eval_rcnn TEST mode

[ssolchoi]

eval_rcnn cgf has a test option. Should I change it to test when submitting with kitti performance? (However, if you change it to test, an error occurs saying there is no train_mask. so.. Should I put train_mask in the test data?)

I want to create data that submits to KITTI. please tell me how i can do it.

[Test CAR command line] CUDA_VISIBLE_DEVICES=0 python eval_rcnn.py --cfg_file cfgs/CAR_EPNet_plus_plus.yaml --eval_mode rcnn --test \ --output_dir ./log/CAR_EPNet_plus_plus/test_results/ \ --data_path ../data/ \ --ckpt ./log/CAR_EPNet_plus_plus/ckpt/checkpoint_epoch_48.pth \ --set LI_FUSION.ENABLED True LI_FUSION.ADD_Image_Attention True CROSS_FUSION True USE_P2I_GATE True \ DEEP_RCNN_FUSION False USE_IMAGE_LOSS True IMAGE_WEIGHT 1.0 USE_IMAGE_SCORE True USE_IMG_DENSE_LOSS True USE_MC_LOSS True \ MC_LOSS_WEIGHT 1.0 I2P_Weight 0.5 P2I_Weight 0.5 ADD_MC_MASK True MC_MASK_THRES 0.2 SAVE_MODEL_PREP 0.8

Thank you for your reply.

[happinesslz]

@ssolchoi Thanks for you attention! For submitting the results to KITTI test benchmark, please modify the configure in SPLIT: val --> SPLIT: test

[ssolchoi]

Thank you!

But I already did that but got an error. (both CAR_EPNet_plus_plus.yaml , config.py)

Error 1 was resolved in the following way. https://stackoverflow.com/questions/54058256/runtimeerror-errors-in-loading-state-dict-for-resnet

But the problem of Error 2 remains. Should I change the run command..?

maybe, should I need to change the "--set" option as below? LI_FUSION.ENABLED False

Looking at the code, it seems that test_mask is required to evel_rcnn in test mode.. Can you give it to me?

please help me..

[ CAR_EPNet_plus_plus.yaml ] TEST: SPLIT: test RPN_PRE_NMS_TOP_N: 9000 RPN_POST_NMS_TOP_N: 100 RPN_NMS_THRESH: 0.8 RPN_DISTANCE_BASED_PROPOSE: True BBOX_AVG_BY_BIN: True RY_WITH_BIN: False

[ Commend line ] CUDA_VISIBLE_DEVICES=0 python eval_rcnn.py --cfg_file cfgs/CAR_EPNet_plus_plus.yaml --eval_mode rcnn_online --test --output_dir ./epnet_plus_plus_released_trained_models/CAR/eval_results/ --data_path ../data/ --ckpt ./pre_trained_model/epnet_plus_plus_released_trained_models/CAR/checkpoint_epoch_43.pth --set LI_FUSION.ENABLED True LI_FUSION.ADD_Image_Attention True CROSS_FUSION True USE_P2I_GATE True DEEP_RCNN_FUSION False USE_IMAGE_LOSS True IMAGE_WEIGHT 1.0 USE_IMAGE_SCORE True

