dict(\n                type='BaseTransformerLayer',\n                attn_cfgs=[\n                    dict(\n                        type='MultiheadAttention',\n                        embed_dims=256,\n                        num_heads=8,\n                        dropout=0.0)\n                ],\n                feedforward_channels=512,\n                ffn_dropout=0.0,\n                operation_order=('cross_attn', 'norm', 'ffn', 'norm'))),\n        ego_map_decoder=dict(\n            type='CustomTransformerDecoder',\n            num_layers=1,\n            return_intermediate=False,\n            transformerlayers=dict(\n                type='BaseTransformerLayer',\n                attn_cfgs=[\n                    dict(\n                        type='MultiheadAttention',\n                        embed_dims=256,\n                        num_heads=8,\n                        dropout=0.0)\n                ],\n                feedforward_channels=512,\n                ffn_dropout=0.0,\n                operation_order=('cross_attn', 'norm', 'ffn', 'norm'))),\n        motion_decoder=dict(\n            type='CustomTransformerDecoder',\n            num_layers=1,\n            return_intermediate=False,\n            transformerlayers=dict(\n                type='BaseTransformerLayer',\n                attn_cfgs=[\n                    dict(\n                        type='MultiheadAttention',\n                        embed_dims=256,\n                        num_heads=8,\n                        dropout=0.0)\n                ],\n                feedforward_channels=512,\n                ffn_dropout=0.0,\n                operation_order=('cross_attn', 'norm', 'ffn', 'norm'))),\n        motion_map_decoder=dict(\n            type='CustomTransformerDecoder',\n            num_layers=1,\n            return_intermediate=False,\n            transformerlayers=dict(\n                type='BaseTransformerLayer',\n                attn_cfgs=[\n                    dict(\n                        type='MultiheadAttention',\n                        embed_dims=256,\n                        num_heads=8,\n                        dropout=0.0)\n                ],\n                feedforward_channels=512,\n                ffn_dropout=0.0,\n                operation_order=('cross_attn', 'norm', 'ffn', 'norm'))),\n        use_pe=True,\n        bev_h=100,\n        bev_w=100,\n        num_query=300,\n        num_classes=9,\n        in_channels=256,\n        sync_cls_avg_factor=True,\n        with_box_refine=True,\n        as_two_stage=False,\n        map_num_vec=100,\n        map_num_classes=6,\n        map_num_pts_per_vec=20,\n        map_num_pts_per_gt_vec=20,\n        map_query_embed_type='instance_pts',\n        map_transform_method='minmax',\n        map_gt_shift_pts_pattern='v2',\n        map_dir_interval=1,\n        map_code_size=2,\n        map_code_weights=[1.0, 1.0, 1.0, 1.0],\n        transformer=dict(\n            type='VADPerceptionTransformer',\n            map_num_vec=100,\n            map_num_pts_per_vec=20,\n            rotate_prev_bev=True,\n            use_shift=True,\n            use_can_bus=True,\n            embed_dims=256,\n            encoder=dict(\n                type='BEVFormerEncoder',\n                num_layers=3,\n                pc_range=[-15.0, -30.0, -2.0, 15.0, 30.0, 2.0],\n                num_points_in_pillar=4,\n                return_intermediate=False,\n                transformerlayers=dict(\n                    