2021-11-08 09:43:58,630 INFO **********************Start logging********************** 2021-11-08 09:43:58,630 INFO CUDA_VISIBLE_DEVICES=1 2021-11-08 09:43:58,630 INFO total_batch_size: 1 2021-11-08 09:43:58,630 INFO cfg_file configs/stereo/kitti_models/liga.3d-and-bev.yaml 2021-11-08 09:43:58,630 INFO batch_size 1 2021-11-08 09:43:58,630 INFO epochs 60 2021-11-08 09:43:58,630 INFO workers 2 2021-11-08 09:43:58,630 INFO exp_name dev 2021-11-08 09:43:58,630 INFO fix_random_seed True 2021-11-08 09:43:58,630 INFO ckpt_save_interval 1 2021-11-08 09:43:58,630 INFO max_ckpt_save_num 30 2021-11-08 09:43:58,630 INFO merge_all_iters_to_one_epoch False 2021-11-08 09:43:58,630 INFO max_waiting_mins 0 2021-11-08 09:43:58,630 INFO save_to_file True 2021-11-08 09:43:58,630 INFO ckpt None 2021-11-08 09:43:58,630 INFO start_epoch 0 2021-11-08 09:43:58,630 INFO continue_train False 2021-11-08 09:43:58,630 INFO launcher pytorch 2021-11-08 09:43:58,630 INFO tcp_port 18858 2021-11-08 09:43:58,630 INFO sync_bn True 2021-11-08 09:43:58,630 INFO find_unused_parameters False 2021-11-08 09:43:58,630 INFO local_rank 0 2021-11-08 09:43:58,630 INFO set_cfgs None 2021-11-08 09:43:58,630 INFO trainval False 2021-11-08 09:43:58,631 INFO imitation 2d 2021-11-08 09:43:58,631 INFO cfg.ROOT_DIR: /home/chenyinan/Projects/LIGA-Stereo_DEV 2021-11-08 09:43:58,631 INFO cfg.LOCAL_RANK: 0 2021-11-08 09:43:58,631 INFO cfg.CLASS_NAMES: ['Car', 'Pedestrian', 'Cyclist'] 2021-11-08 09:43:58,631 INFO cfg.DATA_CONFIG = edict() 2021-11-08 09:43:58,631 INFO cfg.DATA_CONFIG.DATASET: StereoKittiDataset 2021-11-08 09:43:58,632 INFO cfg.DATA_CONFIG.DATA_PATH: ./data/kitti 2021-11-08 09:43:58,632 INFO cfg.DATA_CONFIG.FLIP: True 2021-11-08 09:43:58,632 INFO cfg.DATA_CONFIG.FORCE_FLIP: False 2021-11-08 09:43:58,632 INFO cfg.DATA_CONFIG.POINT_CLOUD_RANGE: [2, -30.4, -3, 59.6, 30.4, 1] 2021-11-08 09:43:58,632 INFO cfg.DATA_CONFIG.VOXEL_SIZE: [0.05, 0.05, 0.1] 2021-11-08 09:43:58,632 INFO cfg.DATA_CONFIG.STEREO_VOXEL_SIZE: [0.2, 0.2, 0.2] 2021-11-08 09:43:58,632 INFO cfg.DATA_CONFIG.DATA_SPLIT = edict() 2021-11-08 09:43:58,633 INFO cfg.DATA_CONFIG.DATA_SPLIT.train: train 2021-11-08 09:43:58,633 INFO cfg.DATA_CONFIG.DATA_SPLIT.test: val 2021-11-08 09:43:58,633 INFO cfg.DATA_CONFIG.INFO_PATH = edict() 2021-11-08 09:43:58,634 INFO cfg.DATA_CONFIG.INFO_PATH.train: ['kitti_infos_train.pkl'] 2021-11-08 09:43:58,634 INFO cfg.DATA_CONFIG.INFO_PATH.test: ['kitti_infos_val.pkl'] 2021-11-08 09:43:58,634 INFO cfg.DATA_CONFIG.USE_VAN: True 2021-11-08 09:43:58,634 INFO cfg.DATA_CONFIG.USE_PERSON_SITTING: True 2021-11-08 09:43:58,634 INFO cfg.DATA_CONFIG.FOV_POINTS_ONLY: True 2021-11-08 09:43:58,634 INFO cfg.DATA_CONFIG.BOXES_GT_IN_CAM2_VIEW: False 2021-11-08 09:43:58,635 INFO cfg.DATA_CONFIG.GENERATE_CORNER_HEATMAP: False 2021-11-08 09:43:58,635 INFO cfg.DATA_CONFIG.CAT_REFLECT_DIM: False 2021-11-08 09:43:58,635 INFO cfg.DATA_CONFIG.TRAIN_DATA_AUGMENTOR: [{'NAME': 'random_crop', 'MIN_REL_X': 0, 'MAX_REL_X': 0, 'MIN_REL_Y': 1.0, 'MAX_REL_Y': 1.0, 'MAX_CROP_H': 320, 'MAX_CROP_W': 1280}, {'NAME': 'filter_truncated', 'AREA_RATIO_THRESH': None, 'AREA_2D_RATIO_THRESH': None, 'GT_TRUNCATED_THRESH': 0.98}] 2021-11-08 09:43:58,635 INFO cfg.DATA_CONFIG.TEST_DATA_AUGMENTOR: [{'NAME': 'random_crop', 'MIN_REL_X': 0, 'MAX_REL_X': 0, 'MIN_REL_Y': 1.0, 'MAX_REL_Y': 1.0, 'MAX_CROP_H': 320, 'MAX_CROP_W': 1280}] 2021-11-08 09:43:58,635 INFO cfg.DATA_CONFIG.POINT_FEATURE_ENCODING = edict() 2021-11-08 09:43:58,636 INFO cfg.DATA_CONFIG.POINT_FEATURE_ENCODING.encoding_type: absolute_coordinates_encoding 2021-11-08 09:43:58,636 INFO cfg.DATA_CONFIG.POINT_FEATURE_ENCODING.used_feature_list: ['x', 'y', 'z'] 2021-11-08 09:43:58,636 INFO cfg.DATA_CONFIG.POINT_FEATURE_ENCODING.src_feature_list: ['x', 'y', 'z'] 2021-11-08 09:43:58,636 INFO cfg.DATA_CONFIG.DATA_PROCESSOR: [{'NAME': 'mask_points_and_boxes_outside_range', 'REMOVE_OUTSIDE_BOXES': True}, {'NAME': 'transform_points_to_voxels', 'VOXEL_SIZE': [0.05, 0.05, 0.1], 'MAX_POINTS_PER_VOXEL': 5, 'MAX_NUMBER_OF_VOXELS': {'train': 40000, 'test': 40000}}] 2021-11-08 09:43:58,636 INFO cfg.DATA_CONFIG._BASE_CONFIG_: ./configs/stereo/dataset_configs/kitti_dataset_fused.yaml 2021-11-08 09:43:58,636 INFO cfg.MODEL = edict() 2021-11-08 09:43:58,637 INFO cfg.MODEL.NAME: stereo_LIGA 2021-11-08 09:43:58,637 INFO cfg.MODEL.LIDAR_MODEL = edict() 2021-11-08 09:43:58,638 INFO cfg.MODEL.LIDAR_MODEL.NAME: SECONDNet 2021-11-08 09:43:58,638 INFO cfg.MODEL.LIDAR_MODEL.RETURN_BATCH_DICT: True 2021-11-08 09:43:58,638 INFO cfg.MODEL.LIDAR_MODEL.PRETRAINED_MODEL: ./ckpt/second_s4_hg.iouloss.ep78.backbone-no-final-bnrelu.input-only-xyz.default-lr-policy-with-wd-decay-78ep.pth 2021-11-08 09:43:58,638 INFO cfg.MODEL.LIDAR_MODEL.VFE = edict() 2021-11-08 09:43:58,638 INFO cfg.MODEL.LIDAR_MODEL.VFE.NAME: MeanVFE 2021-11-08 09:43:58,638 INFO cfg.MODEL.LIDAR_MODEL.BACKBONE_3D = edict() 2021-11-08 09:43:58,638 INFO cfg.MODEL.LIDAR_MODEL.BACKBONE_3D.NAME: VoxelBackBone4xNoFinalBnReLU 2021-11-08 09:43:58,638 INFO cfg.MODEL.LIDAR_MODEL.MAP_TO_BEV = edict() 2021-11-08 09:43:58,638 INFO cfg.MODEL.LIDAR_MODEL.MAP_TO_BEV.NAME: HeightCompression 2021-11-08 09:43:58,638 INFO cfg.MODEL.LIDAR_MODEL.MAP_TO_BEV.NUM_BEV_FEATURES: 160 2021-11-08 09:43:58,638 INFO cfg.MODEL.LIDAR_MODEL.BACKBONE_2D = edict() 2021-11-08 09:43:58,639 INFO cfg.MODEL.LIDAR_MODEL.BACKBONE_2D.NAME: HgBEVBackbone 2021-11-08 09:43:58,639 INFO cfg.MODEL.LIDAR_MODEL.BACKBONE_2D.num_channels: 64 2021-11-08 09:43:58,639 INFO cfg.MODEL.LIDAR_MODEL.BACKBONE_2D.GN: False 2021-11-08 09:43:58,639 INFO