Open sherrygp opened 5 years ago
Sorry, I didn't test the codes on GPU2080. Mabye you should update to CUDA 10 as said here https://github.com/pytorch/pytorch/issues/17334.
@sshaoshuai
Hi,my computer evn is cuda10.1,cudnn 7.5 and GPU2080ti , pytorch1.0.1, do you know why this problem is caused?
python eval_rcnn.py --cfg_file cfgs/default.yaml --ckpt PointRCNN.pth --batch_size 1 --eval_mode rcnn --set RPN.LOC_XZ_FINE False
2019-05-05 16:40:05,047 INFO **Start logging**
2019-05-05 16:40:05,047 INFO cfg_file cfgs/default.yaml
2019-05-05 16:40:05,047 INFO eval_mode rcnn
2019-05-05 16:40:05,047 INFO eval_all False
2019-05-05 16:40:05,047 INFO test False
2019-05-05 16:40:05,047 INFO ckpt PointRCNN.pth
2019-05-05 16:40:05,047 INFO rpn_ckpt None
2019-05-05 16:40:05,047 INFO rcnn_ckpt None
2019-05-05 16:40:05,047 INFO batch_size 1
2019-05-05 16:40:05,047 INFO workers 4
2019-05-05 16:40:05,047 INFO extra_tag default
2019-05-05 16:40:05,047 INFO output_dir None
2019-05-05 16:40:05,047 INFO ckpt_dir None
2019-05-05 16:40:05,047 INFO save_result False
2019-05-05 16:40:05,048 INFO save_rpn_feature False
2019-05-05 16:40:05,048 INFO random_select True
2019-05-05 16:40:05,048 INFO start_epoch 0
2019-05-05 16:40:05,048 INFO rcnn_eval_roi_dir None
2019-05-05 16:40:05,048 INFO rcnn_eval_feature_dir None
2019-05-05 16:40:05,048 INFO set_cfgs ['RPN.LOC_XZ_FINE', 'False']
2019-05-05 16:40:05,048 INFO cfg.TAG: default
2019-05-05 16:40:05,048 INFO cfg.CLASSES: Car
2019-05-05 16:40:05,048 INFO cfg.INCLUDE_SIMILAR_TYPE: True
2019-05-05 16:40:05,048 INFO cfg.AUG_DATA: True
2019-05-05 16:40:05,048 INFO cfg.AUG_METHOD_LIST: ['rotation', 'scaling', 'flip']
2019-05-05 16:40:05,048 INFO cfg.AUG_METHOD_PROB: [1.0, 1.0, 0.5]
2019-05-05 16:40:05,048 INFO cfg.AUG_ROT_RANGE: 18
2019-05-05 16:40:05,048 INFO cfg.GT_AUG_ENABLED: True
2019-05-05 16:40:05,048 INFO cfg.GT_EXTRA_NUM: 15
2019-05-05 16:40:05,048 INFO cfg.GT_AUG_RAND_NUM: True
2019-05-05 16:40:05,049 INFO cfg.GT_AUG_APPLY_PROB: 1.0
2019-05-05 16:40:05,049 INFO cfg.GT_AUG_HARD_RATIO: 0.6
2019-05-05 16:40:05,049 INFO cfg.PC_REDUCE_BY_RANGE: True
2019-05-05 16:40:05,049 INFO cfg.PC_AREA_SCOPE: [[-40. 40. ]
[ -1. 3. ]
[ 0. 70.4]]
2019-05-05 16:40:05,049 INFO cfg.CLS_MEAN_SIZE: [[1.5256319 1.6285675 3.8831165]]
2019-05-05 16:40:05,049 INFO
cfg.RPN = edict()
2019-05-05 16:40:05,049 INFO cfg.RPN.ENABLED: True
2019-05-05 16:40:05,049 INFO cfg.RPN.FIXED: True
2019-05-05 16:40:05,049 INFO cfg.RPN.USE_INTENSITY: False
2019-05-05 16:40:05,049 INFO cfg.RPN.LOC_XZ_FINE: False
2019-05-05 16:40:05,049 INFO cfg.RPN.LOC_SCOPE: 3.0
2019-05-05 16:40:05,050 INFO cfg.RPN.LOC_BIN_SIZE: 0.5
2019-05-05 16:40:05,050 INFO cfg.RPN.NUM_HEAD_BIN: 12
2019-05-05 16:40:05,050 INFO cfg.RPN.BACKBONE: pointnet2_msg
2019-05-05 16:40:05,050 INFO cfg.RPN.USE_BN: True
2019-05-05 16:40:05,050 INFO cfg.RPN.NUM_POINTS: 16384
2019-05-05 16:40:05,050 INFO
cfg.RPN.SA_CONFIG = edict()
2019-05-05 16:40:05,050 INFO cfg.RPN.SA_CONFIG.NPOINTS: [4096, 1024, 256, 64]
2019-05-05 16:40:05,050 INFO cfg.RPN.SA_CONFIG.RADIUS: [[0.1, 0.5], [0.5, 1.0], [1.0, 2.0], [2.0, 4.0]]
2019-05-05 16:40:05,050 INFO cfg.RPN.SA_CONFIG.NSAMPLE: [[16, 32], [16, 32], [16, 32], [16, 32]]
2019-05-05 16:40:05,050 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]]]
2019-05-05 16:40:05,050 INFO cfg.RPN.FP_MLPS: [[128, 128], [256, 256], [512, 512], [512, 512]]
2019-05-05 16:40:05,050 INFO cfg.RPN.CLS_FC: [128]
2019-05-05 16:40:05,050 INFO cfg.RPN.REG_FC: [128]
2019-05-05 16:40:05,050 INFO cfg.RPN.DP_RATIO: 0.5
2019-05-05 16:40:05,050 INFO cfg.RPN.LOSS_CLS: SigmoidFocalLoss
2019-05-05 16:40:05,050 INFO cfg.RPN.FG_WEIGHT: 15
2019-05-05 16:40:05,051 INFO cfg.RPN.FOCAL_ALPHA: [0.25, 0.75]
2019-05-05 16:40:05,051 INFO cfg.RPN.FOCAL_GAMMA: 2.0
2019-05-05 16:40:05,051 INFO cfg.RPN.REG_LOSS_WEIGHT: [1.0, 1.0, 1.0, 1.0]
2019-05-05 16:40:05,051 INFO cfg.RPN.LOSS_WEIGHT: [1.0, 1.0]
2019-05-05 16:40:05,051 INFO cfg.RPN.NMS_TYPE: normal
2019-05-05 16:40:05,051 INFO cfg.RPN.SCORE_THRESH: 0.3
2019-05-05 16:40:05,051 INFO
cfg.RCNN = edict()
2019-05-05 16:40:05,051 INFO cfg.RCNN.ENABLED: True
2019-05-05 16:40:05,051 INFO cfg.RCNN.USE_RPN_FEATURES: True
2019-05-05 16:40:05,051 INFO cfg.RCNN.USE_MASK: True
2019-05-05 16:40:05,051 INFO cfg.RCNN.MASK_TYPE: seg
2019-05-05 16:40:05,051 INFO cfg.RCNN.USE_INTENSITY: False
2019-05-05 16:40:05,051 INFO cfg.RCNN.USE_DEPTH: True
2019-05-05 16:40:05,051 INFO cfg.RCNN.USE_SEG_SCORE: False
2019-05-05 16:40:05,051 INFO cfg.RCNN.ROI_SAMPLE_JIT: True
2019-05-05 16:40:05,051 INFO cfg.RCNN.ROI_FG_AUG_TIMES: 10
2019-05-05 16:40:05,052 INFO cfg.RCNN.REG_AUG_METHOD: multiple
2019-05-05 16:40:05,052 INFO cfg.RCNN.POOL_EXTRA_WIDTH: 1.0
2019-05-05 16:40:05,052 INFO cfg.RCNN.LOC_SCOPE: 1.5
2019-05-05 16:40:05,052 INFO cfg.RCNN.LOC_BIN_SIZE: 0.5
2019-05-05 16:40:05,052 INFO cfg.RCNN.NUM_HEAD_BIN: 9
2019-05-05 16:40:05,052 INFO cfg.RCNN.LOC_Y_BY_BIN: False
2019-05-05 16:40:05,052 INFO cfg.RCNN.LOC_Y_SCOPE: 0.5
2019-05-05 16:40:05,052 INFO cfg.RCNN.LOC_Y_BIN_SIZE: 0.25
2019-05-05 16:40:05,052 INFO cfg.RCNN.SIZE_RES_ON_ROI: False
2019-05-05 16:40:05,052 INFO cfg.RCNN.USE_BN: False
2019-05-05 16:40:05,052 INFO cfg.RCNN.DP_RATIO: 0.0
2019-05-05 16:40:05,052 INFO cfg.RCNN.BACKBONE: pointnet
2019-05-05 16:40:05,052 INFO cfg.RCNN.XYZ_UP_LAYER: [128, 128]
2019-05-05 16:40:05,052 INFO cfg.RCNN.NUM_POINTS: 512
2019-05-05 16:40:05,052 INFO
cfg.RCNN.SA_CONFIG = edict()
2019-05-05 16:40:05,052 INFO cfg.RCNN.SA_CONFIG.NPOINTS: [128, 32, -1]
2019-05-05 16:40:05,052 INFO cfg.RCNN.SA_CONFIG.RADIUS: [0.2, 0.4, 100]
2019-05-05 16:40:05,053 INFO cfg.RCNN.SA_CONFIG.NSAMPLE: [64, 64, 64]
2019-05-05 16:40:05,053 INFO cfg.RCNN.SA_CONFIG.MLPS: [[128, 128, 128], [128, 128, 256], [256, 256, 512]]
2019-05-05 16:40:05,053 INFO cfg.RCNN.CLS_FC: [256, 256]
2019-05-05 16:40:05,053 INFO cfg.RCNN.REG_FC: [256, 256]
2019-05-05 16:40:05,053 INFO cfg.RCNN.LOSS_CLS: BinaryCrossEntropy
2019-05-05 16:40:05,053 INFO cfg.RCNN.FOCAL_ALPHA: [0.25, 0.75]
2019-05-05 16:40:05,053 INFO cfg.RCNN.FOCAL_GAMMA: 2.0
2019-05-05 16:40:05,053 INFO cfg.RCNN.CLS_WEIGHT: [1. 1. 1.]