[ Error 1 ] 2023-03-03 07:29:01,154 INFO **Start logging** 2023-03-03 07:29:01,154 INFO cfg_file cfgs/CAR_EPNet_plus_plus.yaml 2023-03-03 07:29:01,154 INFO eval_mode rcnn_online 2023-03-03 07:29:01,154 INFO eval_all False 2023-03-03 07:29:01,154 INFO test True 2023-03-03 07:29:01,154 INFO ckpt ./pre_trained_model/epnet_plus_plus_released_trained_models/CAR/checkpoint_epoch_43.pth 2023-03-03 07:29:01,154 INFO rpn_ckpt None 2023-03-03 07:29:01,154 INFO rcnn_ckpt None 2023-03-03 07:29:01,154 INFO batch_size 1 2023-03-03 07:29:01,154 INFO workers 4 2023-03-03 07:29:01,154 INFO extra_tag default 2023-03-03 07:29:01,154 INFO output_dir ./epnet_plus_plus_released_trained_models/CAR/eval_results/ 2023-03-03 07:29:01,154 INFO ckpt_dir None 2023-03-03 07:29:01,154 INFO data_path ../data/ 2023-03-03 07:29:01,155 INFO save_result False 2023-03-03 07:29:01,155 INFO save_rpn_feature False 2023-03-03 07:29:01,155 INFO random_select True 2023-03-03 07:29:01,155 INFO start_epoch 0 2023-03-03 07:29:01,155 INFO max_waiting_mins 30 2023-03-03 07:29:01,155 INFO rcnn_eval_roi_dir None 2023-03-03 07:29:01,155 INFO rcnn_eval_feature_dir None 2023-03-03 07:29:01,155 INFO set_cfgs None 2023-03-03 07:29:01,155 INFO model_type base 2023-03-03 07:29:01,155 INFO cfg.TAG: CAR_EPNet_plus_plus 2023-03-03 07:29:01,155 INFO cfg.CLASSES: Car 2023-03-03 07:29:01,155 INFO cfg.INCLUDE_SIMILAR_TYPE: True 2023-03-03 07:29:01,155 INFO cfg.AUG_DATA: True 2023-03-03 07:29:01,155 INFO cfg.AUG_METHOD_LIST: ['rotation', 'scaling', 'flip'] 2023-03-03 07:29:01,155 INFO cfg.AUG_METHOD_PROB: [1.0, 1.0, 0.5] 2023-03-03 07:29:01,156 INFO cfg.AUG_ROT_RANGE: 18 2023-03-03 07:29:01,156 INFO cfg.GT_AUG_ENABLED: False 2023-03-03 07:29:01,156 INFO cfg.GT_EXTRA_NUM: 15 2023-03-03 07:29:01,156 INFO cfg.GT_AUG_RAND_NUM: True 2023-03-03 07:29:01,156 INFO cfg.GT_AUG_APPLY_PROB: 1.0 2023-03-03 07:29:01,156 INFO cfg.GT_AUG_HARD_RATIO: 0.6 2023-03-03 07:29:01,156 INFO cfg.PC_REDUCE_BY_RANGE: True 2023-03-03 07:29:01,156 INFO cfg.PC_AREA_SCOPE: [[-40. 40. ] [ -1. 3. ] [ 0. 70.4]] 2023-03-03 07:29:01,157 INFO cfg.CLS_MEAN_SIZE: [[1.5256319 1.6285675 3.8831165]] 2023-03-03 07:29:01,157 INFO cfg.USE_IOU_BRANCH: True 2023-03-03 07:29:01,157 INFO cfg.USE_IM_DEPTH: False 2023-03-03 07:29:01,157 INFO cfg.USE_PSEUDO_LIDAR: False 2023-03-03 07:29:01,157 INFO cfg.CROSS_FUSION: False 2023-03-03 07:29:01,157 INFO cfg.INPUT_CROSS_FUSION: False 2023-03-03 07:29:01,157 INFO cfg.USE_KNN_FUSION: False 2023-03-03 07:29:01,157 INFO cfg.USE_SELF_ATTENTION: False 2023-03-03 07:29:01,157 INFO cfg.DEEP_RCNN_FUSION: False 2023-03-03 07:29:01,157 INFO cfg.USE_IMAGE_LOSS: False 2023-03-03 07:29:01,157 INFO cfg.IMAGE_WEIGHT: 1.0 2023-03-03 07:29:01,157 INFO cfg.USE_IMAGE_LOSS_TYPE: CrossEntropyLoss 2023-03-03 07:29:01,157 INFO cfg.USE_IMAGE_SCORE: False 2023-03-03 07:29:01,157 INFO cfg.USE_IMG_DENSE_LOSS: False 2023-03-03 07:29:01,157 INFO cfg.USE_KL_LOSS: False 2023-03-03 07:29:01,158 INFO cfg.USE_KL_LOSS_TYPE: KL 2023-03-03 07:29:01,158 INFO cfg.MC_LOSS_WEIGHT: 1.0 2023-03-03 07:29:01,158 INFO cfg.SAVE_MODEL_PREP: 0.8 2023-03-03 07:29:01,158 INFO cfg.USE_P2I_GATE: False 2023-03-03 07:29:01,158 INFO cfg.STACK_CROSS_FUSION: False 2023-03-03 07:29:01,158 INFO cfg.USE_IMAGE_RES: False 2023-03-03 07:29:01,158 INFO cfg.RCNN_IMG_CHANNEL: 32 2023-03-03 07:29:01,158 INFO cfg.ONLY_USE_IMAGE_FEAT: False 2023-03-03 07:29:01,158 INFO cfg.USE_POINT_ATT_FEATURE: False 2023-03-03 07:29:01,158 INFO cfg.USE_POINT_FEATURE_RES: False 2023-03-03 07:29:01,158 INFO cfg.I2P_Weight: 0.5 2023-03-03 07:29:01,158 INFO cfg.P2I_Weight: 0.5 2023-03-03 07:29:01,158 INFO cfg.USE_MC_LOSS: False 2023-03-03 07:29:01,158 INFO cfg.ADD_MC_MASK: False 2023-03-03 07:29:01,158 INFO cfg.MC_MASK_THRES: 0.45 2023-03-03 07:29:01,159 INFO cfg.USE_PURE_IMG_BACKBONE: False 2023-03-03 07:29:01,159 INFO cfg.USE_PAINTING_SCORE: False 2023-03-03 07:29:01,159 INFO cfg.USE_PAINTING_FEAT: False 2023-03-03 07:29:01,159 INFO