cfg.MODEL.BACKBONE_3D = edict() 2021-11-08 09:43:58,639 INFO cfg.MODEL.BACKBONE_3D.NAME: LigaBackbone 2021-11-08 09:43:58,639 INFO cfg.MODEL.BACKBONE_3D.maxdisp: 288 2021-11-08 09:43:58,639 INFO cfg.MODEL.BACKBONE_3D.downsample_disp: 4 2021-11-08 09:43:58,639 INFO cfg.MODEL.BACKBONE_3D.GN: True 2021-11-08 09:43:58,639 INFO cfg.MODEL.BACKBONE_3D.img_feature_attentionbydisp: True 2021-11-08 09:43:58,639 INFO cfg.MODEL.BACKBONE_3D.voxel_attentionbydisp: False 2021-11-08 09:43:58,639 INFO cfg.MODEL.BACKBONE_3D.cat_img_feature: True 2021-11-08 09:43:58,639 INFO cfg.MODEL.BACKBONE_3D.num_3dconvs: 1 2021-11-08 09:43:58,639 INFO cfg.MODEL.BACKBONE_3D.feature_backbone = edict() 2021-11-08 09:43:58,639 INFO cfg.MODEL.BACKBONE_3D.feature_backbone.type: ResNet 2021-11-08 09:43:58,639 INFO cfg.MODEL.BACKBONE_3D.feature_backbone.depth: 34 2021-11-08 09:43:58,639 INFO cfg.MODEL.BACKBONE_3D.feature_backbone.num_stages: 4 2021-11-08 09:43:58,639 INFO cfg.MODEL.BACKBONE_3D.feature_backbone.out_indices: [0, 1, 2, 3] 2021-11-08 09:43:58,639 INFO cfg.MODEL.BACKBONE_3D.feature_backbone.frozen_stages: -1 2021-11-08 09:43:58,639 INFO cfg.MODEL.BACKBONE_3D.feature_backbone.norm_cfg = edict() 2021-11-08 09:43:58,639 INFO cfg.MODEL.BACKBONE_3D.feature_backbone.norm_cfg.type: BN 2021-11-08 09:43:58,639 INFO cfg.MODEL.BACKBONE_3D.feature_backbone.norm_cfg.requires_grad: True 2021-11-08 09:43:58,639 INFO cfg.MODEL.BACKBONE_3D.feature_backbone.norm_eval: False 2021-11-08 09:43:58,639 INFO cfg.MODEL.BACKBONE_3D.feature_backbone.style: pytorch 2021-11-08 09:43:58,639 INFO cfg.MODEL.BACKBONE_3D.feature_backbone.with_max_pool: False 2021-11-08 09:43:58,639 INFO cfg.MODEL.BACKBONE_3D.feature_backbone.deep_stem: False 2021-11-08 09:43:58,639 INFO cfg.MODEL.BACKBONE_3D.feature_backbone.block_with_final_relu: False 2021-11-08 09:43:58,639 INFO cfg.MODEL.BACKBONE_3D.feature_backbone.base_channels: 64 2021-11-08 09:43:58,639 INFO cfg.MODEL.BACKBONE_3D.feature_backbone.strides: [1, 2, 1, 1] 2021-11-08 09:43:58,639 INFO cfg.MODEL.BACKBONE_3D.feature_backbone.dilations: [1, 1, 2, 4] 2021-11-08 09:43:58,639 INFO cfg.MODEL.BACKBONE_3D.feature_backbone.num_channels_factor: [1, 2, 2, 2] 2021-11-08 09:43:58,639 INFO cfg.MODEL.BACKBONE_3D.feature_backbone_pretrained: torchvision://resnet34 2021-11-08 09:43:58,640 INFO cfg.MODEL.BACKBONE_3D.feature_neck = edict() 2021-11-08 09:43:58,640 INFO cfg.MODEL.BACKBONE_3D.feature_neck.GN: True 2021-11-08 09:43:58,640 INFO cfg.MODEL.BACKBONE_3D.feature_neck.in_dims: [3, 64, 128, 128, 128] 2021-11-08 09:43:58,640 INFO cfg.MODEL.BACKBONE_3D.feature_neck.start_level: 2 2021-11-08 09:43:58,640 INFO cfg.MODEL.BACKBONE_3D.feature_neck.stereo_dim: [32, 32] 2021-11-08 09:43:58,640 INFO cfg.MODEL.BACKBONE_3D.feature_neck.with_upconv: True 2021-11-08 09:43:58,640 INFO cfg.MODEL.BACKBONE_3D.feature_neck.cat_img_feature: True 2021-11-08 09:43:58,640 INFO cfg.MODEL.BACKBONE_3D.feature_neck.sem_dim: [128, 32] 2021-11-08 09:43:58,640 INFO cfg.MODEL.BACKBONE_3D.sem_neck = edict() 2021-11-08 09:43:58,640 INFO cfg.MODEL.BACKBONE_3D.sem_neck.type: FPN 2021-11-08 09:43:58,640 INFO cfg.MODEL.BACKBONE_3D.sem_neck.in_channels: [32] 2021-11-08 09:43:58,640 INFO cfg.MODEL.BACKBONE_3D.sem_neck.out_channels: 64 2021-11-08 09:43:58,640 INFO cfg.MODEL.BACKBONE_3D.sem_neck.start_level: 0 2021-11-08 09:43:58,640 INFO cfg.MODEL.BACKBONE_3D.sem_neck.add_extra_convs: on_output 2021-11-08 09:43:58,640 INFO cfg.MODEL.BACKBONE_3D.sem_neck.num_outs: 5 2021-11-08 09:43:58,640 INFO cfg.MODEL.BACKBONE_3D.cost_volume: [{'type': 'concat', 'downsample': 4}] 2021-11-08 09:43:58,640 INFO cfg.MODEL.BACKBONE_3D.cv_dim: 32 2021-11-08 09:43:58,640 INFO cfg.MODEL.BACKBONE_3D.rpn3d_dim: 32 2021-11-08 09:43:58,640 INFO cfg.MODEL.BACKBONE_3D.downsampled_depth_offset: 0.5 2021-11-08 09:43:58,640 INFO cfg.MODEL.BACKBONE_3D.use_stereo_out_type: feature 2021-11-08 09:43:58,640 INFO cfg.MODEL.BACKBONE_3D.num_hg: 1 2021-11-08 09:43:58,640 INFO cfg.MODEL.DENSE_HEAD_2D = edict() 2021-11-08 09:43:58,640 INFO cfg.MODEL.DENSE_HEAD_2D.NAME: MMDet2DHead 2021-11-08 09:43:58,640 INFO cfg.MODEL.DENSE_HEAD_2D.use_3d_center: True 2021-11-08 09:43:58,640 INFO cfg.MODEL.DENSE_HEAD_2D.cfg = edict() 2021-11-08 09:43:58,640 INFO cfg.MODEL.DENSE_HEAD_2D.cfg.type: ATSSAdvHead 2021-11-08 09:43:58,640 INFO cfg.MODEL.DENSE_HEAD_2D.cfg.reg_class_agnostic: False 2021-11-08 09:43:58,640 INFO cfg.MODEL.DENSE_HEAD_2D.cfg.seperate_extra_reg_branch: False 2021-11-08 09:43:58,640 INFO cfg.MODEL.DENSE_HEAD_2D.cfg.num_classes: 3 2021-11-08 09:43:58,641 INFO cfg.MODEL.DENSE_HEAD_2D.cfg.in_channels: 64 2021-11-08 09:43:58,641 INFO cfg.MODEL.DENSE_HEAD_2D.cfg.stacked_convs: 4 2021-11-08 09:43:58,641 INFO cfg.MODEL.DENSE_HEAD_2D.cfg.feat_channels: 64 2021-11-08 09:43:58,641 INFO cfg.MODEL.DENSE_HEAD_2D.cfg.anchor_generator = edict() 2021-11-08 09:43:58,641 INFO cfg.MODEL.DENSE_HEAD_2D.cfg.anchor_generator.type: AnchorGenerator 2021-11-08 09:43:58,641 INFO cfg.MODEL.DENSE_HEAD_2D.cfg.anchor_generator.ratios: [1.0] 2021-11-08 09:43:58,641 INFO cfg.MODEL.DENSE_HEAD_2D.cfg.anchor_generator.octave_base_scale: 16 2021-11-08 09:43:58,641 INFO cfg.MODEL.DENSE_HEAD_2D.cfg.anchor_generator.scales_per_octave: 1 2021-11-08 09:43:58,641 INFO cfg.MODEL.DENSE_HEAD_2D.cfg.anchor_generator.strides: [4, 8, 16, 32, 64] 2021-11-08 09:43:58,641 INFO cfg.MODEL.DENSE_HEAD_2D.cfg.num_extra_reg_channel: 0 2021-11-08 09:43:58,641 INFO cfg.MODEL.DENSE_HEAD_2D.cfg.bbox_coder = edict() 2021-11-08 09:43:58,644 INFO cfg.MODEL.DENSE_HEAD_2D.cfg.bbox_coder.type: DeltaXYWHBBoxCoder 2021-11-08 09:43:58,644 INFO