2019-05-05 16:40:05,053 INFO cfg.RCNN.CLS_FG_THRESH: 0.6
2019-05-05 16:40:05,053 INFO cfg.RCNN.CLS_BG_THRESH: 0.45
2019-05-05 16:40:05,053 INFO cfg.RCNN.CLS_BG_THRESH_LO: 0.05
2019-05-05 16:40:05,053 INFO cfg.RCNN.REG_FG_THRESH: 0.55
2019-05-05 16:40:05,053 INFO cfg.RCNN.FG_RATIO: 0.5
2019-05-05 16:40:05,053 INFO cfg.RCNN.ROI_PER_IMAGE: 64
2019-05-05 16:40:05,054 INFO cfg.RCNN.HARD_BG_RATIO: 0.8
2019-05-05 16:40:05,054 INFO cfg.RCNN.SCORE_THRESH: 0.3
2019-05-05 16:40:05,054 INFO cfg.RCNN.NMS_THRESH: 0.1
2019-05-05 16:40:05,054 INFO
cfg.TRAIN = edict()
2019-05-05 16:40:05,054 INFO cfg.TRAIN.SPLIT: train
2019-05-05 16:40:05,054 INFO cfg.TRAIN.VAL_SPLIT: smallval
2019-05-05 16:40:05,054 INFO cfg.TRAIN.LR: 0.002
2019-05-05 16:40:05,054 INFO cfg.TRAIN.LR_CLIP: 1e-05
2019-05-05 16:40:05,054 INFO cfg.TRAIN.LR_DECAY: 0.5
2019-05-05 16:40:05,054 INFO cfg.TRAIN.DECAY_STEP_LIST: [100, 150, 180, 200]
2019-05-05 16:40:05,054 INFO cfg.TRAIN.LR_WARMUP: True
2019-05-05 16:40:05,054 INFO cfg.TRAIN.WARMUP_MIN: 0.0002
2019-05-05 16:40:05,054 INFO cfg.TRAIN.WARMUP_EPOCH: 1
2019-05-05 16:40:05,054 INFO cfg.TRAIN.BN_MOMENTUM: 0.1
2019-05-05 16:40:05,054 INFO cfg.TRAIN.BN_DECAY: 0.5
2019-05-05 16:40:05,054 INFO cfg.TRAIN.BNM_CLIP: 0.01
2019-05-05 16:40:05,054 INFO cfg.TRAIN.BN_DECAY_STEP_LIST: [1000]
2019-05-05 16:40:05,055 INFO cfg.TRAIN.OPTIMIZER: adam_onecycle
2019-05-05 16:40:05,055 INFO cfg.TRAIN.WEIGHT_DECAY: 0.001
2019-05-05 16:40:05,055 INFO cfg.TRAIN.MOMENTUM: 0.9
2019-05-05 16:40:05,055 INFO cfg.TRAIN.MOMS: [0.95, 0.85]
2019-05-05 16:40:05,055 INFO cfg.TRAIN.DIV_FACTOR: 10.0
2019-05-05 16:40:05,055 INFO cfg.TRAIN.PCT_START: 0.4
2019-05-05 16:40:05,055 INFO cfg.TRAIN.GRAD_NORM_CLIP: 1.0
2019-05-05 16:40:05,055 INFO cfg.TRAIN.RPN_PRE_NMS_TOP_N: 9000
2019-05-05 16:40:05,055 INFO cfg.TRAIN.RPN_POST_NMS_TOP_N: 512
2019-05-05 16:40:05,055 INFO cfg.TRAIN.RPN_NMS_THRESH: 0.85
2019-05-05 16:40:05,055 INFO cfg.TRAIN.RPN_DISTANCE_BASED_PROPOSE: True
2019-05-05 16:40:05,055 INFO
cfg.TEST = edict()
2019-05-05 16:40:05,055 INFO cfg.TEST.SPLIT: val
2019-05-05 16:40:05,055 INFO cfg.TEST.RPN_PRE_NMS_TOP_N: 9000
2019-05-05 16:40:05,055 INFO cfg.TEST.RPN_POST_NMS_TOP_N: 100
2019-05-05 16:40:05,055 INFO cfg.TEST.RPN_NMS_THRESH: 0.8
2019-05-05 16:40:05,056 INFO cfg.TEST.RPN_DISTANCE_BASED_PROPOSE: True
2019-05-05 16:40:05,057 INFO Load testing samples from ../data/KITTI/object/training
2019-05-05 16:40:05,057 INFO Done: total test samples 3769
2019-05-05 16:40:07,559 INFO ==> Loading from checkpoint 'PointRCNN.pth'
2019-05-05 16:40:07,618 INFO ==> Done
2019-05-05 16:40:07,620 INFO ---- EPOCH no_number JOINT EVALUATION ----
2019-05-05 16:40:07,620 INFO ==> Output file: ../output/rcnn/default/eval/epoch_no_number/val
eval: 0%| | 0/3769 [00:00<?, ?it/s]CUDA kernel failed : a PTX JIT compilation failed
Sorry, I didn't have a GPU2080ti and I never saw this error before...
