The accuracy is 0

[jiameixia1202]

My server does not have 8 GPUs. When I use 4 GPUs for training (without any modification to the network), I just change num-gpus to 4. After training, the segmentation accuracy is always 0.00. Do you need to modify other parameters?

[haderalim]

Did you update of the registered coco dataset? And what parameters and its values, you used in config file?

[jiameixia1202]

The following are the parameters we used in training without any modification. I don't quite understand what you mean by "the registered coco dataset"

/root/anaconda3/envs/detectron2/bin/python3.6 /jmx/centermask2/train_net.py --config-file configs/centermask/centermask_lite_V_19_eSE_FPN_ms_4x.yaml --num-gpus 4
Command Line Args: Namespace(config_file='configs/centermask/centermask_lite_V_19_eSE_FPN_ms_4x.yaml', dist_url='tcp://127.0.0.1:49152', eval_only=False, machine_rank=0, num_gpus=4, num_machines=1, opts=[], resume=False)
Config 'configs/centermask/centermask_lite_V_19_eSE_FPN_ms_4x.yaml' has no VERSION. Assuming it to be compatible with latest v2.
Config 'configs/centermask/centermask_lite_V_19_eSE_FPN_ms_4x.yaml' has no VERSION. Assuming it to be compatible with latest v2.
Config 'configs/centermask/centermask_lite_V_19_eSE_FPN_ms_4x.yaml' has no VERSION. Assuming it to be compatible with latest v2.
Config 'configs/centermask/centermask_lite_V_19_eSE_FPN_ms_4x.yaml' has no VERSION. Assuming it to be compatible with latest v2.
[12/08 18:45:25 detectron2]: Rank of current process: 0. World size: 4
[12/08 18:45:27 detectron2]: Environment info:

---

sys.platform linux
Python 3.6.12 |Anaconda, Inc.| (default, Sep 8 2020, 23:10:56) [GCC 7.3.0]
numpy 1.19.4
detectron2 0.3 @/jmx/detectron2/detectron2
Compiler GCC 7.4
CUDA compiler CUDA 10.1
detectron2 arch flags 6.1
DETECTRON2_ENV_MODULE
PyTorch 1.6.0+cu101 @/root/anaconda3/envs/detectron2/lib/python3.6/site-packages/torch
PyTorch debug build False
GPU available True
GPU 0,1,2,3 GeForce GTX 1080 (arch=6.1)
CUDA_HOME /usr/local/cuda
Pillow 8.0.1
torchvision 0.7.0+cu101 @/root/anaconda3/envs/detectron2/lib/python3.6/site-packages/torchvision
torchvision arch flags 3.5, 5.0, 6.0, 7.0, 7.5
fvcore 0.1.2.post20201122
cv2 4.4.0

---

PyTorch built with:

• GCC 7.3
• C++ Version: 201402
• Intel(R) Math Kernel Library Version 2019.0.5 Product Build 20190808 for Intel(R) 64 architecture applications
• Intel(R) MKL-DNN v1.5.0 (Git Hash e2ac1fac44c5078ca927cb9b90e1b3066a0b2ed0)
• OpenMP 201511 (a.k.a. OpenMP 4.5)
• NNPACK is enabled
• CPU capability usage: AVX2
• CUDA Runtime 10.1
• NVCC architecture flags: -gencode;arch=compute_37,code=sm_37;-gencode;arch=compute_50,code=sm_50;-gencode;arch=compute_60,code=sm_60;-gencode;arch=compute_70,code=sm_70;-gencode;arch=compute_75,code=sm_75
• CuDNN 7.6.3
• Magma 2.5.2
• Build settings: BLAS=MKL, BUILD_TYPE=Release, CXX_FLAGS= -Wno-deprecated -fvisibility-inlines-hidden -DUSE_PTHREADPOOL -fopenmp -DNDEBUG -DUSE_FBGEMM -DUSE_QNNPACK -DUSE_PYTORCH_QNNPACK -DUSE_XNNPACK -DUSE_VULKAN_WRAPPER -O2 -fPIC -Wno-narrowing -Wall -Wextra -Werror=return-type -Wno-missing-field-initializers -Wno-type-limits -Wno-array-bounds -Wno-unknown-pragmas -Wno-sign-compare -Wno-unused-parameter -Wno-unused-variable -Wno-unused-function -Wno-unused-result -Wno-unused-local-typedefs -Wno-strict-overflow -Wno-strict-aliasing -Wno-error=deprecated-declarations -Wno-stringop-overflow -Wno-error=pedantic -Wno-error=redundant-decls -Wno-error=old-style-cast -fdiagnostics-color=always -faligned-new -Wno-unused-but-set-variable -Wno-maybe-uninitialized -fno-math-errno -fno-trapping-math -Werror=format -Wno-stringop-overflow, PERF_WITH_AVX=1, PERF_WITH_AVX2=1, PERF_WITH_AVX512=1, USE_CUDA=ON, USE_EXCEPTION_PTR=1, USE_GFLAGS=OFF, USE_GLOG=OFF, USE_MKL=ON, USE_MKLDNN=ON, USE_MPI=OFF, USE_NCCL=ON, USE_NNPACK=ON, USE_OPENMP=ON, USE_STATIC_DISPATCH=OFF,