cfg.MODEL.DENSE_HEAD_2D.cfg.bbox_coder.target_means: [0.0, 0.0, 0.0, 0.0] 2021-11-08 09:43:58,646 INFO cfg.MODEL.DENSE_HEAD_2D.cfg.bbox_coder.target_stds: [0.1, 0.1, 0.2, 0.2] 2021-11-08 09:43:58,647 INFO cfg.MODEL.DENSE_HEAD_2D.cfg.loss_cls = edict() 2021-11-08 09:43:58,647 INFO cfg.MODEL.DENSE_HEAD_2D.cfg.loss_cls.type: FocalLoss 2021-11-08 09:43:58,647 INFO cfg.MODEL.DENSE_HEAD_2D.cfg.loss_cls.use_sigmoid: True 2021-11-08 09:43:58,647 INFO cfg.MODEL.DENSE_HEAD_2D.cfg.loss_cls.gamma: 2.0 2021-11-08 09:43:58,647 INFO cfg.MODEL.DENSE_HEAD_2D.cfg.loss_cls.alpha: 0.25 2021-11-08 09:43:58,647 INFO cfg.MODEL.DENSE_HEAD_2D.cfg.loss_cls.loss_weight: 1.0 2021-11-08 09:43:58,648 INFO cfg.MODEL.DENSE_HEAD_2D.cfg.loss_bbox = edict() 2021-11-08 09:43:58,648 INFO cfg.MODEL.DENSE_HEAD_2D.cfg.loss_bbox.type: GIoULoss 2021-11-08 09:43:58,648 INFO cfg.MODEL.DENSE_HEAD_2D.cfg.loss_bbox.loss_weight: 2.0 2021-11-08 09:43:58,648 INFO cfg.MODEL.DENSE_HEAD_2D.cfg.loss_centerness = edict() 2021-11-08 09:43:58,649 INFO cfg.MODEL.DENSE_HEAD_2D.cfg.loss_centerness.type: CrossEntropyLoss 2021-11-08 09:43:58,649 INFO cfg.MODEL.DENSE_HEAD_2D.cfg.loss_centerness.use_sigmoid: True 2021-11-08 09:43:58,649 INFO cfg.MODEL.DENSE_HEAD_2D.cfg.loss_centerness.loss_weight: 1.0 2021-11-08 09:43:58,649 INFO cfg.MODEL.DENSE_HEAD_2D.cfg.train_cfg = edict() 2021-11-08 09:43:58,649 INFO cfg.MODEL.DENSE_HEAD_2D.cfg.train_cfg.assigner = edict() 2021-11-08 09:43:58,650 INFO cfg.MODEL.DENSE_HEAD_2D.cfg.train_cfg.assigner.type: ATSS3DCenterAssigner 2021-11-08 09:43:58,650 INFO cfg.MODEL.DENSE_HEAD_2D.cfg.train_cfg.assigner.topk: 9 2021-11-08 09:43:58,650 INFO cfg.MODEL.DENSE_HEAD_2D.cfg.train_cfg.allowed_border: -1 2021-11-08 09:43:58,650 INFO cfg.MODEL.DENSE_HEAD_2D.cfg.train_cfg.pos_weight: -1 2021-11-08 09:43:58,650 INFO cfg.MODEL.DENSE_HEAD_2D.cfg.train_cfg.append_3d_centers: True 2021-11-08 09:43:58,650 INFO cfg.MODEL.DENSE_HEAD_2D.cfg.train_cfg.debug: False 2021-11-08 09:43:58,650 INFO cfg.MODEL.DENSE_HEAD_2D.cfg.test_cfg = edict() 2021-11-08 09:43:58,650 INFO cfg.MODEL.DENSE_HEAD_2D.cfg.test_cfg.nms_pre: 1000 2021-11-08 09:43:58,651 INFO cfg.MODEL.DENSE_HEAD_2D.cfg.test_cfg.min_bbox_size: 0 2021-11-08 09:43:58,651 INFO cfg.MODEL.DENSE_HEAD_2D.cfg.test_cfg.score_thr: 0.05 2021-11-08 09:43:58,651 INFO cfg.MODEL.DENSE_HEAD_2D.cfg.test_cfg.nms = edict() 2021-11-08 09:43:58,651 INFO cfg.MODEL.DENSE_HEAD_2D.cfg.test_cfg.nms.type: nms 2021-11-08 09:43:58,651 INFO cfg.MODEL.DENSE_HEAD_2D.cfg.test_cfg.nms.iou_threshold: 0.6 2021-11-08 09:43:58,651 INFO cfg.MODEL.DENSE_HEAD_2D.cfg.test_cfg.max_per_img: 100 2021-11-08 09:43:58,651 INFO cfg.MODEL.MAP_TO_BEV = edict() 2021-11-08 09:43:58,652 INFO cfg.MODEL.MAP_TO_BEV.NAME: HeightCompression 2021-11-08 09:43:58,652 INFO cfg.MODEL.MAP_TO_BEV.NUM_BEV_FEATURES: 160 2021-11-08 09:43:58,652 INFO cfg.MODEL.MAP_TO_BEV.SPARSE_INPUT: False 2021-11-08 09:43:58,652 INFO cfg.MODEL.BACKBONE_2D = edict() 2021-11-08 09:43:58,652 INFO cfg.MODEL.BACKBONE_2D.NAME: HgBEVBackbone 2021-11-08 09:43:58,652 INFO cfg.MODEL.BACKBONE_2D.num_channels: 64 2021-11-08 09:43:58,652 INFO cfg.MODEL.BACKBONE_2D.GN: True 2021-11-08 09:43:58,652 INFO cfg.MODEL.DENSE_HEAD = edict() 2021-11-08 09:43:58,653 INFO cfg.MODEL.DENSE_HEAD.NAME: DetHead 2021-11-08 09:43:58,653 INFO cfg.MODEL.DENSE_HEAD.NUM_CONVS: 2 2021-11-08 09:43:58,653 INFO cfg.MODEL.DENSE_HEAD.GN: True 2021-11-08 09:43:58,653 INFO cfg.MODEL.DENSE_HEAD.CLASS_AGNOSTIC: False 2021-11-08 09:43:58,653 INFO cfg.MODEL.DENSE_HEAD.USE_DIRECTION_CLASSIFIER: True 2021-11-08 09:43:58,654 INFO cfg.MODEL.DENSE_HEAD.DIR_OFFSET: 0.78539 2021-11-08 09:43:58,654 INFO cfg.MODEL.DENSE_HEAD.DIR_LIMIT_OFFSET: 0.0 2021-11-08 09:43:58,654 INFO cfg.MODEL.DENSE_HEAD.NUM_DIR_BINS: 2 2021-11-08 09:43:58,654 INFO cfg.MODEL.DENSE_HEAD.CLAMP_VALUE: 10.0 2021-11-08 09:43:58,655 INFO cfg.MODEL.DENSE_HEAD.xyz_for_angles: True 2021-11-08 09:43:58,655 INFO cfg.MODEL.DENSE_HEAD.hwl_for_angles: True 2021-11-08 09:43:58,655 INFO cfg.MODEL.DENSE_HEAD.do_feature_imitation: True 2021-11-08 09:43:58,655 INFO cfg.MODEL.DENSE_HEAD.imitation_cfg: [{'lidar_feature_layer': 'spatial_features_2d', 'stereo_feature_layer': 'spatial_features_2d', 'normalize': 'cw_scale', 'layer': 'conv2d', 'channel': 64, 'ksize': 1, 'use_relu': False, 'mode': 'inbox'}, {'lidar_feature_layer': 'volume_features', 'stereo_feature_layer': 'volume_features', 'normalize': 'cw_scale', 'layer': 'conv3d', 'channel': 32, 'ksize': 1, 'use_relu': False, 'mode': 'inbox'}] 2021-11-08 09:43:58,656 INFO cfg.MODEL.DENSE_HEAD.ANCHOR_GENERATOR_CONFIG: [{'class_name': 'Car', 'anchor_sizes': [[3.9, 1.6, 1.56]], 'anchor_rotations': [0, 1.57], 'anchor_bottom_heights': [-1.78], 'align_center': False, 'feature_map_stride': 1, 'matched_threshold': 0.6, 'unmatched_threshold': 0.45}, {'class_name': 'Pedestrian', 'anchor_sizes': [[0.8, 0.6, 1.73]], 'anchor_rotations': [0, 1.57], 'anchor_bottom_heights': [-0.6], 'align_center': False, 'feature_map_stride': 1, 'matched_threshold': 0.5, 'unmatched_threshold': 0.35}, {'class_name': 'Cyclist', 'anchor_sizes': [[1.76, 0.6, 1.73]], 'anchor_rotations': [0, 1.57], 'anchor_bottom_heights': [-0.6], 'align_center': False, 'feature_map_stride': 1, 'matched_threshold': 0.5, 'unmatched_threshold': 0.35}] 2021-11-08 09:43:58,656 INFO cfg.MODEL.DENSE_HEAD.TARGET_ASSIGNER_CONFIG = edict() 2021-11-08 09:43:58,656 INFO cfg.MODEL.DENSE_HEAD.TARGET_ASSIGNER_CONFIG.NAME: AxisAlignedTargetAssigner 2021-11-08 09:43:58,656 INFO cfg.MODEL.DENSE_HEAD.TARGET_ASSIGNER_CONFIG.POS_FRACTION: -1.0 2021-11-08 09:43:58,656 INFO