Hello, I am not able to run eval_rcnn.py I keep having this error
python eval_rcnn.py --cfg_file cfgs/default.yaml --ckpt PointRCNN.pth --batch_size 1 --eval_mode rcnn --set RPN.LOC_XZ_FINE False
2019-05-28 12:34:31,425 INFO **Start logging**
2019-05-28 12:34:31,425 INFO cfg_file cfgs/default.yaml
2019-05-28 12:34:31,425 INFO eval_mode rcnn
2019-05-28 12:34:31,425 INFO eval_all False
2019-05-28 12:34:31,425 INFO test False
2019-05-28 12:34:31,425 INFO ckpt pretrained_model/PointRCNN.pth
2019-05-28 12:34:31,425 INFO rpn_ckpt None
2019-05-28 12:34:31,425 INFO rcnn_ckpt None
2019-05-28 12:34:31,425 INFO batch_size 1
2019-05-28 12:34:31,425 INFO workers 4
2019-05-28 12:34:31,425 INFO extra_tag default
2019-05-28 12:34:31,426 INFO output_dir None
2019-05-28 12:34:31,426 INFO ckpt_dir None
2019-05-28 12:34:31,426 INFO save_result False
2019-05-28 12:34:31,426 INFO save_rpn_feature False
2019-05-28 12:34:31,426 INFO random_select True
2019-05-28 12:34:31,426 INFO start_epoch 0
2019-05-28 12:34:31,426 INFO rcnn_eval_roi_dir None
2019-05-28 12:34:31,426 INFO rcnn_eval_feature_dir None
2019-05-28 12:34:31,426 INFO set_cfgs ['RPN.LOC_XZ_FINE', 'False']
2019-05-28 12:34:31,426 INFO cfg.TAG: default
2019-05-28 12:34:31,426 INFO cfg.CLASSES: Car
2019-05-28 12:34:31,426 INFO cfg.INCLUDE_SIMILAR_TYPE: True
2019-05-28 12:34:31,426 INFO cfg.AUG_DATA: True
2019-05-28 12:34:31,426 INFO cfg.AUG_METHOD_LIST: ['rotation', 'scaling', 'flip']
2019-05-28 12:34:31,426 INFO cfg.AUG_METHOD_PROB: [1.0, 1.0, 0.5]
2019-05-28 12:34:31,426 INFO cfg.AUG_ROT_RANGE: 18
2019-05-28 12:34:31,426 INFO cfg.GT_AUG_ENABLED: True
2019-05-28 12:34:31,426 INFO cfg.GT_EXTRA_NUM: 15
2019-05-28 12:34:31,426 INFO cfg.GT_AUG_RAND_NUM: True
2019-05-28 12:34:31,426 INFO cfg.GT_AUG_APPLY_PROB: 1.0
2019-05-28 12:34:31,426 INFO cfg.GT_AUG_HARD_RATIO: 0.6
2019-05-28 12:34:31,426 INFO cfg.PC_REDUCE_BY_RANGE: True
2019-05-28 12:34:31,427 INFO cfg.PC_AREA_SCOPE: [[-40. 40. ]
[ -1. 3. ]
[ 0. 70.4]]
2019-05-28 12:34:31,427 INFO cfg.CLS_MEAN_SIZE: [[1.5256319 1.6285675 3.8831165]]
2019-05-28 12:34:31,427 INFO
cfg.RPN = edict()
2019-05-28 12:34:31,427 INFO cfg.RPN.ENABLED: True
2019-05-28 12:34:31,427 INFO cfg.RPN.FIXED: True
2019-05-28 12:34:31,427 INFO cfg.RPN.USE_INTENSITY: False
2019-05-28 12:34:31,427 INFO cfg.RPN.LOC_XZ_FINE: False
2019-05-28 12:34:31,427 INFO cfg.RPN.LOC_SCOPE: 3.0
2019-05-28 12:34:31,427 INFO cfg.RPN.LOC_BIN_SIZE: 0.5
2019-05-28 12:34:31,427 INFO cfg.RPN.NUM_HEAD_BIN: 12
2019-05-28 12:34:31,427 INFO cfg.RPN.BACKBONE: pointnet2_msg
2019-05-28 12:34:31,427 INFO cfg.RPN.USE_BN: True
2019-05-28 12:34:31,427 INFO cfg.RPN.NUM_POINTS: 16384
2019-05-28 12:34:31,427 INFO
cfg.RPN.SA_CONFIG = edict()
2019-05-28 12:34:31,427 INFO cfg.RPN.SA_CONFIG.NPOINTS: [4096, 1024, 256, 64]
2019-05-28 12:34:31,427 INFO cfg.RPN.SA_CONFIG.RADIUS: [[0.1, 0.5], [0.5, 1.0], [1.0, 2.0], [2.0, 4.0]]
2019-05-28 12:34:31,427 INFO cfg.RPN.SA_CONFIG.NSAMPLE: [[16, 32], [16, 32], [16, 32], [16, 32]]
2019-05-28 12:34:31,427 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]]]
2019-05-28 12:34:31,428 INFO cfg.RPN.FP_MLPS: [[128, 128], [256, 256], [512, 512], [512, 512]]
2019-05-28 12:34:31,428 INFO cfg.RPN.CLS_FC: [128]
2019-05-28 12:34:31,428 INFO cfg.RPN.REG_FC: [128]
2019-05-28 12:34:31,428 INFO cfg.RPN.DP_RATIO: 0.5
2019-05-28 12:34:31,428 INFO cfg.RPN.LOSS_CLS: SigmoidFocalLoss
2019-05-28 12:34:31,428 INFO cfg.RPN.FG_WEIGHT: 15
2019-05-28 12:34:31,428 INFO cfg.RPN.FOCAL_ALPHA: [0.25, 0.75]
2019-05-28 12:34:31,428 INFO cfg.RPN.FOCAL_GAMMA: 2.0
2019-05-28 12:34:31,428 INFO cfg.RPN.REG_LOSS_WEIGHT: [1.0, 1.0, 1.0, 1.0]
2019-05-28 12:34:31,428 INFO cfg.RPN.LOSS_WEIGHT: [1.0, 1.0]
2019-05-28 12:34:31,428 INFO cfg.RPN.NMS_TYPE: normal
2019-05-28 12:34:31,428 INFO cfg.RPN.SCORE_THRESH: 0.3
2019-05-28 12:34:31,428 INFO
cfg.RCNN = edict()
2019-05-28 12:34:31,428 INFO cfg.RCNN.ENABLED: True
2019-05-28 12:34:31,428 INFO cfg.RCNN.USE_RPN_FEATURES: True
2019-05-28 12:34:31,428 INFO cfg.RCNN.USE_MASK: True
2019-05-28 12:34:31,428 INFO cfg.RCNN.MASK_TYPE: seg
2019-05-28 12:34:31,428 INFO cfg.RCNN.USE_INTENSITY: False
2019-05-28 12:34:31,428 INFO cfg.RCNN.USE_DEPTH: True
2019-05-28 12:34:31,428 INFO cfg.RCNN.USE_SEG_SCORE: False
2019-05-28 12:34:31,428 INFO cfg.RCNN.ROI_SAMPLE_JIT: True
2019-05-28 12:34:31,428 INFO cfg.RCNN.ROI_FG_AUG_TIMES: 10
2019-05-28 12:34:31,428 INFO cfg.RCNN.REG_AUG_METHOD: multiple
2019-05-28 12:34:31,428 INFO cfg.RCNN.POOL_EXTRA_WIDTH: 1.0
2019-05-28 12:34:31,428 INFO cfg.RCNN.LOC_SCOPE: 1.5
2019-05-28 12:34:31,429 INFO cfg.RCNN.LOC_BIN_SIZE: 0.5
2019-05-28 12:34:31,429 INFO cfg.RCNN.NUM_HEAD_BIN: 9
2019-05-28 12:34:31,429 INFO cfg.RCNN.LOC_Y_BY_BIN: False
2019-05-28 12:34:31,429 INFO cfg.RCNN.LOC_Y_SCOPE: 0.5
2019-05-28 12:34:31,429 INFO cfg.RCNN.LOC_Y_BIN_SIZE: 0.25
2019-05-28 12:34:31,429 INFO cfg.RCNN.SIZE_RES_ON_ROI: False
2019-05-28 12:34:31,429 INFO cfg.RCNN.USE_BN: False
2019-05-28 12:34:31,429 INFO cfg.RCNN.DP_RATIO: 0.0
2019-05-28 12:34:31,429 INFO cfg.RCNN.BACKBONE: pointnet
2019-05-28 12:34:31,429 INFO cfg.RCNN.XYZ_UP_LAYER: [128, 128]
2019-05-28 12:34:31,429 INFO cfg.RCNN.NUM_POINTS: 512
2019-05-28 12:34:31,429 INFO
cfg.RCNN.SA_CONFIG = edict()
2019-05-28 12:34:31,429 INFO cfg.RCNN.SA_CONFIG.NPOINTS: [128, 32, -1]
2019-05-28 12:34:31,429 INFO cfg.RCNN.SA_CONFIG.RADIUS: [0.2, 0.4, 100]
2019-05-28 12:34:31,429 INFO cfg.RCNN.SA_CONFIG.NSAMPLE: [64, 64, 64]
2019-05-28 12:34:31,429 INFO cfg.RCNN.SA_CONFIG.MLPS: [[128, 128, 128], [128, 128, 256], [256, 256, 512]]
2019-05-28 12:34:31,429 INFO cfg.RCNN.CLS_FC: [256, 256]
2019-05-28 12:34:31,429 INFO cfg.RCNN.REG_FC: [256, 256]
2019-05-28 12:34:31,429 INFO cfg.RCNN.LOSS_CLS: BinaryCrossEntropy
2019-05-28 12:34:31,429 INFO cfg.RCNN.FOCAL_ALPHA: [0.25, 0.75]
2019-05-28 12:34:31,429 INFO cfg.RCNN.FOCAL_GAMMA: 2.0
2019-05-28 12:34:31,429 INFO cfg.RCNN.CLS_WEIGHT: [1. 1. 1.]