[12/08 18:45:27 detectron2]: Command line arguments: Namespace(config_file='configs/centermask/centermask_lite_V_19_eSE_FPN_ms_4x.yaml', dist_url='tcp://127.0.0.1:49152', eval_only=False, machine_rank=0, num_gpus=4, num_machines=1, opts=[], resume=False)
[12/08 18:45:27 detectron2]: Contents of args.config_file=configs/centermask/centermask_lite_V_19_eSE_FPN_ms_4x.yaml:
BASE: "Base-CenterMask-Lite-VoVNet.yaml"

MODEL:
WEIGHTS: "/jmx/centermask2/models/vovnet19_ese_detectron2.pth"
VOVNET:
CONV_BODY : "V-19-eSE"
SOLVER:
STEPS: (300000, 340000)
MAX_ITER: 360000
OUTPUT_DIR: "output/centermask/CenterMask-Lite-V-19-ms-4x"

[12/08 18:45:27 detectron2]: Running with full config:
CUDNN_BENCHMARK: False
DATALOADER:
ASPECT_RATIO_GROUPING: True
FILTER_EMPTY_ANNOTATIONS: True
NUM_WORKERS: 0
REPEAT_THRESHOLD: 0.0
SAMPLER_TRAIN: TrainingSampler
DATASETS:
PRECOMPUTED_PROPOSAL_TOPK_TEST: 1000
PRECOMPUTED_PROPOSAL_TOPK_TRAIN: 2000
PROPOSAL_FILES_TEST: ()
PROPOSAL_FILES_TRAIN: ()
TEST: ('coco_2017_val',)
TRAIN: ('coco_2017_train',)
GLOBAL:
HACK: 1.0
INPUT:
CROP:
ENABLED: False
SIZE: [0.9, 0.9]
TYPE: relative_range
FORMAT: BGR
MASK_FORMAT: polygon
MAX_SIZE_TEST: 1000
MAX_SIZE_TRAIN: 1000
MIN_SIZE_TEST: 600
MIN_SIZE_TRAIN: (580, 600)
MIN_SIZE_TRAIN_SAMPLING: choice
RANDOM_FLIP: horizontal
MODEL:
ANCHOR_GENERATOR:
ANGLES: [[-90, 0, 90]]
ASPECT_RATIOS: [[0.5, 1.0, 2.0]]
NAME: DefaultAnchorGenerator
OFFSET: 0.0
SIZES: [[32, 64, 128, 256, 512]]
BACKBONE:
FREEZE_AT: 0
NAME: build_fcos_vovnet_fpn_backbone
DEVICE: cuda
FCOS:
CENTER_SAMPLE: True
FPN_STRIDES: [8, 16, 32, 64, 128]
INFERENCE_TH_TEST: 0.05
INFERENCE_TH_TRAIN: 0.05
IN_FEATURES: ['p3', 'p4', 'p5', 'p6', 'p7']
LOC_LOSS_TYPE: giou
LOSS_ALPHA: 0.25
LOSS_GAMMA: 2.0
NMS_TH: 0.6
NORM: GN
NUM_BOX_CONVS: 2
NUM_CLASSES: 80
NUM_CLS_CONVS: 2
NUM_SHARE_CONVS: 0
POST_NMS_TOPK_TEST: 50
POST_NMS_TOPK_TRAIN: 100
POS_RADIUS: 1.5
PRE_NMS_TOPK_TEST: 1000
PRE_NMS_TOPK_TRAIN: 1000
PRIOR_PROB: 0.01
SIZES_OF_INTEREST: [64, 128, 256, 512]
THRESH_WITH_CTR: False
TOP_LEVELS: 2
USE_DEFORMABLE: False
USE_RELU: True
USE_SCALE: True
FPN:
FUSE_TYPE: sum
IN_FEATURES: ['stage3', 'stage4', 'stage5']
NORM:
OUT_CHANNELS: 128
KEYPOINT_ON: False
LOAD_PROPOSALS: False
MASKIOU_LOSS_WEIGHT: 1.0
MASKIOU_ON: True
MASK_ON: True
META_ARCHITECTURE: GeneralizedRCNN
MOBILENET: False
PANOPTIC_FPN:
COMBINE:
ENABLED: True
INSTANCES_CONFIDENCE_THRESH: 0.5
OVERLAP_THRESH: 0.5
STUFF_AREA_LIMIT: 4096
INSTANCE_LOSS_WEIGHT: 1.0
PIXEL_MEAN: [103.53, 116.28, 123.675]
PIXEL_STD: [1.0, 1.0, 1.0]
PROPOSAL_GENERATOR:
MIN_SIZE: 0
NAME: FCOS
RESNETS:
DEFORM_MODULATED: False
DEFORM_NUM_GROUPS: 1
DEFORM_ON_PER_STAGE: [False, False, False, False]
DEPTH: 50
NORM: FrozenBN
NUM_GROUPS: 1
OUT_FEATURES: ['res4']
RES2_OUT_CHANNELS: 256
RES5_DILATION: 1
STEM_OUT_CHANNELS: 64
STRIDE_IN_1X1: True
WIDTH_PER_GROUP: 64
RETINANET:
BBOX_REG_LOSS_TYPE: smooth_l1
BBOX_REG_WEIGHTS: (1.0, 1.0, 1.0, 1.0)
FOCAL_LOSS_ALPHA: 0.25
FOCAL_LOSS_GAMMA: 2.0
IN_FEATURES: ['p3', 'p4', 'p5', 'p6', 'p7']