cfg.MODEL.DENSE_HEAD.TARGET_ASSIGNER_CONFIG.SAMPLE_SIZE: 512 2021-11-08 09:43:58,656 INFO cfg.MODEL.DENSE_HEAD.TARGET_ASSIGNER_CONFIG.NORM_BY_NUM_EXAMPLES: False 2021-11-08 09:43:58,656 INFO cfg.MODEL.DENSE_HEAD.TARGET_ASSIGNER_CONFIG.MATCH_HEIGHT: False 2021-11-08 09:43:58,657 INFO cfg.MODEL.DENSE_HEAD.TARGET_ASSIGNER_CONFIG.BOX_CODER: ResidualCoder 2021-11-08 09:43:58,657 INFO cfg.MODEL.DENSE_HEAD.TARGET_ASSIGNER_CONFIG.BOX_CODER_CONFIG = edict() 2021-11-08 09:43:58,657 INFO cfg.MODEL.DENSE_HEAD.TARGET_ASSIGNER_CONFIG.BOX_CODER_CONFIG.div_by_diagonal: True 2021-11-08 09:43:58,657 INFO cfg.MODEL.DENSE_HEAD.TARGET_ASSIGNER_CONFIG.BOX_CODER_CONFIG.use_corners: False 2021-11-08 09:43:58,657 INFO cfg.MODEL.DENSE_HEAD.TARGET_ASSIGNER_CONFIG.BOX_CODER_CONFIG.use_tanh: False 2021-11-08 09:43:58,658 INFO cfg.MODEL.DENSE_HEAD.LOSS_CONFIG = edict() 2021-11-08 09:43:58,658 INFO cfg.MODEL.DENSE_HEAD.LOSS_CONFIG.REG_LOSS_TYPE: WeightedSmoothL1Loss 2021-11-08 09:43:58,658 INFO cfg.MODEL.DENSE_HEAD.LOSS_CONFIG.IOU_LOSS_TYPE: IOU3dLoss 2021-11-08 09:43:58,658 INFO cfg.MODEL.DENSE_HEAD.LOSS_CONFIG.IMITATION_LOSS_TYPE: WeightedL2WithSigmaLoss 2021-11-08 09:43:58,658 INFO cfg.MODEL.DENSE_HEAD.LOSS_CONFIG.LOSS_WEIGHTS = edict() 2021-11-08 09:43:58,658 INFO cfg.MODEL.DENSE_HEAD.LOSS_CONFIG.LOSS_WEIGHTS.cls_weight: 1.0 2021-11-08 09:43:58,658 INFO cfg.MODEL.DENSE_HEAD.LOSS_CONFIG.LOSS_WEIGHTS.loc_weight: 0.5 2021-11-08 09:43:58,658 INFO cfg.MODEL.DENSE_HEAD.LOSS_CONFIG.LOSS_WEIGHTS.dir_weight: 0.2 2021-11-08 09:43:58,658 INFO cfg.MODEL.DENSE_HEAD.LOSS_CONFIG.LOSS_WEIGHTS.iou_weight: 1.0 2021-11-08 09:43:58,658 INFO cfg.MODEL.DENSE_HEAD.LOSS_CONFIG.LOSS_WEIGHTS.imitation_weight: 1.0 2021-11-08 09:43:58,658 INFO cfg.MODEL.DENSE_HEAD.LOSS_CONFIG.LOSS_WEIGHTS.code_weights: [1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0] 2021-11-08 09:43:58,658 INFO cfg.MODEL.DEPTH_LOSS_HEAD = edict() 2021-11-08 09:43:58,658 INFO cfg.MODEL.DEPTH_LOSS_HEAD.LOSS_TYPE = edict() 2021-11-08 09:43:58,658 INFO cfg.MODEL.DEPTH_LOSS_HEAD.LOSS_TYPE.ce: 1.0 2021-11-08 09:43:58,658 INFO cfg.MODEL.DEPTH_LOSS_HEAD.WEIGHTS: [1.0] 2021-11-08 09:43:58,658 INFO cfg.MODEL.POST_PROCESSING = edict() 2021-11-08 09:43:58,658 INFO cfg.MODEL.POST_PROCESSING.RECALL_THRESH_LIST: [0.3, 0.5, 0.7] 2021-11-08 09:43:58,658 INFO cfg.MODEL.POST_PROCESSING.SCORE_THRESH: 0.1 2021-11-08 09:43:58,658 INFO cfg.MODEL.POST_PROCESSING.OUTPUT_RAW_SCORE: False 2021-11-08 09:43:58,658 INFO cfg.MODEL.POST_PROCESSING.EVAL_METRIC: kitti 2021-11-08 09:43:58,658 INFO cfg.MODEL.POST_PROCESSING.NMS_CONFIG = edict() 2021-11-08 09:43:58,658 INFO cfg.MODEL.POST_PROCESSING.NMS_CONFIG.MULTI_CLASSES_NMS: True 2021-11-08 09:43:58,658 INFO cfg.MODEL.POST_PROCESSING.NMS_CONFIG.NMS_TYPE: nms_gpu 2021-11-08 09:43:58,658 INFO cfg.MODEL.POST_PROCESSING.NMS_CONFIG.NMS_THRESH: 0.25 2021-11-08 09:43:58,658 INFO cfg.MODEL.POST_PROCESSING.NMS_CONFIG.NMS_PRE_MAXSIZE: 4096 2021-11-08 09:43:58,658 INFO cfg.MODEL.POST_PROCESSING.NMS_CONFIG.NMS_POST_MAXSIZE: 500 2021-11-08 09:43:58,658 INFO cfg.OPTIMIZATION = edict() 2021-11-08 09:43:58,658 INFO cfg.OPTIMIZATION.BATCH_SIZE_PER_GPU: 1 2021-11-08 09:43:58,658 INFO cfg.OPTIMIZATION.NUM_EPOCHS: 60 2021-11-08 09:43:58,658 INFO cfg.OPTIMIZATION.OPTIMIZER: adamw 2021-11-08 09:43:58,659 INFO cfg.OPTIMIZATION.LR: 0.001 2021-11-08 09:43:58,659 INFO cfg.OPTIMIZATION.WEIGHT_DECAY: 0.0001 2021-11-08 09:43:58,659 INFO cfg.OPTIMIZATION.MOMENTUM: 0.9 2021-11-08 09:43:58,659 INFO cfg.OPTIMIZATION.DIV_FACTOR: 10 2021-11-08 09:43:58,659 INFO cfg.OPTIMIZATION.DECAY_STEP_LIST: [50] 2021-11-08 09:43:58,659 INFO cfg.OPTIMIZATION.LR_DECAY: 0.1 2021-11-08 09:43:58,659 INFO cfg.OPTIMIZATION.LR_CLIP: 1e-07 2021-11-08 09:43:58,659 INFO cfg.OPTIMIZATION.LR_WARMUP: True 2021-11-08 09:43:58,659 INFO cfg.OPTIMIZATION.WARMUP_EPOCH: 1 2021-11-08 09:43:58,659 INFO cfg.OPTIMIZATION.GRAD_NORM_CLIP: 10 2021-11-08 09:43:58,659 INFO cfg.TAG: liga.3d-and-bev 2021-11-08 09:43:58,659 INFO cfg.EXP_GROUP_PATH: stereo_kitti_models 2021-11-08 09:43:58,855 INFO boxes_gt_in_cam2_view False 2021-11-08 09:43:58,869 INFO Loading KITTI dataset 2021-11-08 09:43:59,059 INFO Total samples for KITTI dataset: 3712 2021-11-08 09:43:59,075 INFO ==> Loading parameters from checkpoint ./ckpt/second_s4_hg.iouloss.ep78.backbone-no-final-bnrelu.input-only-xyz.default-lr-policy-with-wd-decay-78ep.pth to CPU 2021-11-08 09:43:59,098 INFO ==> Checkpoint trained from version: liga+0.1.0+7aa7b92+py60b444b 2021-11-08 09:43:59,160 INFO Not Loaded weight dense_head.rpn3d_cls_convs.0.0.0.weight: torch.Size([64, 64, 3, 3]) 2021-11-08 09:43:59,160 INFO Not Loaded weight dense_head.rpn3d_cls_convs.0.0.1.weight: torch.Size([64]) 2021-11-08 09:43:59,161 INFO Not Loaded weight dense_head.rpn3d_cls_convs.0.0.1.bias: torch.Size([64]) 2021-11-08 09:43:59,161 INFO Not Loaded weight dense_head.rpn3d_cls_convs.0.0.1.running_mean: torch.Size([64]) 2021-11-08 09:43:59,161 INFO Not Loaded weight dense_head.rpn3d_cls_convs.0.0.1.running_var: torch.Size([64]) 2021-11-08 09:43:59,161 INFO Not Loaded weight dense_head.rpn3d_cls_convs.0.0.1.num_batches_tracked: torch.Size([]) 2021-11-08 09:43:59,161 INFO Not Loaded weight dense_head.rpn3d_cls_convs.1.0.0.weight: torch.Size([64, 64, 3, 3]) 2021-11-08 09:43:59,161 INFO Not Loaded weight dense_head.rpn3d_cls_convs.1.0.1.weight: torch.Size([64]) 2021-11-08 09:43:59,161 INFO Not Loaded weight dense_head.rpn3d_cls_convs.1.0.1.bias: torch.Size([64]) 2021-11-08 09:43:59,162 INFO Not Loaded weight dense_head.rpn3d_cls_convs.1.0.1.running_mean: torch.Size([64]) 2021-11-08 09:43:59,162 INFO Not Loaded