2019-05-28 12:34:31,430 INFO cfg.RCNN.CLS_FG_THRESH: 0.6
2019-05-28 12:34:31,430 INFO cfg.RCNN.CLS_BG_THRESH: 0.45
2019-05-28 12:34:31,430 INFO cfg.RCNN.CLS_BG_THRESH_LO: 0.05
2019-05-28 12:34:31,430 INFO cfg.RCNN.REG_FG_THRESH: 0.55
2019-05-28 12:34:31,430 INFO cfg.RCNN.FG_RATIO: 0.5
2019-05-28 12:34:31,430 INFO cfg.RCNN.ROI_PER_IMAGE: 64
2019-05-28 12:34:31,430 INFO cfg.RCNN.HARD_BG_RATIO: 0.8
2019-05-28 12:34:31,430 INFO cfg.RCNN.SCORE_THRESH: 0.3
2019-05-28 12:34:31,430 INFO cfg.RCNN.NMS_THRESH: 0.1
2019-05-28 12:34:31,430 INFO
cfg.TRAIN = edict()
2019-05-28 12:34:31,430 INFO cfg.TRAIN.SPLIT: train
2019-05-28 12:34:31,430 INFO cfg.TRAIN.VAL_SPLIT: smallval
2019-05-28 12:34:31,430 INFO cfg.TRAIN.LR: 0.002
2019-05-28 12:34:31,430 INFO cfg.TRAIN.LR_CLIP: 1e-05
2019-05-28 12:34:31,430 INFO cfg.TRAIN.LR_DECAY: 0.5
2019-05-28 12:34:31,430 INFO cfg.TRAIN.DECAY_STEP_LIST: [100, 150, 180, 200]
2019-05-28 12:34:31,430 INFO cfg.TRAIN.LR_WARMUP: True
2019-05-28 12:34:31,430 INFO cfg.TRAIN.WARMUP_MIN: 0.0002
2019-05-28 12:34:31,430 INFO cfg.TRAIN.WARMUP_EPOCH: 1
2019-05-28 12:34:31,430 INFO cfg.TRAIN.BN_MOMENTUM: 0.1
2019-05-28 12:34:31,430 INFO cfg.TRAIN.BN_DECAY: 0.5
2019-05-28 12:34:31,430 INFO cfg.TRAIN.BNM_CLIP: 0.01
2019-05-28 12:34:31,430 INFO cfg.TRAIN.BN_DECAY_STEP_LIST: [1000]
2019-05-28 12:34:31,430 INFO cfg.TRAIN.OPTIMIZER: adam_onecycle
2019-05-28 12:34:31,430 INFO cfg.TRAIN.WEIGHT_DECAY: 0.001
2019-05-28 12:34:31,431 INFO cfg.TRAIN.MOMENTUM: 0.9
2019-05-28 12:34:31,431 INFO cfg.TRAIN.MOMS: [0.95, 0.85]
2019-05-28 12:34:31,431 INFO cfg.TRAIN.DIV_FACTOR: 10.0
2019-05-28 12:34:31,431 INFO cfg.TRAIN.PCT_START: 0.4
2019-05-28 12:34:31,431 INFO cfg.TRAIN.GRAD_NORM_CLIP: 1.0
2019-05-28 12:34:31,431 INFO cfg.TRAIN.RPN_PRE_NMS_TOP_N: 9000
2019-05-28 12:34:31,431 INFO cfg.TRAIN.RPN_POST_NMS_TOP_N: 512
2019-05-28 12:34:31,431 INFO cfg.TRAIN.RPN_NMS_THRESH: 0.85
2019-05-28 12:34:31,431 INFO cfg.TRAIN.RPN_DISTANCE_BASED_PROPOSE: True
2019-05-28 12:34:31,431 INFO
cfg.TEST = edict()
2019-05-28 12:34:31,431 INFO cfg.TEST.SPLIT: val
2019-05-28 12:34:31,431 INFO cfg.TEST.RPN_PRE_NMS_TOP_N: 9000
2019-05-28 12:34:31,431 INFO cfg.TEST.RPN_POST_NMS_TOP_N: 100
2019-05-28 12:34:31,431 INFO cfg.TEST.RPN_NMS_THRESH: 0.8
2019-05-28 12:34:31,431 INFO cfg.TEST.RPN_DISTANCE_BASED_PROPOSE: True
2019-05-28 12:34:31,432 INFO Load testing samples from ../data/KITTI/object/training
2019-05-28 12:34:31,432 INFO Done: total test samples 3769
2019-05-28 12:34:33,657 INFO ==> Loading from checkpoint 'pretrained_model/PointRCNN.pth'
2019-05-28 12:34:33,705 INFO ==> Done
2019-05-28 12:34:33,707 INFO ---- EPOCH no_number JOINT EVALUATION ----
2019-05-28 12:34:33,707 INFO ==> Output file: ../output/rcnn/default/eval/epoch_no_number/val
eval: 0%| | 0/3769 [00:00<?, ?it/s]Traceback (most recent call last):
File "eval_rcnn.py", line 902, in
Thanks in advance for the reply, I use gtx 1070 and cuda 9.0 cudnn 7.2.1
Hello, I am not able to run eval_rcnn.py I keep having this error
python eval_rcnn.py --cfg_file cfgs/default.yaml --ckpt PointRCNN.pth --batch_size 1 --eval_mode rcnn --set RPN.LOC_XZ_FINE False 2019-05-28 12:34:31,425 INFO **Start logging** 2019-05-28 12:34:31,425 INFO cfg_file cfgs/default.yaml 2019-05-28 12:34:31,425 INFO eval_mode rcnn 2019-05-28 12:34:31,425 INFO eval_all False 2019-05-28 12:34:31,425 INFO test False 2019-05-28 12:34:31,425 INFO ckpt pretrained_model/PointRCNN.pth 2019-05-28 12:34:31,425 INFO rpn_ckpt None 2019-05-28 12:34:31,425 INFO rcnn_ckpt None 2019-05-28 12:34:31,425 INFO batch_size 1 2019-05-28 12:34:31,425 INFO workers 4 2019-05-28 12:34:31,425 INFO extra_tag default 2019-05-28 12:34:31,426 INFO output_dir None 2019-05-28 12:34:31,426 INFO ckpt_dir None 2019-05-28 12:34:31,426 INFO save_result False 2019-05-28 12:34:31,426 INFO save_rpn_feature False 2019-05-28 12:34:31,426 INFO random_select True 2019-05-28 12:34:31,426 INFO start_epoch 0 2019-05-28 12:34:31,426 INFO rcnn_eval_roi_dir None 2019-05-28 12:34:31,426 INFO rcnn_eval_feature_dir None 2019-05-28 12:34:31,426 INFO set_cfgs ['RPN.LOC_XZ_FINE', 'False'] 2019-05-28 12:34:31,426 INFO cfg.TAG: default 2019-05-28 12:34:31,426 INFO cfg.CLASSES: Car 2019-05-28 12:34:31,426 INFO cfg.INCLUDE_SIMILAR_TYPE: True 2019-05-28 12:34:31,426 INFO cfg.AUG_DATA: True 2019-05-28 12:34:31,426 INFO cfg.AUG_METHOD_LIST: ['rotation', 'scaling', 'flip'] 2019-05-28 12:34:31,426 INFO cfg.AUG_METHOD_PROB: [1.0, 1.0, 0.5] 2019-05-28 12:34:31,426 INFO cfg.AUG_ROT_RANGE: 18 2019-05-28 12:34:31,426 INFO cfg.GT_AUG_ENABLED: True 2019-05-28 12:34:31,426 INFO cfg.GT_EXTRA_NUM: 15 2019-05-28 12:34:31,426 INFO cfg.GT_AUG_RAND_NUM: True 2019-05-28 12:34:31,426 INFO cfg.GT_AUG_APPLY_PROB: 1.0 2019-05-28 12:34:31,426 INFO cfg.GT_AUG_HARD_RATIO: 0.6 2019-05-28 12:34:31,426 INFO cfg.PC_REDUCE_BY_RANGE: True 2019-05-28 12:34:31,427 INFO cfg.PC_AREA_SCOPE: [[-40. 