type='BEVFormerLayer',\n                    attn_cfgs=[\n                        dict(\n                            type='TemporalSelfAttention',\n                            embed_dims=256,\n                            num_levels=1),\n                        dict(\n                            type='SpatialCrossAttention',\n                            pc_range=[-15.0, -30.0, -2.0, 15.0, 30.0, 2.0],\n                            deformable_attention=dict(\n                                type='MSDeformableAttention3D',\n                                embed_dims=256,\n                                num_points=8,\n                                num_levels=1),\n                            embed_dims=256)\n                    ],\n                    feedforward_channels=512,\n                    ffn_dropout=0.0,\n                    operation_order=('self_attn', 'norm', 'cross_attn', 'norm',\n                                     'ffn', 'norm'))),\n            decoder=dict(\n                type='DetectionTransformerDecoder',\n                num_layers=3,\n                return_intermediate=True,\n                transformerlayers=dict(\n                    type='DetrTransformerDecoderLayer',\n                    attn_cfgs=[\n                        dict(\n                            type='MultiheadAttention',\n                            embed_dims=256,\n                            num_heads=8,\n                            dropout=0.0),\n                        dict(\n                            type='CustomMSDeformableAttention',\n                            embed_dims=256,\n                            num_levels=1)\n                    ],\n                    feedforward_channels=512,\n                    ffn_dropout=0.0,\n                    operation_order=('self_attn', 'norm', 'cross_attn', 'norm',\n                                     'ffn', 'norm'))),\n            map_decoder=dict(\n                type='MapDetectionTransformerDecoder',\n                num_layers=3,\n                return_intermediate=True,\n                transformerlayers=dict(\n                    type='DetrTransformerDecoderLayer',\n                    attn_cfgs=[\n                        dict(\n                            type='MultiheadAttention',\n                            embed_dims=256,\n                            num_heads=8,\n                            dropout=0.0),\n                        dict(\n                            type='CustomMSDeformableAttention',\n                            embed_dims=256,\n                            num_levels=1)\n                    ],\n                    feedforward_channels=512,\n                    ffn_dropout=0.0,\n                    operation_order=('self_attn', 'norm', 'cross_attn', 'norm',\n                                     'ffn', 'norm')))),\n        bbox_coder=dict(\n            type='CustomNMSFreeCoder',\n            post_center_range=[-20, -35, -10.0, 20, 35, 10.0],\n            pc_range=[-15.0, -30.0, -2.0, 15.0, 30.0, 2.0],\n            max_num=100,\n            voxel_size=[0.15, 0.15, 4],\n            num_classes=9),\n        map_bbox_coder=dict(\n            type='MapNMSFreeCoder',\n            post_center_range=[-20, -35, -20, -35, 20, 35, 20, 35],\n            pc_range=[-15.0, -30.0, -2.0, 15.0, 30.0, 2.0],\n            max_num=50,\n            voxel_size=[0.15, 0.15, 4],\n            num_classes=6),\n        