weight dense_head.rpn3d_cls_convs.1.0.1.running_var: torch.Size([64]) 2021-11-08 09:43:59,162 INFO Not Loaded weight dense_head.rpn3d_cls_convs.1.0.1.num_batches_tracked: torch.Size([]) 2021-11-08 09:43:59,162 INFO Not Loaded weight dense_head.rpn3d_bbox_convs.0.0.0.weight: torch.Size([64, 64, 3, 3]) 2021-11-08 09:43:59,162 INFO Not Loaded weight dense_head.rpn3d_bbox_convs.0.0.1.weight: torch.Size([64]) 2021-11-08 09:43:59,162 INFO Not Loaded weight dense_head.rpn3d_bbox_convs.0.0.1.bias: torch.Size([64]) 2021-11-08 09:43:59,162 INFO Not Loaded weight dense_head.rpn3d_bbox_convs.0.0.1.running_mean: torch.Size([64]) 2021-11-08 09:43:59,162 INFO Not Loaded weight dense_head.rpn3d_bbox_convs.0.0.1.running_var: torch.Size([64]) 2021-11-08 09:43:59,163 INFO Not Loaded weight dense_head.rpn3d_bbox_convs.0.0.1.num_batches_tracked: torch.Size([]) 2021-11-08 09:43:59,163 INFO Not Loaded weight dense_head.rpn3d_bbox_convs.1.0.0.weight: torch.Size([64, 64, 3, 3]) 2021-11-08 09:43:59,163 INFO Not Loaded weight dense_head.rpn3d_bbox_convs.1.0.1.weight: torch.Size([64]) 2021-11-08 09:43:59,163 INFO Not Loaded weight dense_head.rpn3d_bbox_convs.1.0.1.bias: torch.Size([64]) 2021-11-08 09:43:59,163 INFO Not Loaded weight dense_head.rpn3d_bbox_convs.1.0.1.running_mean: torch.Size([64]) 2021-11-08 09:43:59,163 INFO Not Loaded weight dense_head.rpn3d_bbox_convs.1.0.1.running_var: torch.Size([64]) 2021-11-08 09:43:59,163 INFO Not Loaded weight dense_head.rpn3d_bbox_convs.1.0.1.num_batches_tracked: torch.Size([]) 2021-11-08 09:43:59,163 INFO Not Loaded weight dense_head.conv_cls.weight: torch.Size([18, 64, 3, 3]) 2021-11-08 09:43:59,164 INFO Not Loaded weight dense_head.conv_cls.bias: torch.Size([18]) 2021-11-08 09:43:59,164 INFO Not Loaded weight dense_head.conv_box.weight: torch.Size([42, 64, 3, 3]) 2021-11-08 09:43:59,164 INFO Not Loaded weight dense_head.conv_box.bias: torch.Size([42]) 2021-11-08 09:43:59,164 INFO Not Loaded weight dense_head.conv_dir_cls.weight: torch.Size([12, 64, 1, 1]) 2021-11-08 09:43:59,164 INFO Not Loaded weight dense_head.conv_dir_cls.bias: torch.Size([12]) 2021-11-08 09:43:59,165 INFO ==> Done (loaded 110/110) 2021-11-08 09:43:59,545 INFO DistributedDataParallel( (module): LIGA( (lidar_model): SECONDNet( (vfe): MeanVFE() (backbone_3d): VoxelBackBone4xNoFinalBnReLU( (conv_input): SparseSequential( (0): SubMConv3d() (1): SyncBatchNorm(16, eps=0.001, momentum=0.01, affine=True, track_running_stats=True) (2): ReLU() ) (conv1): SparseSequential( (0): SparseSequential( (0): SubMConv3d() (1): SyncBatchNorm(16, eps=0.001, momentum=0.01, affine=True, track_running_stats=True) (2): ReLU() ) ) (conv2): SparseSequential( (0): SparseSequential( (0): SparseConv3d() (1): SyncBatchNorm(32, eps=0.001, momentum=0.01, affine=True, track_running_stats=True) (2): ReLU() ) (1): SparseSequential( (0): SubMConv3d() (1): SyncBatchNorm(32, eps=0.001, momentum=0.01, affine=True, track_running_stats=True) (2): ReLU() ) (2): SparseSequential( (0): SubMConv3d() (1): SyncBatchNorm(32, eps=0.001, momentum=0.01, affine=True, track_running_stats=True) (2): ReLU() ) ) (conv3): SparseSequential( (0): SparseSequential( (0): SparseConv3d() (1): SyncBatchNorm(64, eps=0.001, momentum=0.01, affine=True, track_running_stats=True) (2): ReLU() ) (1): SparseSequential( (0): SubMConv3d() (1): SyncBatchNorm(64, eps=0.001, momentum=0.01, affine=True, track_running_stats=True) (2): ReLU() ) (2): SparseSequential( (0): SubMConv3d() (1): SyncBatchNorm(64, eps=0.001, momentum=0.01, affine=True, track_running_stats=True) (2): ReLU() ) ) (conv4): SparseSequential( (0): SparseSequential( (0): SparseConv3d() (1): SyncBatchNorm(64, eps=0.001, momentum=0.01, affine=True, track_running_stats=True) (2): ReLU() ) (1): SparseSequential( (0): SubMConv3d() (1): SyncBatchNorm(64, eps=0.001, momentum=0.01, affine=True, track_running_stats=True) (2): ReLU() ) (2): SparseSequential( (0): SubMConv3d() (1): SyncBatchNorm(64, eps=0.001, momentum=0.01, affine=True, track_running_stats=True) (2): ReLU() ) ) (conv_out): SparseSequential( (0): SparseConv3d() ) ) (map_to_bev_module): HeightCompression() (backbone_2d): HgBEVBackbone( (rpn3d_conv2): Sequential( (0): Sequential( (0): Conv2d(160, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False) (1): SyncBatchNorm(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True) ) (1): ReLU(inplace=True) ) (rpn3d_conv3): hourglass2d( (conv1): Sequential( (0): Sequential( (0): Conv2d(64, 128, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False) (1): SyncBatchNorm(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True) ) (1): ReLU(inplace=True) ) (conv2): Sequential( (0): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False) (1): SyncBatchNorm(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True) ) (conv3): Sequential( (0): Sequential( (0): Conv2d(128, 128, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False) (1): SyncBatchNorm(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True) ) (1): ReLU(inplace=True) ) (conv4): Sequential( (0): Sequential( (0): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False) (1): SyncBatchNorm(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True) ) (1): ReLU(inplace=True) ) (conv5): Sequential( (0): ConvTranspose2d(128, 128, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), output_padding=(1, 1), bias=False) (1): SyncBatchNorm(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True) ) (conv6): Sequential( (0): ConvTranspose2d(128, 64, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), output_padding=(1, 