40. ] [ -1. 3. ] [ 0. 70.4]] 2019-05-28 12:34:31,427 INFO cfg.CLS_MEAN_SIZE: [[1.5256319 1.6285675 3.8831165]] 2019-05-28 12:34:31,427 INFO cfg.RPN = edict() 2019-05-28 12:34:31,427 INFO cfg.RPN.ENABLED: True 2019-05-28 12:34:31,427 INFO cfg.RPN.FIXED: True 2019-05-28 12:34:31,427 INFO cfg.RPN.USE_INTENSITY: False 2019-05-28 12:34:31,427 INFO cfg.RPN.LOC_XZ_FINE: False 2019-05-28 12:34:31,427 INFO cfg.RPN.LOC_SCOPE: 3.0 2019-05-28 12:34:31,427 INFO cfg.RPN.LOC_BIN_SIZE: 0.5 2019-05-28 12:34:31,427 INFO cfg.RPN.NUM_HEAD_BIN: 12 2019-05-28 12:34:31,427 INFO cfg.RPN.BACKBONE: pointnet2_msg 2019-05-28 12:34:31,427 INFO cfg.RPN.USE_BN: True 2019-05-28 12:34:31,427 INFO cfg.RPN.NUM_POINTS: 16384 2019-05-28 12:34:31,427 INFO cfg.RPN.SA_CONFIG = edict() 2019-05-28 12:34:31,427 INFO cfg.RPN.SA_CONFIG.NPOINTS: [4096, 1024, 256, 64] 2019-05-28 12:34:31,427 INFO cfg.RPN.SA_CONFIG.RADIUS: [[0.1, 0.5], [0.5, 1.0], [1.0, 2.0], [2.0, 4.0]] 2019-05-28 12:34:31,427 INFO cfg.RPN.SA_CONFIG.NSAMPLE: [[16, 32], [16, 32], [16, 32], [16, 32]] 2019-05-28 12:34:31,427 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]]] 2019-05-28 12:34:31,428 INFO cfg.RPN.FP_MLPS: [[128, 128], [256, 256], [512, 512], [512, 512]] 2019-05-28 12:34:31,428 INFO cfg.RPN.CLS_FC: [128] 2019-05-28 12:34:31,428 INFO cfg.RPN.REG_FC: [128] 2019-05-28 12:34:31,428 INFO cfg.RPN.DP_RATIO: 0.5 2019-05-28 12:34:31,428 INFO cfg.RPN.LOSS_CLS: SigmoidFocalLoss 2019-05-28 12:34:31,428 INFO cfg.RPN.FG_WEIGHT: 15 2019-05-28 12:34:31,428 INFO cfg.RPN.FOCAL_ALPHA: [0.25, 0.75] 2019-05-28 12:34:31,428 INFO cfg.RPN.FOCAL_GAMMA: 2.0 2019-05-28 12:34:31,428 INFO cfg.RPN.REG_LOSS_WEIGHT: [1.0, 1.0, 1.0, 1.0] 2019-05-28 12:34:31,428 INFO cfg.RPN.LOSS_WEIGHT: [1.0, 1.0] 2019-05-28 12:34:31,428 INFO cfg.RPN.NMS_TYPE: normal 2019-05-28 12:34:31,428 INFO cfg.RPN.SCORE_THRESH: 0.3 2019-05-28 12:34:31,428 INFO cfg.RCNN = edict() 2019-05-28 12:34:31,428 INFO cfg.RCNN.ENABLED: True 2019-05-28 12:34:31,428 INFO cfg.RCNN.USE_RPN_FEATURES: True 2019-05-28 12:34:31,428 INFO cfg.RCNN.USE_MASK: True 2019-05-28 12:34:31,428 INFO cfg.RCNN.MASK_TYPE: seg 2019-05-28 12:34:31,428 INFO cfg.RCNN.USE_INTENSITY: False 2019-05-28 12:34:31,428 INFO cfg.RCNN.USE_DEPTH: True 2019-05-28 12:34:31,428 INFO cfg.RCNN.USE_SEG_SCORE: False 2019-05-28 12:34:31,428 INFO cfg.RCNN.ROI_SAMPLE_JIT: True 2019-05-28 12:34:31,428 INFO cfg.RCNN.ROI_FG_AUG_TIMES: 10 2019-05-28 12:34:31,428 INFO cfg.RCNN.REG_AUG_METHOD: multiple 2019-05-28 12:34:31,428 INFO cfg.RCNN.POOL_EXTRA_WIDTH: 1.0 2019-05-28 12:34:31,428 INFO cfg.RCNN.LOC_SCOPE: 1.5 2019-05-28 12:34:31,429 INFO cfg.RCNN.LOC_BIN_SIZE: 0.5 2019-05-28 12:34:31,429 INFO cfg.RCNN.NUM_HEAD_BIN: 9 2019-05-28 12:34:31,429 INFO cfg.RCNN.LOC_Y_BY_BIN: False 2019-05-28 12:34:31,429 INFO cfg.RCNN.LOC_Y_SCOPE: 0.5 2019-05-28 12:34:31,429 INFO cfg.RCNN.LOC_Y_BIN_SIZE: 0.25 2019-05-28 12:34:31,429 INFO cfg.RCNN.SIZE_RES_ON_ROI: False 2019-05-28 12:34:31,429 INFO cfg.RCNN.USE_BN: False 2019-05-28 12:34:31,429 INFO cfg.RCNN.DP_RATIO: 0.0 2019-05-28 12:34:31,429 INFO cfg.RCNN.BACKBONE: pointnet 2019-05-28 12:34:31,429 INFO cfg.RCNN.XYZ_UP_LAYER: [128, 128] 2019-05-28 12:34:31,429 INFO cfg.RCNN.NUM_POINTS: 512 2019-05-28 12:34:31,429 INFO cfg.RCNN.SA_CONFIG = edict() 2019-05-28 12:34:31,429 INFO cfg.RCNN.SA_CONFIG.NPOINTS: [128, 32, -1] 2019-05-28 12:34:31,429 INFO cfg.RCNN.SA_CONFIG.RADIUS: [0.2, 0.4, 100] 2019-05-28 12:34:31,429 INFO cfg.RCNN.SA_CONFIG.NSAMPLE: [64, 64, 64] 2019-05-28 12:34:31,429 INFO cfg.RCNN.SA_CONFIG.MLPS: [[128, 128, 128], [128, 128, 256], [256, 256, 512]] 2019-05-28 12:34:31,429 INFO cfg.RCNN.CLS_FC: [256, 256] 2019-05-28 12:34:31,429 INFO cfg.RCNN.REG_FC: [256, 256] 2019-05-28 12:34:31,429 INFO cfg.RCNN.LOSS_CLS: BinaryCrossEntropy 2019-05-28 12:34:31,429 INFO cfg.RCNN.FOCAL_ALPHA: [0.25, 0.75] 2019-05-28 12:34:31,429 INFO cfg.RCNN.FOCAL_GAMMA: 2.0 2019-05-28 12:34:31,429 INFO cfg.RCNN.CLS_WEIGHT: [1. 1. 1.] 