positional_encoding=dict(\n            type='LearnedPositionalEncoding',\n            num_feats=128,\n            row_num_embed=100,\n            col_num_embed=100),\n        loss_cls=dict(\n            type='FocalLoss',\n            use_sigmoid=True,\n            gamma=2.0,\n            alpha=0.25,\n            loss_weight=2.0),\n        loss_bbox=dict(type='L1Loss', loss_weight=0.25),\n        loss_traj=dict(type='L1Loss', loss_weight=0.2),\n        loss_traj_cls=dict(\n            type='FocalLoss',\n            use_sigmoid=True,\n            gamma=2.0,\n            alpha=0.25,\n            loss_weight=0.2),\n        loss_iou=dict(type='GIoULoss', loss_weight=0.0),\n        loss_map_cls=dict(\n            type='FocalLoss',\n            use_sigmoid=True,\n            gamma=2.0,\n            alpha=0.25,\n            loss_weight=2.0),\n        loss_map_bbox=dict(type='L1Loss', loss_weight=0.0),\n        loss_map_iou=dict(type='GIoULoss', loss_weight=0.0),\n        loss_map_pts=dict(type='PtsL1Loss', loss_weight=1.0),\n        loss_map_dir=dict(type='PtsDirCosLoss', loss_weight=0.005),\n        loss_plan_reg=dict(type='L1Loss', loss_weight=1.0),\n        loss_plan_bound=dict(\n            type='PlanMapBoundLoss', loss_weight=1.0, dis_thresh=1.0),\n        loss_plan_col=dict(type='PlanCollisionLoss', loss_weight=1.0),\n        loss_plan_dir=dict(type='PlanMapDirectionLoss', loss_weight=0.5),\n        loss_vae_gen=dict(type='ProbabilisticLoss', loss_weight=1.0)),\n    train_cfg=dict(\n        pts=dict(\n            grid_size=[512, 512, 1],\n            voxel_size=[0.15, 0.15, 4],\n            point_cloud_range=[-15.0, -30.0, -2.0, 15.0, 30.0, 2.0],\n            out_size_factor=4,\n            assigner=dict(\n                type='HungarianAssigner3D',\n                cls_cost=dict(type='FocalLossCost', weight=2.0),\n                reg_cost=dict(type='BBox3DL1Cost', weight=0.25),\n                iou_cost=dict(type='IoUCost', weight=0.0),\n                pc_range=[-15.0, -30.0, -2.0, 15.0, 30.0, 2.0]),\n            map_assigner=dict(\n                type='MapHungarianAssigner3D',\n                cls_cost=dict(type='FocalLossCost', weight=2.0),\n                reg_cost=dict(\n                    type='BBoxL1Cost', weight=0.0, box_format='xywh'),\n                iou_cost=dict(type='IoUCost', iou_mode='giou', weight=0.0),\n                pts_cost=dict(type='OrderedPtsL1Cost', weight=1.0),\n                pc_range=[-15.0, -30.0, -2.0, 15.0, 30.0, 2.0]))))\ninfo_root = 'data/infos'\nmap_root = 'data/bench2drive/maps'\nmap_file = 'data/infos/b2d_map_infos.pkl'\nann_file_train = 'data/infos/b2d_infos_train.pkl'\nann_file_val = 'data/infos/b2d_infos_val.pkl'\nann_file_test = 'data/infos/b2d_infos_val.pkl'\ninference_only_pipeline = [\n    dict(type='LoadMultiViewImageFromFiles', to_float32=True),\n    dict(\n        type='NormalizeMultiviewImage',\n        mean=[123.675, 116.28, 103.53],\n        std=[58.395, 57.12, 57.375],\n        to_rgb=True),\n    dict(type='PadMultiViewImage', size_divisor=32),\n    dict(\n        type='MultiScaleFlipAug3D',\n        img_scale=(1600, 900),\n        pts_scale_ratio=1,\n        flip=False,\n        transforms=[\n            dict(type='RandomScaleImageMultiViewImage', scales=[0.8]),\n            dict(type='PadMultiViewImage', size_divisor=32),\n            dict(\n                type='VADFormatBundle3D',\n                class_names=[\n                    'car', 'van', 'truck', 