1), bias=False) (1): SyncBatchNorm(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True) ) ) ) (dense_head): None (point_head): None ) (backbone_3d): LigaBackbone( (feature_backbone): ResNet( (conv1): Conv2d(3, 64, kernel_size=(7, 7), stride=(2, 2), padding=(3, 3), bias=False) (bn1): SyncBatchNorm(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True) (relu): ReLU(inplace=True) (layer1): ResLayer( (0): BasicBlock( (conv1): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False) (bn1): SyncBatchNorm(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True) (conv2): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False) (bn2): SyncBatchNorm(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True) (relu): ReLU(inplace=True) ) (1): BasicBlock( (conv1): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False) (bn1): SyncBatchNorm(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True) (conv2): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False) (bn2): SyncBatchNorm(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True) (relu): ReLU(inplace=True) ) (2): BasicBlock( (conv1): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False) (bn1): SyncBatchNorm(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True) (conv2): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False) (bn2): SyncBatchNorm(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True) (relu): ReLU(inplace=True) ) ) (layer2): ResLayer( (0): BasicBlock( (conv1): Conv2d(64, 128, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False) (bn1): SyncBatchNorm(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True) (conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False) (bn2): SyncBatchNorm(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True) (relu): ReLU(inplace=True) (downsample): Sequential( (0): Conv2d(64, 128, kernel_size=(1, 1), stride=(2, 2), bias=False) (1): SyncBatchNorm(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True) ) ) (1): BasicBlock( (conv1): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False) (bn1): SyncBatchNorm(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True) (conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False) (bn2): SyncBatchNorm(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True) (relu): ReLU(inplace=True) ) (2): BasicBlock( (conv1): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False) (bn1): SyncBatchNorm(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True) (conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False) (bn2): SyncBatchNorm(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True) (relu): ReLU(inplace=True) ) (3): BasicBlock( (conv1): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False) (bn1): SyncBatchNorm(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True) (conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False) (bn2): SyncBatchNorm(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True) (relu): ReLU(inplace=True) ) ) (layer3): ResLayer( (0): BasicBlock( (conv1): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(2, 2), dilation=(2, 2), bias=False) (bn1): SyncBatchNorm(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True) (conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False) (bn2): SyncBatchNorm(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True) (relu): ReLU(inplace=True) ) (1): BasicBlock( (conv1): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(2, 2), dilation=(2, 2), bias=False) (bn1): SyncBatchNorm(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True) (conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False) (bn2): SyncBatchNorm(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True) (relu): ReLU(inplace=True) ) (2): BasicBlock( (conv1): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(2, 2), dilation=(2, 2), bias=False) (bn1): SyncBatchNorm(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True) (conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False) (bn2): SyncBatchNorm(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True) (relu): ReLU(inplace=True) ) (3): BasicBlock( (conv1): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(2, 2), dilation=(2, 2), bias=False) (bn1): SyncBatchNorm(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True) (conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False) (bn2): SyncBatchNorm(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True) (relu): ReLU(inplace=True) ) (4): BasicBlock( (conv1): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(2, 2), dilation=(2, 2), bias=False) (bn1): SyncBatchNorm(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True) (conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False) (bn2): SyncBatchNorm(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True) (relu): ReLU(inplace=True) ) (5): BasicBlock( (conv1): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(2, 2), dilation=(2, 2), bias=False) (bn1): SyncBatchNorm(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True) (conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False) (bn2): SyncBatchNorm(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True) (relu): ReLU(inplace=True) ) ) (layer4): ResLayer( (0): BasicBlock( (conv1): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(4, 4), dilation=(4, 4), bias=False) (bn1): SyncBatchNorm(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True) (conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False) (bn2): SyncBatchNorm(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True) (relu): ReLU(inplace=True) ) (1): BasicBlock( (conv1): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(4, 