IOU_LABELS: [0, -1, 1]
IOU_THRESHOLDS: [0.4, 0.5]
NMS_THRESH_TEST: 0.5
NORM:
NUM_CLASSES: 80
NUM_CONVS: 4
PRIOR_PROB: 0.01
SCORE_THRESH_TEST: 0.05
SMOOTH_L1_LOSS_BETA: 0.1
TOPK_CANDIDATES_TEST: 1000
ROI_BOX_CASCADE_HEAD:
BBOX_REG_WEIGHTS: ((10.0, 10.0, 5.0, 5.0), (20.0, 20.0, 10.0, 10.0), (30.0, 30.0, 15.0, 15.0))
IOUS: (0.5, 0.6, 0.7)
ROI_BOX_HEAD:
BBOX_REG_LOSS_TYPE: smooth_l1
BBOX_REG_LOSS_WEIGHT: 1.0
BBOX_REG_WEIGHTS: (10.0, 10.0, 5.0, 5.0)
CLS_AGNOSTIC_BBOX_REG: False
CONV_DIM: 256
FC_DIM: 1024
NAME:
NORM:
NUM_CONV: 0
NUM_FC: 0
POOLER_RESOLUTION: 14
POOLER_SAMPLING_RATIO: 0
POOLER_TYPE: ROIAlignV2
SMOOTH_L1_BETA: 0.0
TRAIN_ON_PRED_BOXES: False
ROI_HEADS:
BATCH_SIZE_PER_IMAGE: 512
IN_FEATURES: ['p3', 'p4', 'p5']
IOU_LABELS: [0, 1]
IOU_THRESHOLDS: [0.5]
NAME: CenterROIHeads
NMS_THRESH_TEST: 0.5
NUM_CLASSES: 80
POSITIVE_FRACTION: 0.25
PROPOSAL_APPEND_GT: True
SCORE_THRESH_TEST: 0.05
ROI_KEYPOINT_HEAD:
ASSIGN_CRITERION: ratio
CONV_DIMS: (512, 512, 512, 512, 512, 512, 512, 512)
IN_FEATURES: ['p2', 'p3', 'p4', 'p5']
LOSS_WEIGHT: 1.0
MIN_KEYPOINTS_PER_IMAGE: 1
NAME: KRCNNConvDeconvUpsampleHead
NORMALIZE_LOSS_BY_VISIBLE_KEYPOINTS: True
NUM_KEYPOINTS: 17
POOLER_RESOLUTION: 14
POOLER_SAMPLING_RATIO: 0
POOLER_TYPE: ROIAlignV2
ROI_MASKIOU_HEAD:
CONV_DIM: 128
NAME: MaskIoUHead
NUM_CONV: 2
ROI_MASK_HEAD:
ASSIGN_CRITERION: ratio
CLS_AGNOSTIC_MASK: False
CONV_DIM: 128
NAME: SpatialAttentionMaskHead
NORM:
NUM_CONV: 2
POOLER_RESOLUTION: 14
POOLER_SAMPLING_RATIO: 0
POOLER_TYPE: ROIAlignV2
RPN:
BATCH_SIZE_PER_IMAGE: 256
BBOX_REG_LOSS_TYPE: smooth_l1
BBOX_REG_LOSS_WEIGHT: 1.0
BBOX_REG_WEIGHTS: (1.0, 1.0, 1.0, 1.0)
BOUNDARY_THRESH: -1
HEAD_NAME: StandardRPNHead
IN_FEATURES: ['res4']
IOU_LABELS: [0, -1, 1]
IOU_THRESHOLDS: [0.3, 0.7]
LOSS_WEIGHT: 1.0
NMS_THRESH: 0.7
POSITIVE_FRACTION: 0.5
POST_NMS_TOPK_TEST: 1000
POST_NMS_TOPK_TRAIN: 2000
PRE_NMS_TOPK_TEST: 6000
PRE_NMS_TOPK_TRAIN: 12000
SMOOTH_L1_BETA: 0.0
SEM_SEG_HEAD:
COMMON_STRIDE: 4
CONVS_DIM: 128
IGNORE_VALUE: 255
IN_FEATURES: ['p2', 'p3', 'p4', 'p5']
LOSS_WEIGHT: 1.0
NAME: SemSegFPNHead
NORM: GN
NUM_CLASSES: 54
VOVNET:
BACKBONE_OUT_CHANNELS: 256
CONV_BODY: V-19-eSE
DEFORMABLE_GROUPS: 1
NORM: FrozenBN
OUT_CHANNELS: 256
OUT_FEATURES: ['stage3', 'stage4', 'stage5']
STAGE_WITH_DCN: (False, False, False, False)
WITH_MODULATED_DCN: False
WEIGHTS: /jmx/centermask2/models/vovnet19_ese_detectron2.pth
OUTPUT_DIR: output/centermask/CenterMask-Lite-V-19-ms-4x
SEED: -1
SOLVER:
AMP:
ENABLED: False
BASE_LR: 0.01
BIAS_LR_FACTOR: 1.0
CHECKPOINT_PERIOD: 10000
CLIP_GRADIENTS:
CLIP_TYPE: value
CLIP_VALUE: 1.0
ENABLED: False
NORM_TYPE: 2.0
GAMMA: 0.1
IMS_PER_BATCH: 16
LR_SCHEDULER_NAME: WarmupMultiStepLR
MAX_ITER: 360000
MOMENTUM: 0.9
NESTEROV: False
REFERENCE_WORLD_SIZE: 0
STEPS: (300000, 340000)
WARMUP_FACTOR: 0.001
WARMUP_ITERS: 1000
WARMUP_METHOD: linear
WEIGHT_DECAY: 0.0001
WEIGHT_DECAY_BIAS: 0.0001
WEIGHT_DECAY_NORM: 0.0
TEST:
AUG:
ENABLED: False
FLIP: True
MAX_SIZE: 4000