'bicycle', 'traffic_sign',\n                    'traffic_cone', 'traffic_light', 'pedestrian', 'others'\n                ],\n                with_label=False,\n                with_ego=True),\n            dict(type='CustomCollect3D', keys=['img', 'ego_fut_cmd'])\n        ])\n]\noptimizer = dict(\n    type='AdamW',\n    lr=0.0002,\n    paramwise_cfg=dict(custom_keys=dict(img_backbone=dict(lr_mult=0.1))),\n    weight_decay=0.01)\noptimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))\nlr_config = dict(\n    by_epoch=False,\n    policy='CosineAnnealing',\n    warmup='linear',\n    warmup_iters=500,\n    warmup_ratio=0.3333333333333333,\n    min_lr_ratio=0.001)\nrunner = dict(type='EpochBasedRunner', max_epochs=6)\nfind_unused_parameters = True\ncustom_hooks = [\n    dict(type='CustomSetEpochInfoHook'),\n    dict(\n        type='MLflowHook',\n        tracking_uri='https://dagshub.com/YSH-research/YSH-Bench2Drive.mlflow',\n        exp_name='GenAD-Bench2Drive',\n        log_interval=100,\n        log_model=True,\n        tags=dict(model='GenAD', dataset='Bench2Drive'))\n]\ngpu_ids = range(0, 1)\n", "CLASSES": ["car", "van", "truck", "bicycle", "traffic_sign", "traffic_cone", "traffic_light", "pedestrian", "others"], "PALETTE": null, "env_info": "MMCV: 0.0.1", "seed": 0, "exp_name": "GenAD_config_b2d.py", "epoch": 2, "iter": 469538, "time": "Thu Feb 12 09:40:00 2026", "hook_msgs": {"last_ckpt": "/home/Humble/Humble/EndtoEnd/Bench2DriveZoo/work_dirs/GenAD_config_b2d/epoch_1.pth"}}
+{"mode": "train", "epoch": 3, "iter": 50, "lr": 0.00015, "memory": 9562, "data_time": 0.18522, "loss_cls": 0.06763, "loss_bbox": 0.33598, "loss_traj": 0.37795, "loss_traj_cls": 0.03457, "loss_map_cls": 0.08361, "loss_map_bbox": 0.0, "loss_map_iou": 0.0, "loss_map_pts": 0.35584, "loss_map_dir": 0.0125, "loss_plan_reg": 0.05548, "loss_plan_bound": 0.0, "loss_plan_col": 0.03964, "loss_plan_dir": 0.00129, "d0.loss_cls": 0.11048, "d0.loss_bbox": 0.40059, "d1.loss_cls": 0.07206, "d1.loss_bbox": 0.34304, "d0.loss_map_cls": 0.10725, "d0.loss_map_bbox": 0.0, "d0.loss_map_iou": 0.0, "d0.loss_map_pts": 0.44727, "d0.loss_map_dir": 0.01566, "d1.loss_map_cls": 0.08936, "d1.loss_map_bbox": 0.0, "d1.loss_map_iou": 0.0, "d1.loss_map_pts": 0.37008, "d1.loss_map_dir": 0.01322, "loss_vae_gen": 1e-05, "loss": 3.33352, "grad_norm": 75.77571, "time": 0.67214}
+{"mode": "train", "epoch": 3, "iter": 100, "lr": 0.00015, "memory": 9562, "data_time": 0.03896, "loss_cls": 0.04944, "loss_bbox": 0.29215, "loss_traj": 0.47163, "loss_traj_cls": 0.03024, "loss_map_cls": 0.07525, "loss_map_bbox": 0.0, "loss_map_iou": 0.0, "loss_map_pts": 0.36258, "loss_map_dir": 0.01369, "loss_plan_reg": 0.04196, "loss_plan_bound": 0.0, "loss_plan_col": 0.04547, "loss_plan_dir": 0.00517, "d0.loss_cls": 0.09057, "d0.loss_bbox": 0.34719, "d1.loss_cls": 0.051, "d1.loss_bbox": 0.29328, "d0.loss_map_cls": 0.09273, "d0.loss_map_bbox": 0.0, "d0.loss_map_iou": 0.0, "d0.loss_map_pts": 0.45886, "d0.loss_map_dir": 0.01734, "d1.loss_map_cls": 0.07638, "d1.loss_map_bbox": 0.0, "d1.loss_map_iou": 0.0, "d1.loss_map_pts": 0.37779, "d1.loss_map_dir": 0.01437, "loss_vae_gen": 1e-05, "loss": 3.20709, "grad_norm": 47.1782, "time": 0.54354}