4), dilation=(4, 4), bias=False) (bn1): SyncBatchNorm(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True) (conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False) (bn2): SyncBatchNorm(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True) (relu): ReLU(inplace=True) ) (2): BasicBlock( (conv1): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(4, 4), dilation=(4, 4), bias=False) (bn1): SyncBatchNorm(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True) (conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False) (bn2): SyncBatchNorm(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True) (relu): ReLU(inplace=True) ) ) ) (feature_neck): feature_extraction_neck( (spp_branches): ModuleList( (0): Sequential( (0): AvgPool2d(kernel_size=(64, 64), stride=(64, 64), padding=0) (1): Sequential( (0): Conv2d(128, 32, kernel_size=(1, 1), stride=(1, 1), bias=False) (1): GroupNorm(32, 32, eps=1e-05, affine=True) ) (2): ReLU(inplace=True) ) (1): Sequential( (0): AvgPool2d(kernel_size=(32, 32), stride=(32, 32), padding=0) (1): Sequential( (0): Conv2d(128, 32, kernel_size=(1, 1), stride=(1, 1), bias=False) (1): GroupNorm(32, 32, eps=1e-05, affine=True) ) (2): ReLU(inplace=True) ) (2): Sequential( (0): AvgPool2d(kernel_size=(16, 16), stride=(16, 16), padding=0) (1): Sequential( (0): Conv2d(128, 32, kernel_size=(1, 1), stride=(1, 1), bias=False) (1): GroupNorm(32, 32, eps=1e-05, affine=True) ) (2): ReLU(inplace=True) ) (3): Sequential( (0): AvgPool2d(kernel_size=(8, 8), stride=(8, 8), padding=0) (1): Sequential( (0): Conv2d(128, 32, kernel_size=(1, 1), stride=(1, 1), bias=False) (1): GroupNorm(32, 32, eps=1e-05, affine=True) ) (2): ReLU(inplace=True) ) ) (upconv_module): upconv_module( (conv): ModuleList( (0): Sequential( (0): Conv2d(512, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False) (1): SyncBatchNorm(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True) ) (1): Sequential( (0): Conv2d(64, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False) (1): SyncBatchNorm(32, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True) ) ) (redir): ModuleList( (0): Sequential( (0): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False) (1): SyncBatchNorm(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True) ) (1): Sequential( (0): Conv2d(3, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False) (1): SyncBatchNorm(32, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True) ) ) (up): Upsample(scale_factor=2.0, mode=bilinear) ) (lastconv): Sequential( (0): Sequential( (0): Conv2d(32, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False) (1): GroupNorm(32, 32, eps=1e-05, affine=True) ) (1): ReLU(inplace=True) (2): Conv2d(32, 32, kernel_size=(1, 1), stride=(1, 1), bias=False) ) (rpnconv): Sequential( (0): Sequential( (0): Conv2d(512, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False) (1): GroupNorm(32, 128, eps=1e-05, affine=True) ) (1): ReLU(inplace=True) (2): Sequential( (0): Conv2d(128, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False) (1): GroupNorm(32, 32, eps=1e-05, affine=True) ) (3): ReLU(inplace=True) ) ) (sem_neck): FPN( (lateral_convs): ModuleList( (0): ConvModule( (conv): Conv2d(32, 64, kernel_size=(1, 1), stride=(1, 1)) ) ) (fpn_convs): ModuleList( (0): ConvModule( (conv): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1)) ) (1): ConvModule( (conv): Conv2d(64, 64, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1)) ) (2): ConvModule( (conv): Conv2d(64, 64, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1)) ) (3): ConvModule( (conv): Conv2d(64, 64, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1)) ) (4): ConvModule( (conv): Conv2d(64, 64, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1)) ) ) ) (build_cost): BuildCostVolume (dres0): Sequential( (0): Sequential( (0): Conv3d(64, 32, kernel_size=(3, 3, 3), stride=(1, 1, 1), padding=(1, 1, 1), bias=False) (1): GroupNorm(32, 32, eps=1e-05, affine=True) ) (1): ReLU(inplace=True) ) (dres1): Sequential( (0): Sequential( (0): Conv3d(32, 32, kernel_size=(3, 3, 3), stride=(1, 1, 1), padding=(1, 1, 1), bias=False) (1): GroupNorm(32, 32, eps=1e-05, affine=True) ) ) (hg_stereo): ModuleList( (0): hourglass( (conv1): Sequential( (0): Sequential( (0): Conv3d(32, 64, kernel_size=(3, 3, 3), stride=(2, 2, 2), padding=(1, 1, 1), bias=False) (1): GroupNorm(32, 64, eps=1e-05, affine=True) ) (1): ReLU(inplace=True) ) (conv2): Sequential( (0): Conv3d(64, 64, kernel_size=(3, 3, 3), stride=(1, 1, 1), padding=(1, 1, 1), bias=False) (1): GroupNorm(32, 64, eps=1e-05, affine=True) ) (conv3): Sequential( (0): Sequential( (0): Conv3d(64, 64, kernel_size=(3, 3, 3), stride=(2, 2, 2), padding=(1, 1, 1), bias=False) (1): GroupNorm(32, 64, eps=1e-05, affine=True) ) (1): ReLU(inplace=True) ) (conv4): Sequential( (0): Sequential( (0): Conv3d(64, 64, kernel_size=(3, 3, 3), stride=(1, 1, 1), padding=(1, 1, 1), bias=False) (1): GroupNorm(32, 64, eps=1e-05, affine=True) ) (1): ReLU(inplace=True) ) (conv5): Sequential( (0): ConvTranspose3d(64, 64, kernel_size=(3, 3, 3), stride=(2, 2, 2), padding=(1, 1, 1), output_padding=(1, 1, 1), bias=False) (1): GroupNorm(32, 64, eps=1e-05, affine=True) ) (conv6): Sequential( (0): ConvTranspose3d(64, 32, kernel_size=(3, 3, 3), stride=(2, 2, 2), padding=(1, 1, 1), output_padding=(1, 1, 1), bias=False) (1): GroupNorm(32, 32, eps=1e-05, affine=True) ) ) ) (pred_stereo): ModuleList( (0): Sequential( (0): Sequential( (0): Conv3d(32, 