2019-05-28 12:34:31,430 INFO cfg.RCNN.CLS_FG_THRESH: 0.6 2019-05-28 12:34:31,430 INFO cfg.RCNN.CLS_BG_THRESH: 0.45 2019-05-28 12:34:31,430 INFO cfg.RCNN.CLS_BG_THRESH_LO: 0.05 2019-05-28 12:34:31,430 INFO cfg.RCNN.REG_FG_THRESH: 0.55 2019-05-28 12:34:31,430 INFO cfg.RCNN.FG_RATIO: 0.5 2019-05-28 12:34:31,430 INFO cfg.RCNN.ROI_PER_IMAGE: 64 2019-05-28 12:34:31,430 INFO cfg.RCNN.HARD_BG_RATIO: 0.8 2019-05-28 12:34:31,430 INFO cfg.RCNN.SCORE_THRESH: 0.3 2019-05-28 12:34:31,430 INFO cfg.RCNN.NMS_THRESH: 0.1 2019-05-28 12:34:31,430 INFO cfg.TRAIN = edict() 2019-05-28 12:34:31,430 INFO cfg.TRAIN.SPLIT: train 2019-05-28 12:34:31,430 INFO cfg.TRAIN.VAL_SPLIT: smallval 2019-05-28 12:34:31,430 INFO cfg.TRAIN.LR: 0.002 2019-05-28 12:34:31,430 INFO cfg.TRAIN.LR_CLIP: 1e-05 2019-05-28 12:34:31,430 INFO cfg.TRAIN.LR_DECAY: 0.5 2019-05-28 12:34:31,430 INFO cfg.TRAIN.DECAY_STEP_LIST: [100, 150, 180, 200] 2019-05-28 12:34:31,430 INFO cfg.TRAIN.LR_WARMUP: True 2019-05-28 12:34:31,430 INFO cfg.TRAIN.WARMUP_MIN: 0.0002 2019-05-28 12:34:31,430 INFO cfg.TRAIN.WARMUP_EPOCH: 1 2019-05-28 12:34:31,430 INFO cfg.TRAIN.BN_MOMENTUM: 0.1 2019-05-28 12:34:31,430 INFO cfg.TRAIN.BN_DECAY: 0.5 2019-05-28 12:34:31,430 INFO cfg.TRAIN.BNM_CLIP: 0.01 2019-05-28 12:34:31,430 INFO cfg.TRAIN.BN_DECAY_STEP_LIST: [1000] 2019-05-28 12:34:31,430 INFO cfg.TRAIN.OPTIMIZER: adam_onecycle 2019-05-28 12:34:31,430 INFO cfg.TRAIN.WEIGHT_DECAY: 0.001 2019-05-28 12:34:31,431 INFO cfg.TRAIN.MOMENTUM: 0.9 2019-05-28 12:34:31,431 INFO cfg.TRAIN.MOMS: [0.95, 0.85] 2019-05-28 12:34:31,431 INFO cfg.TRAIN.DIV_FACTOR: 10.0 2019-05-28 12:34:31,431 INFO cfg.TRAIN.PCT_START: 0.4 2019-05-28 12:34:31,431 INFO cfg.TRAIN.GRAD_NORM_CLIP: 1.0 2019-05-28 12:34:31,431 INFO cfg.TRAIN.RPN_PRE_NMS_TOP_N: 9000 2019-05-28 12:34:31,431 INFO cfg.TRAIN.RPN_POST_NMS_TOP_N: 512 2019-05-28 12:34:31,431 INFO cfg.TRAIN.RPN_NMS_THRESH: 0.85 2019-05-28 12:34:31,431 INFO cfg.TRAIN.RPN_DISTANCE_BASED_PROPOSE: True 2019-05-28 12:34:31,431 INFO cfg.TEST = edict() 2019-05-28 12:34:31,431 INFO cfg.TEST.SPLIT: val 2019-05-28 12:34:31,431 INFO cfg.TEST.RPN_PRE_NMS_TOP_N: 9000 2019-05-28 12:34:31,431 INFO cfg.TEST.RPN_POST_NMS_TOP_N: 100 2019-05-28 12:34:31,431 INFO cfg.TEST.RPN_NMS_THRESH: 0.8 2019-05-28 12:34:31,431 INFO cfg.TEST.RPN_DISTANCE_BASED_PROPOSE: True 2019-05-28 12:34:31,432 INFO Load testing samples from ../data/KITTI/object/training 2019-05-28 12:34:31,432 INFO Done: total test samples 3769 2019-05-28 12:34:33,657 INFO ==> Loading from checkpoint 'pretrained_model/PointRCNN.pth' 2019-05-28 12:34:33,705 INFO ==> Done 2019-05-28 12:34:33,707 INFO ---- EPOCH no_number JOINT EVALUATION ---- 2019-05-28 12:34:33,707 INFO ==> Output file: ../output/rcnn/default/eval/epoch_no_number/val eval: 0%| | 0/3769 [00:00<?, ?it/s]Traceback (most recent call last): File "eval_rcnn.py", line 902, in eval_single_ckpt(root_result_dir) File "eval_rcnn.py", line 765, in eval_single_ckpt eval_one_epoch(model, test_loader, epoch_id, root_result_dir, logger) File "eval_rcnn.py", line 692, in eval_one_epoch ret_dict = eval_one_epoch_joint(model, dataloader, epoch_id, result_dir, logger) File "eval_rcnn.py", line 486, in eval_one_epoch_joint for data in dataloader: File "/home/ekele/anaconda3/envs/gpurl/lib/python3.6/site-packages/torch/utils/data/dataloader.py", line 637, in next return self._process_next_batch(batch) File "/home/ekele/anaconda3/envs/gpurl/lib/python3.6/site-packages/torch/utils/data/dataloader.py", line 658, in _process_next_batch raise batch.exc_type(batch.exc_msg) AssertionError: Traceback (most recent call last): File "/home/ekele/anaconda3/envs/gpurl/lib/python3.6/site-packages/torch/utils/data/dataloader.py", line 138, in _worker_loop samples = collate_fn([dataset[i] for i in batch_indices]) File "/home/ekele/anaconda3/envs/gpurl/lib/python3.6/site-packages/torch/utils/data/dataloader.py", line 138, in samples = collate_fn([dataset[i] for i in batch_indices]) File "/home/ekele/PointRCNN/tools/../lib/datasets/kitti_rcnn_dataset.py", line 234, in getitem return self.get_rpn_sample(index) File "/home/ekele/PointRCNN/tools/../lib/datasets/kitti_rcnn_dataset.py", line 249, in get_rpn_sample calib = self.get_calib(sample_id) File "/home/ekele/PointRCNN/tools/../lib/datasets/kitti_rcnn_dataset.py", line 133, in get_calib return super().get_calib(idx % 10000) File "/home/ekele/PointRCNN/tools/../lib/datasets/kitti_dataset.py", line 47, in get_calib assert os.path.exists(calib_file) AssertionError
hello, I got the same problems with you when I try to run the eval_rcnn.py. Did you fixed this bug?
Could you go into kitti_dataset.py and add a print(calib_file)
infront of the
assert os.path.exists(calib_file)
Then check if that file exists. Maybe you put the calib folder somewhere different.
Could you go into kitti_dataset.py and add a
print(calib_file)
infront of theassert os.path.exists(calib_file)
Then check if that file exists. Maybe you put the calib folder somewhere different.
Thank you for your help, I just download the [Download Velodyne point clouds, if you want to use laser information (29 GB)] in KITTI website。 Maybe I need to download another data?
Could you go into kitti_dataset.py and add a
print(calib_file)
infront of theassert os.path.exists(calib_file)
Then check if that file exists. Maybe you put the calib folder somewhere different.