MIN_SIZES: (400, 500, 600, 700, 800, 900, 1000, 1100, 1200)
DETECTIONS_PER_IMAGE: 100
EVAL_PERIOD: 0
EXPECTED_RESULTS: []
KEYPOINT_OKS_SIGMAS: []
PRECISE_BN:
ENABLED: False
NUM_ITER: 200
VERSION: 2
VIS_PERIOD: 0
[12/08 18:45:27 detectron2]: Full config saved to output/centermask/CenterMask-Lite-V-19-ms-4x/config.yaml
[12/08 18:45:27 d2.utils.env]: Using a generated random seed 27801612
[12/08 18:45:28 d2.engine.defaults]: Model:
GeneralizedRCNN(
(backbone): FPN(
(fpn_lateral3): Conv2d(512, 128, kernel_size=(1, 1), stride=(1, 1))
(fpn_output3): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
(fpn_lateral4): Conv2d(768, 128, kernel_size=(1, 1), stride=(1, 1))
(fpn_output4): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
(fpn_lateral5): Conv2d(1024, 128, kernel_size=(1, 1), stride=(1, 1))
(fpn_output5): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
(top_block): LastLevelP6P7(
(p6): Conv2d(128, 128, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1))
(p7): Conv2d(128, 128, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1))
)
(bottom_up): VoVNet(
(stem): Sequential(
(stem_1/conv): Conv2d(3, 64, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)
(stem_1/norm): FrozenBatchNorm2d(num_features=64, eps=1e-05)
(stem_1/relu): ReLU(inplace=True)
(stem_2/conv): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(stem_2/norm): FrozenBatchNorm2d(num_features=64, eps=1e-05)
(stem_2/relu): ReLU(inplace=True)
(stem_3/conv): Conv2d(64, 128, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)
(stem_3/norm): FrozenBatchNorm2d(num_features=128, eps=1e-05)
(stem_3/relu): ReLU(inplace=True)
)
(stage2): _OSA_stage(
(OSA2_1): _OSA_module(
(layers): ModuleList(
(0): Sequential(
(OSA2_1_0/conv): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(OSA2_1_0/norm): FrozenBatchNorm2d(num_features=128, eps=1e-05)
(OSA2_1_0/relu): ReLU(inplace=True)
)
(1): Sequential(
(OSA2_1_1/conv): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(OSA2_1_1/norm): FrozenBatchNorm2d(num_features=128, eps=1e-05)
(OSA2_1_1/relu): ReLU(inplace=True)
)
(2): Sequential(
(OSA2_1_2/conv): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(OSA2_1_2/norm): FrozenBatchNorm2d(num_features=128, eps=1e-05)
(OSA2_1_2/relu): ReLU(inplace=True)
)
)
(concat): Sequential(
(OSA2_1_concat/conv): Conv2d(512, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)
(OSA2_1_concat/norm): FrozenBatchNorm2d(num_features=256, eps=1e-05)
(OSA2_1_concat/relu): ReLU(inplace=True)
)
(ese): eSEModule(
(avg_pool): AdaptiveAvgPool2d(output_size=1)
(fc): Conv2d(256, 256, kernel_size=(1, 1), stride=(1, 1))
(hsigmoid): Hsigmoid()
)
)
)
(stage3): _OSA_stage(
(Pooling): MaxPool2d(kernel_size=3, stride=2, padding=0, dilation=1, ceil_mode=True)
(OSA3_1): _OSA_module(
(layers): ModuleList(
(0): Sequential(
(OSA3_1_0/conv): Conv2d(256, 160, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(OSA3_1_0/norm): FrozenBatchNorm2d(num_features=160, eps=1e-05)
(OSA3_1_0/relu): ReLU(inplace=True)
)
(1): Sequential(