cfg.LI_FUSION = edict() 2023-03-03 07:29:01,159 INFO cfg.LI_FUSION.ENABLED: True 2023-03-03 07:29:01,159 INFO cfg.LI_FUSION.IMG_FEATURES_CHANNEL: 128 2023-03-03 07:29:01,159 INFO cfg.LI_FUSION.ADD_Image_Attention: True 2023-03-03 07:29:01,159 INFO cfg.LI_FUSION.IMG_CHANNELS: [3, 64, 128, 256, 512] 2023-03-03 07:29:01,159 INFO cfg.LI_FUSION.POINT_CHANNELS: [96, 256, 512, 1024] 2023-03-03 07:29:01,159 INFO cfg.LI_FUSION.DeConv_Reduce: [16, 16, 16, 16] 2023-03-03 07:29:01,159 INFO cfg.LI_FUSION.DeConv_Kernels: [2, 4, 8, 16] 2023-03-03 07:29:01,159 INFO cfg.LI_FUSION.DeConv_Strides: [2, 4, 8, 16] 2023-03-03 07:29:01,159 INFO
cfg.RPN = edict() 2023-03-03 07:29:01,159 INFO cfg.RPN.ENABLED: True 2023-03-03 07:29:01,159 INFO cfg.RPN.FIXED: False 2023-03-03 07:29:01,159 INFO cfg.RPN.USE_INTENSITY: False 2023-03-03 07:29:01,159 INFO cfg.RPN.USE_RGB: False 2023-03-03 07:29:01,160 INFO cfg.RPN.LOC_XZ_FINE: True 2023-03-03 07:29:01,160 INFO cfg.RPN.LOC_SCOPE: 3.0 2023-03-03 07:29:01,160 INFO cfg.RPN.LOC_BIN_SIZE: 0.5 2023-03-03 07:29:01,160 INFO cfg.RPN.NUM_HEAD_BIN: 12 2023-03-03 07:29:01,160 INFO cfg.RPN.BACKBONE: pointnet2_msg 2023-03-03 07:29:01,160 INFO cfg.RPN.USE_BN: True 2023-03-03 07:29:01,160 INFO cfg.RPN.NUM_POINTS: 16384 2023-03-03 07:29:01,160 INFO
cfg.RPN.SA_CONFIG = edict() 2023-03-03 07:29:01,160 INFO cfg.RPN.SA_CONFIG.ATTN_DIM: 128 2023-03-03 07:29:01,160 INFO cfg.RPN.SA_CONFIG.ATTN: [0, 0, 128, 128] 2023-03-03 07:29:01,160 INFO cfg.RPN.SA_CONFIG.NPOINTS: [4096, 1024, 256, 64] 2023-03-03 07:29:01,160 INFO cfg.RPN.SA_CONFIG.RADIUS: [[0.1, 0.5], [0.5, 1.0], [1.0, 2.0], [2.0, 4.0]] 2023-03-03 07:29:01,160 INFO cfg.RPN.SA_CONFIG.NSAMPLE: [[16, 32], [16, 32], [16, 32], [16, 32]] 2023-03-03 07:29:01,160 INFO cfg.RPN.SA_CONFIG.MLPS: [[[16, 16, 32], [32, 32, 64]], [[64, 64, 128], [64, 96, 128]], [[128, 196, 256], [128, 196, 256]], [[256, 256, 512], [256, 384, 512]]] 2023-03-03 07:29:01,160 INFO cfg.RPN.FP_MLPS: [[128, 128], [256, 256], [512, 512], [512, 512]] 2023-03-03 07:29:01,161 INFO cfg.RPN.CLS_FC: [128] 2023-03-03 07:29:01,161 INFO cfg.RPN.REG_FC: [128] 2023-03-03 07:29:01,161 INFO cfg.RPN.DP_RATIO: 0.5 2023-03-03 07:29:01,161 INFO cfg.RPN.LOSS_CLS: SigmoidFocalLoss 2023-03-03 07:29:01,161 INFO cfg.RPN.FG_WEIGHT: 15 2023-03-03 07:29:01,161 INFO cfg.RPN.FOCAL_ALPHA: [0.25, 0.75] 2023-03-03 07:29:01,161 INFO cfg.RPN.FOCAL_GAMMA: 2.0 2023-03-03 07:29:01,161 INFO cfg.RPN.REG_LOSS_WEIGHT: [1.0, 1.0, 1.0, 1.0] 2023-03-03 07:29:01,161 INFO cfg.RPN.LOSS_WEIGHT: [1.0, 1.0] 2023-03-03 07:29:01,161 INFO cfg.RPN.NMS_TYPE: normal 2023-03-03 07:29:01,161 INFO cfg.RPN.SCORE_THRESH: 0.2 2023-03-03 07:29:01,161 INFO