+{"mode": "train", "epoch": 3, "iter": 150, "lr": 0.00015, "memory": 9562, "data_time": 0.05514, "loss_cls": 0.06265, "loss_bbox": 0.33035, "loss_traj": 0.6509, "loss_traj_cls": 0.02888, "loss_map_cls": 0.09342, "loss_map_bbox": 0.0, "loss_map_iou": 0.0, "loss_map_pts": 0.40021, "loss_map_dir": 0.01597, "loss_plan_reg": 0.04514, "loss_plan_bound": 0.0, "loss_plan_col": 0.02267, "loss_plan_dir": 0.00679, "d0.loss_cls": 0.12421, "d0.loss_bbox": 0.38275, "d1.loss_cls": 0.07982, "d1.loss_bbox": 0.32443, "d0.loss_map_cls": 0.12501, "d0.loss_map_bbox": 0.0, "d0.loss_map_iou": 0.0, "d0.loss_map_pts": 0.5185, "d0.loss_map_dir": 0.01935, "d1.loss_map_cls": 0.09653, "d1.loss_map_bbox": 0.0, "d1.loss_map_iou": 0.0, "d1.loss_map_pts": 0.41387, "d1.loss_map_dir": 0.01664, "loss_vae_gen": 0.0, "loss": 3.75811, "grad_norm": 55.78294, "time": 0.52356}
+{"mode": "train", "epoch": 3, "iter": 200, "lr": 0.00015, "memory": 9562, "data_time": 0.06696, "loss_cls": 0.04281, "loss_bbox": 0.27124, "loss_traj": 0.57379, "loss_traj_cls": 0.0347, "loss_map_cls": 0.07074, "loss_map_bbox": 0.0, "loss_map_iou": 0.0, "loss_map_pts": 0.30735, "loss_map_dir": 0.01249, "loss_plan_reg": 0.0631, "loss_plan_bound": 0.0, "loss_plan_col": 0.02599, "loss_plan_dir": 0.0061, "d0.loss_cls": 0.08082, "d0.loss_bbox": 0.31986, "d1.loss_cls": 0.05215, "d1.loss_bbox": 0.27084, "d0.loss_map_cls": 0.10092, "d0.loss_map_bbox": 0.0, "d0.loss_map_iou": 0.0, "d0.loss_map_pts": 0.40319, "d0.loss_map_dir": 0.01624, "d1.loss_map_cls": 0.07538, "d1.loss_map_bbox": 0.0, "d1.loss_map_iou": 0.0, "d1.loss_map_pts": 0.32069, "d1.loss_map_dir": 0.01324, "loss_vae_gen": 1e-05, "loss": 3.06166, "grad_norm": 45.50508, "time": 0.53713}
+{"mode": "train", "epoch": 3, "iter": 250, "lr": 0.00015, "memory": 9562, "data_time": 0.03395, "loss_cls": 0.08073, "loss_bbox": 0.31743, "loss_traj": 0.50973, "loss_traj_cls": 0.03364, "loss_map_cls": 0.08942, "loss_map_bbox": 0.0, "loss_map_iou": 0.0, "loss_map_pts": 0.35201, "loss_map_dir": 0.01453, "loss_plan_reg": 0.04608, "loss_plan_bound": 0.00265, "loss_plan_col": 0.0, "loss_plan_dir": 0.00382, "d0.loss_cls": 0.11465, "d0.loss_bbox": 0.36684, "d1.loss_cls": 0.08358, "d1.loss_bbox": 0.32539, "d0.loss_map_cls": 0.10871, "d0.loss_map_bbox": 0.0, "d0.loss_map_iou": 0.0, "d0.loss_map_pts": 0.46576, "d0.loss_map_dir": 0.01774, "d1.loss_map_cls": 0.09238, "d1.loss_map_bbox": 0.0, "d1.loss_map_iou": 0.0, "d1.loss_map_pts": 0.36511, "d1.loss_map_dir": 0.01508, "loss_vae_gen": 0.0, "loss": 3.4053, "grad_norm": 60.41092, "time": 0.53668}

GenAD/with_hlc/latest.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:285dbdaa1a252985f680a6ae3c6d3e6dfc77ee2df7d149a28f7a4fbbc164615c
+size 826834590

GenAD/with_hlc/tf_logs/events.out.tfevents.1770264006.ysh-4090.2004455.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d1568489764cbbe650add86aa14ef2bb82c53cd50355d7623e8ee4a34922c989
+size 14632

GenAD/with_hlc/tf_logs/events.out.tfevents.1770272652.ysh-4090.2264287.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:eb7e43821d70a44a8870d428978b2d7784bf3821b0e455356e607f51ec6f02b3
+size 343312

GenAD/with_hlc/tf_logs/events.out.tfevents.1770548824.ysh-4090.110431.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fb0ae1ebfb5d3868dde83c58b96dcbc53c0b4374de476638cf03033c9e2f8c33
+size 469306

GenAD/with_hlc/tf_logs/events.out.tfevents.1770635673.ysh-4090.42668.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2f15f270bd24b82530472085ddb71a7a1030eb5dc18187a0c4ced0980cf68d3c
+size 21923250

GenAD/with_hlc/tf_logs/events.out.tfevents.1770970250.ysh-4090.83090.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a3327b753b7028f901f8e28892cfcf5090095ef8d4506f2b6e46a32917bc49d2
+size 9378