(OSA3_1_1/conv): Conv2d(160, 160, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(OSA3_1_1/norm): FrozenBatchNorm2d(num_features=160, eps=1e-05)
(OSA3_1_1/relu): ReLU(inplace=True)
)
(2): Sequential(
(OSA3_1_2/conv): Conv2d(160, 160, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(OSA3_1_2/norm): FrozenBatchNorm2d(num_features=160, eps=1e-05)
(OSA3_1_2/relu): ReLU(inplace=True)
)
)
(concat): Sequential(
(OSA3_1_concat/conv): Conv2d(736, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)
(OSA3_1_concat/norm): FrozenBatchNorm2d(num_features=512, eps=1e-05)
(OSA3_1_concat/relu): ReLU(inplace=True)
)
(ese): eSEModule(
(avg_pool): AdaptiveAvgPool2d(output_size=1)
(fc): Conv2d(512, 512, kernel_size=(1, 1), stride=(1, 1))
(hsigmoid): Hsigmoid()
)
)
)
(stage4): _OSA_stage(
(Pooling): MaxPool2d(kernel_size=3, stride=2, padding=0, dilation=1, ceil_mode=True)
(OSA4_1): _OSA_module(
(layers): ModuleList(
(0): Sequential(
(OSA4_1_0/conv): Conv2d(512, 192, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(OSA4_1_0/norm): FrozenBatchNorm2d(num_features=192, eps=1e-05)
(OSA4_1_0/relu): ReLU(inplace=True)
)
(1): Sequential(
(OSA4_1_1/conv): Conv2d(192, 192, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(OSA4_1_1/norm): FrozenBatchNorm2d(num_features=192, eps=1e-05)
(OSA4_1_1/relu): ReLU(inplace=True)
)
(2): Sequential(
(OSA4_1_2/conv): Conv2d(192, 192, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(OSA4_1_2/norm): FrozenBatchNorm2d(num_features=192, eps=1e-05)
(OSA4_1_2/relu): ReLU(inplace=True)
)
)
(concat): Sequential(
(OSA4_1_concat/conv): Conv2d(1088, 768, kernel_size=(1, 1), stride=(1, 1), bias=False)
(OSA4_1_concat/norm): FrozenBatchNorm2d(num_features=768, eps=1e-05)
(OSA4_1_concat/relu): ReLU(inplace=True)
)
(ese): eSEModule(
(avg_pool): AdaptiveAvgPool2d(output_size=1)
(fc): Conv2d(768, 768, kernel_size=(1, 1), stride=(1, 1))
(hsigmoid): Hsigmoid()
)
)
)
(stage5): _OSA_stage(
(Pooling): MaxPool2d(kernel_size=3, stride=2, padding=0, dilation=1, ceil_mode=True)
(OSA5_1): _OSA_module(
(layers): ModuleList(
(0): Sequential(
(OSA5_1_0/conv): Conv2d(768, 224, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(OSA5_1_0/norm): FrozenBatchNorm2d(num_features=224, eps=1e-05)
(OSA5_1_0/relu): ReLU(inplace=True)
)
(1): Sequential(
(OSA5_1_1/conv): Conv2d(224, 224, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(OSA5_1_1/norm): FrozenBatchNorm2d(num_features=224, eps=1e-05)
(OSA5_1_1/relu): ReLU(inplace=True)
)
(2): Sequential(
(OSA5_1_2/conv): Conv2d(224, 224, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(OSA5_1_2/norm): FrozenBatchNorm2d(num_features=224, eps=1e-05)
(OSA5_1_2/relu): ReLU(inplace=True)
)
)
(concat): Sequential(
(OSA5_1_concat/conv): Conv2d(1440, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)
(OSA5_1_concat/norm): FrozenBatchNorm2d(num_features=1024, eps=1e-05)
(OSA5_1_concat/relu): ReLU(inplace=True)
)
(ese): eSEModule(
(avg_pool): AdaptiveAvgPool2d(output_size=1)