cfg.RCNN = edict() 2023-03-03 07:29:01,161 INFO cfg.RCNN.ENABLED: True 2023-03-03 07:29:01,161 INFO cfg.RCNN.USE_RPN_FEATURES: True 2023-03-03 07:29:01,161 INFO cfg.RCNN.USE_MASK: True 2023-03-03 07:29:01,161 INFO cfg.RCNN.MASK_TYPE: seg 2023-03-03 07:29:01,162 INFO cfg.RCNN.USE_INTENSITY: False 2023-03-03 07:29:01,162 INFO cfg.RCNN.USE_DEPTH: True 2023-03-03 07:29:01,162 INFO cfg.RCNN.USE_SEG_SCORE: False 2023-03-03 07:29:01,162 INFO cfg.RCNN.ROI_SAMPLE_JIT: True 2023-03-03 07:29:01,162 INFO cfg.RCNN.ROI_FG_AUG_TIMES: 10 2023-03-03 07:29:01,162 INFO cfg.RCNN.REG_AUG_METHOD: multiple 2023-03-03 07:29:01,162 INFO cfg.RCNN.POOL_EXTRA_WIDTH: 0.2 2023-03-03 07:29:01,162 INFO cfg.RCNN.USE_RGB: False 2023-03-03 07:29:01,162 INFO cfg.RCNN.LOC_SCOPE: 1.5 2023-03-03 07:29:01,162 INFO cfg.RCNN.LOC_BIN_SIZE: 0.5 2023-03-03 07:29:01,162 INFO cfg.RCNN.NUM_HEAD_BIN: 9 2023-03-03 07:29:01,162 INFO cfg.RCNN.LOC_Y_BY_BIN: False 2023-03-03 07:29:01,162 INFO cfg.RCNN.LOC_Y_SCOPE: 0.5 2023-03-03 07:29:01,162 INFO cfg.RCNN.LOC_Y_BIN_SIZE: 0.25 2023-03-03 07:29:01,162 INFO cfg.RCNN.SIZE_RES_ON_ROI: False 2023-03-03 07:29:01,163 INFO cfg.RCNN.USE_BN: False 2023-03-03 07:29:01,163 INFO cfg.RCNN.DP_RATIO: 0.0 2023-03-03 07:29:01,163 INFO cfg.RCNN.BACKBONE: pointnet 2023-03-03 07:29:01,163 INFO cfg.RCNN.XYZ_UP_LAYER: [128, 128] 2023-03-03 07:29:01,163 INFO cfg.RCNN.NUM_POINTS: 512 2023-03-03 07:29:01,163 INFO
cfg.RCNN.SA_CONFIG = edict() 2023-03-03 07:29:01,163 INFO cfg.RCNN.SA_CONFIG.NPOINTS: [128, 32, -1] 2023-03-03 07:29:01,163 INFO cfg.RCNN.SA_CONFIG.RADIUS: [0.2, 0.4, 100] 2023-03-03 07:29:01,163 INFO cfg.RCNN.SA_CONFIG.NSAMPLE: [64, 64, 64] 2023-03-03 07:29:01,163 INFO cfg.RCNN.SA_CONFIG.MLPS: [[128, 128, 128], [128, 128, 256], [256, 256, 512]] 2023-03-03 07:29:01,163 INFO cfg.RCNN.CLS_FC: [512, 512] 2023-03-03 07:29:01,163 INFO cfg.RCNN.REG_FC: [512, 512] 2023-03-03 07:29:01,163 INFO cfg.RCNN.LOSS_CLS: BinaryCrossEntropy 2023-03-03 07:29:01,163 INFO cfg.RCNN.FOCAL_ALPHA: [0.25, 0.75] 2023-03-03 07:29:01,163 INFO cfg.RCNN.FOCAL_GAMMA: 2.0 2023-03-03 07:29:01,164 INFO cfg.RCNN.CLS_WEIGHT: [1. 1. 1.] 2023-03-03 07:29:01,164 INFO cfg.RCNN.CLS_FG_THRESH: 0.6 2023-03-03 07:29:01,164 INFO cfg.RCNN.CLS_BG_THRESH: 0.45 2023-03-03 07:29:01,164 INFO cfg.RCNN.CLS_BG_THRESH_LO: 0.05 2023-03-03 07:29:01,164 INFO cfg.RCNN.REG_FG_THRESH: 0.55 2023-03-03 07:29:01,164 INFO cfg.RCNN.FG_RATIO: 0.5 2023-03-03 07:29:01,164 INFO cfg.RCNN.ROI_PER_IMAGE: 64 2023-03-03 07:29:01,164 INFO cfg.RCNN.HARD_BG_RATIO: 0.8 2023-03-03 07:29:01,164 INFO cfg.RCNN.IOU_LOSS_TYPE: raw 2023-03-03 07:29:01,164 INFO cfg.RCNN.IOU_ANGLE_POWER: 1 2023-03-03 07:29:01,164 INFO cfg.RCNN.SCORE_THRESH: 0.2 2023-03-03 07:29:01,164 INFO cfg.RCNN.NMS_THRESH: 0.1 2023-03-03 07:29:01,165 INFO
cfg.TRAIN = edict() 2023-03-03 07:29:01,165 INFO cfg.TRAIN.SPLIT: train 2023-03-03 07:29:01,165 INFO cfg.TRAIN.VAL_SPLIT: smallval 2023-03-03 07:29:01,165 INFO cfg.TRAIN.LR: 0.002 2023-03-03 07:29:01,165 INFO cfg.TRAIN.LR_CLIP: 1e-05 2023-03-03 07:29:01,165 INFO cfg.TRAIN.LR_DECAY: 0.5 2023-03-03 07:29:01,165 INFO cfg.TRAIN.DECAY_STEP_LIST: [100, 150, 180, 200] 2023-03-03 07:29:01,165 INFO cfg.TRAIN.LR_WARMUP: True 2023-03-03 07:29:01,165 INFO cfg.TRAIN.WARMUP_MIN: 0.0002 2023-03-03 07:29:01,165 INFO cfg.TRAIN.WARMUP_EPOCH: 1 2023-03-03 07:29:01,165 INFO cfg.TRAIN.BN_MOMENTUM: 0.1 2023-03-03 07:29:01,165 INFO cfg.TRAIN.BN_DECAY: 0.5 2023-03-03 07:29:01,165 INFO cfg.TRAIN.BNM_CLIP: 0.01 2023-03-03 07:29:01,165 INFO cfg.TRAIN.BN_DECAY_STEP_LIST: [1000] 2023-03-03 07:29:01,165 INFO cfg.TRAIN.OPTIMIZER: adam_onecycle 2023-03-03 07:29:01,165 INFO cfg.TRAIN.WEIGHT_DECAY: 0.001 2023-03-03 07:29:01,166 INFO cfg.TRAIN.MOMENTUM: 0.9 2023-03-03 07:29:01,166 INFO cfg.TRAIN.MOMS: [0.95, 0.85] 2023-03-03 07:29:01,166 INFO cfg.TRAIN.DIV_FACTOR: 10.0 2023-03-03 