Thank you for your help, I am fixed this bug(I forgot to download some data).
hello
my computer evn is cuda10.0 and GPU Tesla T4, python3.7, pytorch1.2, when i run the
Inference
, python eval_rcnn.py --cfg_file cfgs/default.yaml --ckpt PointRCNN.pth --batch_size 4 --eval_mode rcnn
, I met a problem, you know why this problem is caused?
yaml_cfg = edict(yaml.load(f)) 2019-11-26 13:07:32,556 INFO **Start logging** 2019-11-26 13:07:32,556 INFO cfg_file cfgs/default.yaml 2019-11-26 13:07:32,556 INFO eval_mode rcnn 2019-11-26 13:07:32,556 INFO eval_all False 2019-11-26 13:07:32,556 INFO test False 2019-11-26 13:07:32,556 INFO ckpt PointRCNN.pth 2019-11-26 13:07:32,556 INFO rpn_ckpt None 2019-11-26 13:07:32,556 INFO rcnn_ckpt None 2019-11-26 13:07:32,556 INFO batch_size 4 2019-11-26 13:07:32,556 INFO workers 4 2019-11-26 13:07:32,557 INFO extra_tag default 2019-11-26 13:07:32,557 INFO output_dir None 2019-11-26 13:07:32,557 INFO ckpt_dir None 2019-11-26 13:07:32,557 INFO save_result False 2019-11-26 13:07:32,557 INFO save_rpn_feature False 2019-11-26 13:07:32,557 INFO random_select True 2019-11-26 13:07:32,557 INFO start_epoch 0 2019-11-26 13:07:32,557 INFO rcnn_eval_roi_dir None 2019-11-26 13:07:32,557 INFO rcnn_eval_feature_dir None 2019-11-26 13:07:32,557 INFO set_cfgs None 2019-11-26 13:07:32,557 INFO cfg.TAG: default 2019-11-26 13:07:32,557 INFO cfg.CLASSES: Car 2019-11-26 13:07:32,557 INFO cfg.INCLUDE_SIMILAR_TYPE: True 2019-11-26 13:07:32,557 INFO cfg.AUG_DATA: True 2019-11-26 13:07:32,557 INFO cfg.AUG_METHOD_LIST: ['rotation', 'scaling', 'flip'] 2019-11-26 13:07:32,557 INFO cfg.AUG_METHOD_PROB: [1.0, 1.0, 0.5] 2019-11-26 13:07:32,557 INFO cfg.AUG_ROT_RANGE: 18 2019-11-26 13:07:32,557 INFO cfg.GT_AUG_ENABLED: True 2019-11-26 13:07:32,557 INFO cfg.GT_EXTRA_NUM: 15 2019-11-26 13:07:32,557 INFO cfg.GT_AUG_RAND_NUM: True 2019-11-26 13:07:32,557 INFO cfg.GT_AUG_APPLY_PROB: 1.0 2019-11-26 13:07:32,557 INFO cfg.GT_AUG_HARD_RATIO: 0.6 2019-11-26 13:07:32,557 INFO cfg.PC_REDUCE_BY_RANGE: True 2019-11-26 13:07:32,558 INFO cfg.PC_AREA_SCOPE: [[-40. 40. ] [ -1. 3. ] [ 0. 70.4]] 2019-11-26 13:07:32,558 INFO cfg.CLS_MEAN_SIZE: [[1.5256319 1.6285675 3.8831165]] 2019-11-26 13:07:32,558 INFO cfg.RPN = edict() 2019-11-26 13:07:32,558 INFO cfg.RPN.ENABLED: True 2019-11-26 13:07:32,558 INFO cfg.RPN.FIXED: True 2019-11-26 13:07:32,558 INFO cfg.RPN.USE_INTENSITY: False 2019-11-26 13:07:32,558 INFO cfg.RPN.LOC_XZ_FINE: True 2019-11-26 13:07:32,558 INFO cfg.RPN.LOC_SCOPE: 3.0 2019-11-26 13:07:32,558 INFO cfg.RPN.LOC_BIN_SIZE: 0.5 2019-11-26 13:07:32,558 INFO cfg.RPN.NUM_HEAD_BIN: 12 2019-11-26 13:07:32,558 INFO cfg.RPN.BACKBONE: pointnet2_msg 2019-11-26 13:07:32,558 INFO cfg.RPN.USE_BN: True 2019-11-26 13:07:32,558 INFO cfg.RPN.NUM_POINTS: 16384 2019-11-26 13:07:32,558 INFO cfg.RPN.SA_CONFIG = edict() 2019-11-26 13:07:32,559 INFO cfg.RPN.SA_CONFIG.NPOINTS: [4096, 1024, 256, 64] 2019-11-26 13:07:32,559 INFO cfg.RPN.SA_CONFIG.RADIUS: [[0.1, 0.5], [0.5, 1.0], [1.0, 2.0], [2.0, 4.0]] 2019-11-26 13:07:32,559 INFO cfg.RPN.SA_CONFIG.NSAMPLE: [[16, 32], [16, 32], [16, 32], [16, 32]] 2019-11-26 13:07:32,559 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]]] 2019-11-26 13:07:32,559 INFO cfg.RPN.FP_MLPS: [[128, 128], [256, 256], [512, 512], [512, 512]] 2019-11-26 13:07:32,559 INFO cfg.RPN.CLS_FC: [128] 2019-11-26 13:07:32,559 INFO cfg.RPN.REG_FC: [128] 2019-11-26 13:07:32,559 INFO cfg.RPN.DP_RATIO: 0.5 2019-11-26 13:07:32,559 INFO cfg.RPN.LOSS_CLS: SigmoidFocalLoss 2019-11-26 13:07:32,559 INFO cfg.RPN.FG_WEIGHT: 15 2019-11-26 13:07:32,559 INFO cfg.RPN.FOCAL_ALPHA: [0.25, 0.75] 2019-11-26 13:07:32,559 INFO cfg.RPN.FOCAL_GAMMA: 2.0 2019-11-26 13:07:32,559 INFO cfg.RPN.REG_LOSS_WEIGHT: [1.0, 1.0, 1.0, 1.0] 2019-11-26 13:07:32,559 INFO cfg.RPN.LOSS_WEIGHT: [1.0, 1.0] 2019-11-26 13:07:32,559 INFO cfg.RPN.NMS_TYPE: normal 2019-11-26 13:07:32,559 INFO cfg.RPN.SCORE_THRESH: 0.3 2019-11-26 13:07:32,559 INFO cfg.RCNN = edict() 2019-11-26 13:07:32,559 INFO cfg.RCNN.ENABLED: True 2019-11-26 13:07:32,559 INFO cfg.RCNN.USE_RPN_FEATURES: True 2019-11-26 13:07:32,559 INFO cfg.RCNN.USE_MASK: True 2019-11-26 13:07:32,559 INFO cfg.RCNN.MASK_TYPE: seg 2019-11-26 13:07:32,559 INFO cfg.RCNN.USE_INTENSITY: False 2019-11-26 13:07:32,559 INFO cfg.RCNN.USE_DEPTH: True 2019-11-26 13:07:32,559 INFO cfg.RCNN.USE_SEG_SCORE: False 2019-11-26 13:07:32,559 INFO cfg.RCNN.ROI_SAMPLE_JIT: True 2019-11-26 13:07:32,559 INFO cfg.RCNN.ROI_FG_AUG_TIMES: 10 2019-11-26 13:07:32,560 INFO cfg.RCNN.REG_AUG_METHOD: multiple 2019-11-26 13:07:32,560 INFO cfg.RCNN.POOL_EXTRA_WIDTH: 1.0 2019-11-26 13:07:32,560 INFO cfg.RCNN.LOC_SCOPE: 1.5 2019-11-26 13:07:32,560 INFO cfg.RCNN.LOC_BIN_SIZE: 0.5 2019-11-26 13:07:32,560 INFO cfg.RCNN.NUM_HEAD_BIN: 9 2019-11-26 13:07:32,560 INFO cfg.RCNN.LOC_Y_BY_BIN: False 2019-11-26 13:07:32,560 INFO cfg.RCNN.LOC_Y_SCOPE: 0.5 2019-11-26 13:07:32,560 INFO cfg.RCNN.LOC_Y_BIN_SIZE: 0.25 2019-11-26 13:07:32,560 INFO cfg.RCNN.SIZE_RES_ON_ROI: False 2019-11-26 13:07:32,560 INFO cfg.RCNN.USE_BN: False 2019-11-26 13:07:32,560 INFO cfg.RCNN.DP_RATIO: 0.0 2019-11-26 13:07:32,560 INFO cfg.RCNN.BACKBONE: pointnet 2019-11-26 13:07:32,560 INFO cfg.RCNN.XYZ_UP_LAYER: [128, 128] 2019-11-26 13:07:32,560 INFO cfg.RCNN.NUM_POINTS: 512 2019-11-26 13:07:32,560 INFO cfg.RCNN.SA_CONFIG = edict() 2019-11-26 13:07:32,560 INFO cfg.RCNN.SA_CONFIG.NPOINTS: [128, 32, -1] 2019-11-26 13:07:32,560 INFO cfg.RCNN.SA_CONFIG.RADIUS: [0.2, 0.4, 100] 2019-11-26 13:07:32,560 INFO cfg.RCNN.SA_CONFIG.NSAMPLE: [64, 64, 64] 2019-11-26 13:07:32,560 INFO cfg.RCNN.SA_CONFIG.MLPS: [[128, 128, 128], [128, 128, 256], [256, 256, 512]] 2019-11-26 13:07:32,560 INFO cfg.RCNN.CLS_FC: [256, 256] 2019-11-26 13:07:32,560 INFO cfg.RCNN.REG_FC: [256, 256] 2019-11-26 13:07:32,560 INFO cfg.RCNN.LOSS_CLS: BinaryCrossEntropy 2019-11-26 13:07:32,560 INFO cfg.RCNN.FOCAL_ALPHA: [0.25, 0.75] 2019-11-26 13:07:32,560 INFO cfg.RCNN.FOCAL_GAMMA: 2.0 2019-11-26 13:07:32,561 INFO cfg.RCNN.CLS_WEIGHT: [1. 