(fc): Conv2d(1024, 1024, kernel_size=(1, 1), stride=(1, 1))
(hsigmoid): Hsigmoid()
)
)
)
)
)
(proposal_generator): FCOS(
(iou_loss): IOULoss()
(fcos_head): FCOSHead(
(cls_tower): Sequential(
(0): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
(1): GroupNorm(32, 128, eps=1e-05, affine=True)
(2): ReLU()
(3): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
(4): GroupNorm(32, 128, eps=1e-05, affine=True)
(5): ReLU()
)
(bbox_tower): Sequential(
(0): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
(1): GroupNorm(32, 128, eps=1e-05, affine=True)
(2): ReLU()
(3): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
(4): GroupNorm(32, 128, eps=1e-05, affine=True)
(5): ReLU()
)
(share_tower): Sequential()
(cls_logits): Conv2d(128, 80, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
(bbox_pred): Conv2d(128, 4, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
(ctrness): Conv2d(128, 1, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
(scales): ModuleList(
(0): Scale()
(1): Scale()
(2): Scale()
(3): Scale()
(4): Scale()
)
)
)
(roi_heads): CenterROIHeads(
(mask_pooler): ROIPooler(
(level_poolers): ModuleList(
(0): ROIAlign(output_size=(14, 14), spatial_scale=0.125, sampling_ratio=0, aligned=True)
(1): ROIAlign(output_size=(14, 14), spatial_scale=0.0625, sampling_ratio=0, aligned=True)
(2): ROIAlign(output_size=(14, 14), spatial_scale=0.03125, sampling_ratio=0, aligned=True)
)
)
(mask_head): SpatialAttentionMaskHead(
(mask_fcn1): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
(mask_fcn2): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
(spatialAtt): SpatialAttention(
(conv): Conv2d(2, 1, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(sigmoid): Sigmoid()
)
(deconv): ConvTranspose2d(128, 128, kernel_size=(2, 2), stride=(2, 2))
(predictor): Conv2d(128, 80, kernel_size=(1, 1), stride=(1, 1))
)
(maskiou_head): MaskIoUHead(
(maskiou_fcn1): Conv2d(129, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
(maskiou_fcn2): Conv2d(128, 128, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1))
(maskiou_fc1): Linear(in_features=6272, out_features=1024, bias=True)
(maskiou_fc2): Linear(in_features=1024, out_features=1024, bias=True)
(maskiou): Linear(in_features=1024, out_features=80, bias=True)
(pooling): MaxPool2d(kernel_size=[2, 2], stride=[2, 2], padding=[0, 0], dilation=[1, 1], ceil_mode=False)
)
)
)
[12/08 18:45:28 d2.data.dataset_mapper]: [DatasetMapper] Augmentations used in training: [ResizeShortestEdge(short_edge_length=(580, 600), max_size=1000, sample_style='choice'), RandomFlip()]
[12/08 18:46:12 d2.data.datasets.coco]: Loading datasets/coco/annotations/instances_train2017.json takes 44.32 seconds.
[12/08 18:46:14 d2.data.datasets.coco]: Loaded 118287 images in COCO format from datasets/coco/annotations/instances_train2017.json
[12/08 18:46:38 d2.data.build]: Removed 1021 images with no usable annotations. 117266 images left.
[12/08 18:46:52 d2.data.build]: Distribution of instances among all 80 categories:

category	#instances	category	#instances	category	#instances
person	257253	bicycle	7056	car	43533
motorcycle	8654	airplane	5129	bus	6061
train	4570	truck	9970	boat	10576
traffic light	12842	fire hydrant	1865	stop sign	1983
parking meter	1283	bench	9820	bird	10542
cat	4766	dog	5500	horse	6567
sheep	9223	cow	8014	elephant	5484
bear	1294	zebra	5269	giraffe	5128
backpack	8714	umbrella	11265	handbag	12342
tie	6448	suitcase	6112	frisbee	2681
skis	6623	snowboard	2681	sports ball	6299
kite	8802	baseball bat	3273	baseball gl..	3747
skateboard	5536	surfboard	6095	tennis racket	4807
bottle	24070	wine glass	7839	cup	20574
fork	5474	knife	7760	spoon	6159
bowl	14323	banana	9195	apple	5776
sandwich	4356	orange	6302	broccoli	7261
carrot	7758	hot dog	2884	pizza	5807
donut	7005	cake	6296	chair	38073
couch	5779	potted plant	8631	bed	4192
dining table	15695	toilet	4149	tv	5803
laptop	4960	mouse	2261	remote	5700
keyboard	2854	cell phone	6422	microwave	1672
oven	3334	toaster	225	sink	5609
refrigerator	2634	book	24077	clock	6320
vase	6577	scissors	1464	teddy bear	4729
hair drier	198	toothbrush	1945
total	849949
[12/08 18:46:52 d2.data.build]: Using training sampler TrainingSampler
[12/08 18:46:53 d2.data.common]: Serializing 117266 elements to byte tensors and concatenating them all ...
[12/08 18:47:04 d2.data.common]: Serialized dataset takes 451.21 MiB
[12/08 18:47:13 d2.engine.defaults]: Model:
GeneralizedRCNN(
(backbone): FPN(