07:29:01,166 INFO cfg.TRAIN.PCT_START: 0.4 2023-03-03 07:29:01,166 INFO cfg.TRAIN.GRAD_NORM_CLIP: 1.0 2023-03-03 07:29:01,166 INFO cfg.TRAIN.RPN_PRE_NMS_TOP_N: 9000 2023-03-03 07:29:01,166 INFO cfg.TRAIN.RPN_POST_NMS_TOP_N: 512 2023-03-03 07:29:01,166 INFO cfg.TRAIN.RPN_NMS_THRESH: 0.85 2023-03-03 07:29:01,166 INFO cfg.TRAIN.RPN_DISTANCE_BASED_PROPOSE: True 2023-03-03 07:29:01,166 INFO cfg.TRAIN.RPN_TRAIN_WEIGHT: 1.0 2023-03-03 07:29:01,166 INFO cfg.TRAIN.RCNN_TRAIN_WEIGHT: 1.0 2023-03-03 07:29:01,166 INFO cfg.TRAIN.CE_WEIGHT: 5.0 2023-03-03 07:29:01,166 INFO cfg.TRAIN.RPN_CE_WEIGHT: 5.0 2023-03-03 07:29:01,166 INFO cfg.TRAIN.IOU_LOSS_TYPE: cls_mask_with_bin 2023-03-03 07:29:01,166 INFO cfg.TRAIN.BBOX_AVG_BY_BIN: True 2023-03-03 07:29:01,166 INFO cfg.TRAIN.RY_WITH_BIN: False 2023-03-03 07:29:01,166 INFO
cfg.TEST = edict() 2023-03-03 07:29:01,167 INFO cfg.TEST.SPLIT: test 2023-03-03 07:29:01,167 INFO cfg.TEST.RPN_PRE_NMS_TOP_N: 9000 2023-03-03 07:29:01,167 INFO cfg.TEST.RPN_POST_NMS_TOP_N: 100 2023-03-03 07:29:01,167 INFO cfg.TEST.RPN_NMS_THRESH: 0.8 2023-03-03 07:29:01,167 INFO cfg.TEST.RPN_DISTANCE_BASED_PROPOSE: True 2023-03-03 07:29:01,167 INFO cfg.TEST.BBOX_AVG_BY_BIN: True 2023-03-03 07:29:01,167 INFO cfg.TEST.RY_WITH_BIN: False 2023-03-03 07:29:01,171 INFO Load testing samples from ../data/KITTI/object/testing 2023-03-03 07:29:01,171 INFO Done: total test samples 7518 2023-03-03 07:29:04,976 INFO ==> Loading from checkpoint './pre_trained_model/epnet_plus_plus_released_trained_models/CAR/checkpoint_epoch_43.pth' Traceback (most recent call last): File "eval_rcnn.py", line 1026, in eval_single_ckpt(root_result_dir, data_path=args.data_path) File "eval_rcnn.py", line 865, in eval_single_ckpt load_ckpt_based_on_args(model, logger) File "eval_rcnn.py", line 817, in load_ckpt_based_on_args train_utils.load_checkpoint(model, filename = args.ckpt, logger = logger) File "/workspace/EPNetV2/tools/../tools/train_utils/train_utils.py", line 85, in load_checkpoint model.load_state_dict(checkpoint['model_state']) File "/root/anaconda3/envs/epnet_2/lib/python3.8/site-packages/torch/nn/modules/module.py", line 1051, in load_state_dict raise RuntimeError('Error(s) in loading state_dict for {}:\n\t{}'.format( RuntimeError: Error(s) in loading state_dict for PointRCNN: Unexpected key(s) in state_dict: "rpn.rpn_image_cls_layer.conv1.weight", "rpn.rpn_image_cls_layer.bn1.weight", "rpn.rpn_image_cls_layer.bn1.bias", "rpn.rpn_image_cls_layer.bn1.running_mean", "rpn.rpn_image_cls_layer.bn1.running_var", "rpn.rpn_image_cls_layer.bn1.num_batches_tracked", "rpn.rpn_image_cls_layer.conv2.weight", "rpn.backbone_net.Cross_Fusion.0.P2IA_Layer.conv1.0.weight", "rpn.backbone_net.Cross_Fusion.0.P2IA_Layer.conv1.0.bias", "rpn.backbone_net.Cross_Fusion.0.P2IA_Layer.conv1.1.weight", "rpn.backbone_net.Cross_Fusion.0.P2IA_Layer.conv1.1.bias", "rpn.backbone_net.Cross_Fusion.0.P2IA_Layer.conv1.1.running_mean", "rpn.backbone_net.Cross_Fusion.0.P2IA_Layer.conv1.1.running_var", "rpn.backbone_net.Cross_Fusion.0.P2IA_Layer.conv1.1.num_batches_tracked", "rpn.backbone_net.Cross_Fusion.0.P2IA_Layer.fc1.weight", "rpn.backbone_net.Cross_Fusion.0.P2IA_Layer.fc1.bias", "rpn.backbone_net.Cross_Fusion.0.P2IA_Layer.fc2.weight", "rpn.backbone_net.Cross_Fusion.0.P2IA_Layer.fc2.bias", "rpn.backbone_net.Cross_Fusion.0.P2IA_Layer.fc3.weight", "rpn.backbone_net.Cross_Fusion.0.P2IA_Layer.fc3.bias", "rpn.backbone_net.Cross_Fusion.0.conv1.weight", "rpn.backbone_net.Cross_Fusion.0.bn1.weight", "rpn.backbone_net.Cross_Fusion.0.bn1.bias", "rpn.backbone_net.Cross_Fusion.0.bn1.running_mean", "rpn.backbone_net.Cross_Fusion.0.bn1.running_var", "rpn.backbone_net.Cross_Fusion.0.bn1.num_batches_tracked", "rpn.backbone_net.Cross_Fusion.1.P2IA_Layer.conv1.0.weight", "rpn.backbone_net.Cross_Fusion.1.P2IA_Layer.conv1.0.bias", "rpn.backbone_net.Cross_Fusion.1.P2IA_Layer.conv1.1.weight", "rpn.backbone_net.Cross_Fusion.1.P2IA_Layer.conv1.1.bias", "rpn.backbone_net.Cross_Fusion.1.P2IA_Layer.conv1.1.running_mean", "rpn.backbone_net.Cross_Fusion.1.P2IA_Layer.conv1.1.running_var", "rpn.backbone_net.Cross_Fusion.1.P2IA_Layer.conv1.1.num_batches_tracked", "rpn.backbone_net.Cross_Fusion.1.P2IA_Layer.fc1.weight", "rpn.backbone_net.Cross_Fusion.1.P2IA_Layer.fc1.bias", "rpn.backbone_net.Cross_Fusion.1.P2IA_Layer.fc2.weight", "rpn.backbone_net.Cross_Fusion.1.P2IA_Layer.fc2.bias", "rpn.backbone_net.Cross_Fusion.1.P2IA_Layer.fc3.weight", "rpn.backbone_net.Cross_Fusion.1.P2IA_Layer.fc3.bias", "rpn.backbone_net.Cross_Fusion.1.conv1.weight", "rpn.backbone_net.Cross_Fusion.1.bn1.weight", "rpn.backbone_net.Cross_Fusion.1.bn1.bias", "rpn.backbone_net.Cross_Fusion.1.bn1.running_mean", "rpn.backbone_net.Cross_Fusion.1.bn1.running_var", "rpn.backbone_net.Cross_Fusion.1.bn1.num_batches_tracked", "rpn.backbone_net.Cross_Fusion.2.P2IA_Layer.conv1.0.weight", "rpn.backbone_net.Cross_Fusion.2.P2IA_Layer.conv1.0.bias", "rpn.backbone_net.Cross_Fusion.2.P2IA_Layer.conv1.1.weight", "rpn.backbone_net.Cross_Fusion.2.P2IA_Layer.conv1.1.bias", "rpn.backbone_net.Cross_Fusion.2.P2IA_Layer.conv1.1.running_mean", "rpn.backbone_net.Cross_Fusion.2.P2IA_Layer.conv1.1.running_var", "rpn.backbone_net.Cross_Fusion.2.P2IA_Layer.conv1.1.num_batches_tracked", "rpn.backbone_net.Cross_Fusion.2.P2IA_Layer.fc1.weight", "rpn.backbone_net.Cross_Fusion.2.P2IA_Layer.fc1.bias", "rpn.backbone_net.Cross_Fusion.2.P2IA_Layer.fc2.weight", "rpn.backbone_net.Cross_Fusion.2.P2IA_Layer.fc2.bias", "rpn.backbone_net.Cross_Fusion.2.P2IA_Layer.fc3.weight", "rpn.backbone_net.Cross_Fusion.2.P2IA_Layer.fc3.bias", "rpn.backbone_net.Cross_Fusion.2.conv1.weight", "rpn.backbone_net.Cross_Fusion.2.bn1.weight", "rpn.backbone_net.Cross_Fusion.2.bn1.bias", "rpn.backbone_net.Cross_Fusion.2.bn1.running_mean", "rpn.backbone_net.Cross_Fusion.2.bn1.running_var", "rpn.backbone_net.Cross_Fusion.2.bn1.num_batches_tracked", "rpn.backbone_net.Cross_Fusion.3.P2IA_Layer.conv1.0.weight", "rpn.backbone_net.Cross_Fusion.3.P2IA_Layer.conv1.0.bias", "rpn.backbone_net.Cross_Fusion.3.P2IA_Layer.conv1.1.weight", "rpn.backbone_net.Cross_Fusion.3.P2IA_Layer.conv1.1.bias", "rpn.backbone_net.Cross_Fusion.3.P2IA_Layer.conv1.1.running_mean", "rpn.backbone_net.Cross_Fusion.3.P2IA_Layer.conv1.1.running_var", "rpn.backbone_net.Cross_Fusion.3.P2IA_Layer.conv1.1.num_batches_tracked", "rpn.backbone_net.Cross_Fusion.3.P2IA_Layer.fc1.weight", "rpn.backbone_net.Cross_Fusion.3.P2IA_Layer.fc1.bias", "rpn.backbone_net.Cross_Fusion.3.P2IA_Layer.fc2.weight", "rpn.backbone_net.Cross_Fusion.3.P2IA_Layer.fc2.bias", "rpn.backbone_net.Cross_Fusion.3.P2IA_Layer.fc3.weight", "rpn.backbone_net.Cross_Fusion.3.P2IA_Layer.fc3.bias", "rpn.backbone_net.Cross_Fusion.3.conv1.weight", "rpn.backbone_net.Cross_Fusion.3.bn1.weight", "rpn.backbone_net.Cross_Fusion.3.bn1.bias", "rpn.backbone_net.Cross_Fusion.3.bn1.running_mean", "rpn.backbone_net.Cross_Fusion.3.bn1.running_var", "rpn.backbone_net.Cross_Fusion.3.bn1.num_batches_tracked".