1. 1.] 2019-11-26 13:07:32,561 INFO cfg.RCNN.CLS_FG_THRESH: 0.6 2019-11-26 13:07:32,561 INFO cfg.RCNN.CLS_BG_THRESH: 0.45 2019-11-26 13:07:32,561 INFO cfg.RCNN.CLS_BG_THRESH_LO: 0.05 2019-11-26 13:07:32,561 INFO cfg.RCNN.REG_FG_THRESH: 0.55 2019-11-26 13:07:32,561 INFO cfg.RCNN.FG_RATIO: 0.5 2019-11-26 13:07:32,561 INFO cfg.RCNN.ROI_PER_IMAGE: 64 2019-11-26 13:07:32,561 INFO cfg.RCNN.HARD_BG_RATIO: 0.8 2019-11-26 13:07:32,561 INFO cfg.RCNN.SCORE_THRESH: 0.3 2019-11-26 13:07:32,561 INFO cfg.RCNN.NMS_THRESH: 0.1 2019-11-26 13:07:32,561 INFO cfg.TRAIN = edict() 2019-11-26 13:07:32,561 INFO cfg.TRAIN.SPLIT: train 2019-11-26 13:07:32,561 INFO cfg.TRAIN.VAL_SPLIT: smallval 2019-11-26 13:07:32,561 INFO cfg.TRAIN.LR: 0.002 2019-11-26 13:07:32,561 INFO cfg.TRAIN.LR_CLIP: 1e-05 2019-11-26 13:07:32,561 INFO cfg.TRAIN.LR_DECAY: 0.5 2019-11-26 13:07:32,561 INFO cfg.TRAIN.DECAY_STEP_LIST: [100, 150, 180, 200] 2019-11-26 13:07:32,561 INFO cfg.TRAIN.LR_WARMUP: True 2019-11-26 13:07:32,561 INFO cfg.TRAIN.WARMUP_MIN: 0.0002 2019-11-26 13:07:32,561 INFO cfg.TRAIN.WARMUP_EPOCH: 1 2019-11-26 13:07:32,561 INFO cfg.TRAIN.BN_MOMENTUM: 0.1 2019-11-26 13:07:32,561 INFO cfg.TRAIN.BN_DECAY: 0.5 2019-11-26 13:07:32,561 INFO cfg.TRAIN.BNM_CLIP: 0.01 2019-11-26 13:07:32,562 INFO cfg.TRAIN.BN_DECAY_STEP_LIST: [1000] 2019-11-26 13:07:32,562 INFO cfg.TRAIN.OPTIMIZER: adam_onecycle 2019-11-26 13:07:32,562 INFO cfg.TRAIN.WEIGHT_DECAY: 0.001 2019-11-26 13:07:32,562 INFO cfg.TRAIN.MOMENTUM: 0.9 2019-11-26 13:07:32,562 INFO cfg.TRAIN.MOMS: [0.95, 0.85] 2019-11-26 13:07:32,562 INFO cfg.TRAIN.DIV_FACTOR: 10.0 2019-11-26 13:07:32,562 INFO cfg.TRAIN.PCT_START: 0.4 2019-11-26 13:07:32,562 INFO cfg.TRAIN.GRAD_NORM_CLIP: 1.0 2019-11-26 13:07:32,562 INFO cfg.TRAIN.RPN_PRE_NMS_TOP_N: 9000 2019-11-26 13:07:32,562 INFO cfg.TRAIN.RPN_POST_NMS_TOP_N: 512 2019-11-26 13:07:32,562 INFO cfg.TRAIN.RPN_NMS_THRESH: 0.85 2019-11-26 13:07:32,562 INFO cfg.TRAIN.RPN_DISTANCE_BASED_PROPOSE: True 2019-11-26 13:07:32,562 INFO cfg.TEST = edict() 2019-11-26 13:07:32,562 INFO cfg.TEST.SPLIT: val 2019-11-26 13:07:32,562 INFO cfg.TEST.RPN_PRE_NMS_TOP_N: 9000 2019-11-26 13:07:32,562 INFO cfg.TEST.RPN_POST_NMS_TOP_N: 100 2019-11-26 13:07:32,562 INFO cfg.TEST.RPN_NMS_THRESH: 0.8 2019-11-26 13:07:32,562 INFO cfg.TEST.RPN_DISTANCE_BASED_PROPOSE: True 2019-11-26 13:07:32,564 INFO Load testing samples from ../data/KITTI/object/training 2019-11-26 13:07:32,564 INFO Done: total test samples 3769 2019-11-26 13:07:34,925 INFO ==> Loading from checkpoint 'PointRCNN.pth' Traceback (most recent call last): File "eval_rcnn.py", line 902, in
eval_single_ckpt(root_result_dir) File "eval_rcnn.py", line 762, in eval_single_ckpt load_ckpt_based_on_args(model, logger) File "eval_rcnn.py", line 719, in load_ckpt_based_on_args train_utils.load_checkpoint(model, filename=args.ckpt, logger=logger) File "/data/cxy/code/PointRCNN/tools/../tools/train_utils/train_utils.py", line 85, in load_checkpoint model.load_state_dict(checkpoint['model_state']) File "/home/WJ2/.conda/envs/cxy/lib/python3.6/site-packages/torch/nn/modules/module.py", line 845, in load_state_dict self.class.name, "\n\t".join(error_msgs))) RuntimeError: Error(s) in loading state_dict for PointRCNN: size mismatch for rpn.rpn_reg_layer.2.conv.weight: copying a param with shape torch.Size([52, 128, 1]) from checkpoint, the shape in current model is torch.Size([76, 128, 1]). size mismatch for rpn.rpn_reg_layer.2.conv.bias: copying a param with shape torch.Size([52]) from checkpoint, the shape in current model is torch.Size([76]).
@chenxyyy It seems like you either modified the config file between training and interference or used a different file whatsoever. It tries to load the parameters from the trained model to an empty model that has been set up according to the config file. If you change the config file after training, some sizes of that empty model may change to the trained one.
Thank you very much. I want to directly use the model provided by the author to evaluate。 Can I do that? @nicolas-schreiber
Well, I guess, just use the original config file from this repo without any changes
I didn't change any config file, but the above error occurred after the code ran.
Could you go into kitti_dataset.py and add a
print(calib_file)
infront of theassert os.path.exists(calib_file)
Then check if that file exists. Maybe you put the calib folder somewhere different.
Thank you for your help, I am fixed this bug(I forgot to download some data).
how did you fixed this bug ?
你可以进入kitti_dataset.py和添加
print(calib_file)
的盈assert os.path.exists(calib_file)
然后检查该文件是否存在。也许您将calib文件夹放在其他位置。
感谢您的帮助,我已修复此错误(我忘记下载一些数据)。
您好,请问您怎么解决的这个问题?能指导一下我吗?
Could you go into kitti_dataset.py and add a
print(calib_file)
infront of theassert os.path.exists(calib_file)
Then check if that file exists. Maybe you put the calib folder somewhere different.
Thank you for your help, I am fixed this bug(I forgot to download some data).
how did you fixed this bug ?
the path is /object, not /Object
I didn't change any config file, but the above error occurred after the code ran.
I have the same problem! any advice?
I have the same problem! any advice?
hello my computer evn is cuda9.0 and GPU2080, do you know why this problem is caused? RuntimeError: cublas runtime error : the GPU program failed to execute at /opt/conda/conda-bld/pytorch_1549628766161/work/aten/src/THC/THCBlas.cu:441