32, kernel_size=(3, 3, 3), stride=(1, 1, 1), padding=(1, 1, 1), bias=False) (1): GroupNorm(32, 32, eps=1e-05, affine=True) ) (1): ReLU(inplace=True) (2): Conv3d(32, 1, kernel_size=(3, 3, 3), stride=(1, 1, 1), padding=(1, 1, 1), bias=False) (3): Upsample(scale_factor=4.0, mode=trilinear) ) ) (dispregression): disparityregression() (rpn3d_convs): Sequential( (0): Sequential( (0): Sequential( (0): Conv3d(64, 32, kernel_size=(3, 3, 3), stride=(1, 1, 1), padding=(1, 1, 1), bias=False) (1): GroupNorm(32, 32, eps=1e-05, affine=True) ) (1): ReLU(inplace=True) ) ) (rpn3d_pool): AvgPool3d(kernel_size=(4, 1, 1), stride=(4, 1, 1), padding=0) ) (map_to_bev_module): HeightCompression() (backbone_2d): HgBEVBackbone( (rpn3d_conv2): Sequential( (0): Sequential( (0): Conv2d(160, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False) (1): GroupNorm(32, 64, eps=1e-05, affine=True) ) (1): ReLU(inplace=True) ) (rpn3d_conv3): hourglass2d( (conv1): Sequential( (0): Sequential( (0): Conv2d(64, 128, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False) (1): GroupNorm(32, 128, eps=1e-05, affine=True) ) (1): ReLU(inplace=True) ) (conv2): Sequential( (0): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False) (1): GroupNorm(32, 128, eps=1e-05, affine=True) ) (conv3): Sequential( (0): Sequential( (0): Conv2d(128, 128, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False) (1): GroupNorm(32, 128, eps=1e-05, affine=True) ) (1): ReLU(inplace=True) ) (conv4): Sequential( (0): Sequential( (0): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False) (1): GroupNorm(32, 128, eps=1e-05, affine=True) ) (1): ReLU(inplace=True) ) (conv5): Sequential( (0): ConvTranspose2d(128, 128, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), output_padding=(1, 1), bias=False) (1): GroupNorm(32, 128, eps=1e-05, affine=True) ) (conv6): Sequential( (0): ConvTranspose2d(128, 64, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), output_padding=(1, 1), bias=False) (1): GroupNorm(32, 64, eps=1e-05, affine=True) ) ) ) (dense_head_2d): MMDet2DHead( (bbox_head): ATSSAdvHead( (loss_cls): FocalLoss() (loss_bbox): GIoULoss() (relu): ReLU(inplace=True) (cls_convs): ModuleList( (0): ConvModule( (conv): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False) (gn): GroupNorm(32, 64, eps=1e-05, affine=True) (activate): ReLU(inplace=True) ) (1): ConvModule( (conv): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False) (gn): GroupNorm(32, 64, eps=1e-05, affine=True) (activate): ReLU(inplace=True) ) (2): ConvModule( (conv): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False) (gn): GroupNorm(32, 64, eps=1e-05, affine=True) (activate): ReLU(inplace=True) ) (3): ConvModule( (conv): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False) (gn): GroupNorm(32, 64, eps=1e-05, affine=True) (activate): ReLU(inplace=True) ) ) (reg_convs): ModuleList( (0): ConvModule( (conv): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False) (gn): GroupNorm(32, 64, eps=1e-05, affine=True) (activate): ReLU(inplace=True) ) (1): ConvModule( (conv): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False) (gn): GroupNorm(32, 64, eps=1e-05, affine=True) (activate): ReLU(inplace=True) ) (2): ConvModule( (conv): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False) (gn): GroupNorm(32, 64, eps=1e-05, affine=True) (activate): ReLU(inplace=True) ) (3): ConvModule( (conv): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False) (gn): GroupNorm(32, 64, eps=1e-05, affine=True) (activate): ReLU(inplace=True) ) ) (atss_cls): Conv2d(64, 3, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1)) (atss_reg): Conv2d(64, 12, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1)) (atss_centerness): Conv2d(64, 1, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1)) (scales): ModuleList( (0): Scale() (1): Scale() (2): Scale() (3): Scale() (4): Scale() ) (loss_centerness): CrossEntropyLoss() ) ) (dense_head): DetHead( (norm_imitation): ModuleDict( (spatial_features_2d): NormalizeLayer() (volume_features): NormalizeLayer() ) (conv_imitation): ModuleList( (0): Conv2d(64, 64, kernel_size=(1, 1), stride=(1, 1)) (1): Conv3d(32, 32, kernel_size=(1, 1, 1), stride=(1, 1, 1)) ) (cls_loss_func): SigmoidFocalClassificationLoss() (reg_loss_func): WeightedSmoothL1Loss() (imitation_loss_func): WeightedL2WithSigmaLoss() (iou_loss_func): IOU3dLoss() (dir_loss_func): WeightedCrossEntropyLoss() (rpn3d_cls_convs): Sequential( (0): Sequential( (0): Sequential( (0): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False) (1): GroupNorm(32, 64, eps=1e-05, affine=True) ) (1): ReLU(inplace=True) ) (1): Sequential( (0): Sequential( (0): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False) (1): GroupNorm(32, 64, eps=1e-05, affine=True) ) (1): ReLU(inplace=True) ) ) (rpn3d_bbox_convs): Sequential( (0): Sequential( (0): Sequential( (0): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False) (1): GroupNorm(32, 64, eps=1e-05, affine=True) ) (1): ReLU(inplace=True) ) (1): Sequential( (0): Sequential( (0): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False) (1): GroupNorm(32, 64, eps=1e-05, affine=True) ) (1): ReLU(inplace=True) ) ) (conv_cls): Conv2d(64, 18, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1)) (conv_box): Conv2d(64, 42, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1)) (conv_dir_cls): Conv2d(64, 12, kernel_size=(1, 1), stride=(1, 1)) ) (depth_loss_head): DepthLossHead() ) ) 2021-11-08 09:43:59,550 INFO *******Start training: /home/chenyinan/Projects/LIGA-Stereo_DEV/outputs/stereo_kitti_models/liga.3d-and-bev.dev ********