[ Error 2 ] ##############USE Fusion_Cross_Conv_Gate(ADD)######### ##############ADDITION PI2 ATTENTION######### 2023-03-03 07:40:20,511 INFO ==> Loading from checkpoint './pre_trained_model/epnet_plus_plus_released_trained_models/CAR/checkpoint_epoch_43.pth' 2023-03-03 07:40:20,766 INFO ==> Done 2023-03-03 07:40:20,771 INFO ---- EPOCH 43 JOINT EVALUATION ---- 2023-03-03 07:40:20,771 INFO ==> Output file: ./epnet_plus_plus_released_trained_models/CAR/eval_results/eval/epoch_43/test/test_mode eval: 0%| | 0/7518 [00:00<?, ?it/s]Traceback (most recent call last): File "eval_rcnn.py", line 1026, in eval_single_ckpt(root_result_dir, data_path=args.data_path) File "eval_rcnn.py", line 868, in eval_single_ckpt eval_one_epoch(model, test_loader, epoch_id, root_result_dir, logger) File "eval_rcnn.py", line 790, in eval_one_epoch ret_dict = eval_one_epoch_joint(model, dataloader, epoch_id, result_dir, logger) File "eval_rcnn.py", line 526, in eval_one_epoch_joint for data in dataloader: File "/root/anaconda3/envs/epnet_2/lib/python3.8/site-packages/torch/utils/data/dataloader.py", line 435, in next data = self._next_data() File "/root/anaconda3/envs/epnet_2/lib/python3.8/site-packages/torch/utils/data/dataloader.py", line 1085, in _next_data return self._process_data(data) File "/root/anaconda3/envs/epnet_2/lib/python3.8/site-packages/torch/utils/data/dataloader.py", line 1111, in _process_data data.reraise() File "/root/anaconda3/envs/epnet_2/lib/python3.8/site-packages/torch/_utils.py", line 428, in reraise raise self.exc_type(msg) FileNotFoundError: Caught FileNotFoundError in DataLoader worker process 0. Original Traceback (most recent call last): File "/root/anaconda3/envs/epnet_2/lib/python3.8/site-packages/torch/utils/data/_utils/worker.py", line 198, in _worker_loop data = fetcher.fetch(index) File "/root/anaconda3/envs/epnet_2/lib/python3.8/site-packages/torch/utils/data/_utils/fetch.py", line 44, in fetch data = [self.dataset[idx] for idx in possibly_batched_index] File "/root/anaconda3/envs/epnet_2/lib/python3.8/site-packages/torch/utils/data/_utils/fetch.py", line 44, in data = [self.dataset[idx] for idx in possibly_batched_index] File "/workspace/EPNetV2/tools/../lib/datasets/kitti_rcnn_dataset.py", line 288, in getitem return self.get_rpn_with_li_fusion(index) File "/workspace/EPNetV2/tools/../lib/datasets/kitti_rcnn_dataset.py", line 317, in get_rpn_with_li_fusion img_seg_mask = self.get_KINS_car_mask(sample_id) File "/workspace/EPNetV2/tools/../lib/datasets/kitti_dataset.py", line 113, in get_KINS_car_mask cat_mask = np.load(os.path.join(self.mask_dir, '%06d.npy' % idx)) File "/root/anaconda3/envs/epnet_2/lib/python3.8/site-packages/numpy/lib/npyio.py", line 407, in load fid = stack.enter_context(open(os_fspath(file), "rb")) FileNotFoundError: [Errno 2] No such file or directory: '../data/KITTI/object/testing/train_mask/000000.npy'

[happinesslz]

@ssolchoi Sorry to reply too late! The command is following: CUDA_VISIBLE_DEVICES=0 python eval_rcnn.py --cfg_file cfgs/CAR_EPNet_plus_plus_for_test.yaml --eval_mode rcnn_online \ --output_dir ./epnet_plus_plus_released_trained_models/CAR/eval_results/ \ --data_path ../data/ --ckpt ./epnet_plus_plus_released_trained_models/CAR/checkpoint_epoch_43.pth --test \ --set LI_FUSION.ENABLED True LI_FUSION.ADD_Image_Attention True CROSS_FUSION True USE_P2I_GATE True \ DEEP_RCNN_FUSION False USE_IMAGE_LOSS True IMAGE_WEIGHT 1.0 USE_IMAGE_SCORE True

Please remember to modify: 1)TEST.SPLIT: (val ---> test) in configure file 2) add '--test' in the command 3) remove the line 317 "img_seg_mask = self.get_KINS_car_mask(sample_id)" in the file of 'kitti_rcnn_dataset.py' to the above line of 526.

[happinesslz]

@ssolchoi Note that we do not use 'test_mask' in EPNet++.

[faziii0]

@ssolchoi @happinesslz Can you help me how you upload results on kitti benchmark as testset have 7518 images and trainval set have 7480. do you submit it on testset or validation set. when i run on testset the evaultion it is all zero but on validation set it is showing 90% AP. WHAT CAN BE THE SOLUTION PLEASE HELP.

2023-12-14 18:12:10,189 INFO -------------------performance of epoch 44--------------------- 2023-12-14 18:12:10,189 INFO 2023-12-14 18:12:10.189209 2023-12-14 18:12:10,189 INFO final average detections: 3.861 2023-12-14 18:12:10,189 INFO final average rpn_iou refined: 0.000 2023-12-14 18:12:10,189 INFO final average cls acc: 0.000 2023-12-14 18:12:10,189 INFO final average cls acc refined: 0.000 2023-12-14 18:12:10,189 INFO total roi bbox recall(thresh=0.100): 0 / 0 = 0.000000 2023-12-14 18:12:10,189 INFO total roi bbox recall(thresh=0.300): 0 / 0 = 0.000000 2023-12-14 18:12:10,189 INFO total roi bbox recall(thresh=0.500): 0 / 0 = 0.000000 2023-12-14 18:12:10,189 INFO total roi bbox recall(thresh=0.700): 0 / 0 = 0.000000 2023-12-14 18:12:10,189 INFO total roi bbox recall(thresh=0.900): 0 / 0 = 0.000000 2023-12-14 18:12:10,189 INFO total bbox recall(thresh=0.100): 0 / 0 = 0.000000 2023-12-14 18:12:10,189 INFO total bbox recall(thresh=0.300): 0 / 0 = 0.000000 2023-12-14 18:12:10,189 INFO total bbox recall(thresh=0.500): 0 / 0 = 0.000000 2023-12-14 18:12:10,189 INFO total bbox recall(thresh=0.700): 0 / 0 = 0.000000 2023-12-14 18:12:10,189 INFO total bbox recall(thresh=0.900): 0 / 0 = 0.000000 2023-12-14 18:12:10,189 INFO result is saved to: ./epnet_plus_plus_released_trained_models/PED/eval_results/eval/epoch_44/test/test_mode

[AnRabbit]

@faziii0 Hello, I apologize for the interruption. May I inquire whether you have resolved the issue regarding the test set submission results showing as zero? If you have found a solution, could you kindly share the approach you took? Thank you, and I wish you all the best!

[faziii0]

@AnRabbit Because we don't have testset labels, thats why we have to upload our predicted labels for testset on to the kitti website to know the results