(fpn_lateral3): Conv2d(512, 128, kernel_size=(1, 1), stride=(1, 1))
(fpn_output3): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
(fpn_lateral4): Conv2d(768, 128, kernel_size=(1, 1), stride=(1, 1))
(fpn_output4): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
(fpn_lateral5): Conv2d(1024, 128, kernel_size=(1, 1), stride=(1, 1))
(fpn_output5): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
(top_block): LastLevelP6P7(
  (p6): Conv2d(128, 128, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1))
  (p7): Conv2d(128, 128, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1))
)
(bottom_up): VoVNet(
  (stem): Sequential(
    (stem_1/conv): Conv2d(3, 64, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)
    (stem_1/norm): FrozenBatchNorm2d(num_features=64, eps=1e-05)
    (stem_1/relu): ReLU(inplace=True)
    (stem_2/conv): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
    (stem_2/norm): FrozenBatchNorm2d(num_features=64, eps=1e-05)
    (stem_2/relu): ReLU(inplace=True)
    (stem_3/conv): Conv2d(64, 128, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)
    (stem_3/norm): FrozenBatchNorm2d(num_features=128, eps=1e-05)
    (stem_3/relu): ReLU(inplace=True)
  )
  (stage2): _OSA_stage(
    (OSA2_1): _OSA_module(
      (layers): ModuleList(
        (0): Sequential(
          (OSA2_1_0/conv): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
          (OSA2_1_0/norm): FrozenBatchNorm2d(num_features=128, eps=1e-05)
          (OSA2_1_0/relu): ReLU(inplace=True)
        )
        (1): Sequential(
          (OSA2_1_1/conv): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
          (OSA2_1_1/norm): FrozenBatchNorm2d(num_features=128, eps=1e-05)
          (OSA2_1_1/relu): ReLU(inplace=True)
        )
        (2): Sequential(
          (OSA2_1_2/conv): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
          (OSA2_1_2/norm): FrozenBatchNorm2d(num_features=128, eps=1e-05)
          (OSA2_1_2/relu): ReLU(inplace=True)
        )
      )
      (concat): Sequential(
        (OSA2_1_concat/conv): Conv2d(512, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)
        (OSA2_1_concat/norm): FrozenBatchNorm2d(num_features=256, eps=1e-05)
        (OSA2_1_concat/relu): ReLU(inplace=True)
      )
      (ese): eSEModule(
        (avg_pool): AdaptiveAvgPool2d(output_size=1)
        (fc): Conv2d(256, 256, kernel_size=(1, 1), stride=(1, 1))
        (hsigmoid): Hsigmoid()
      )
    )
  )
  (stage3): _OSA_stage(
    (Pooling): MaxPool2d(kernel_size=3, stride=2, padding=0, dilation=1, ceil_mode=True)
    (OSA3_1): _OSA_module(
      (layers): ModuleList(
        (0): Sequential(
          (OSA3_1_0/conv): Conv2d(256, 160, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
          (OSA3_1_0/norm): FrozenBatchNorm2d(num_features=160, eps=1e-05)
          (OSA3_1_0/relu): ReLU(inplace=True)
        )
        (1): Sequential(
          (OSA3_1_1/conv): Conv2d(160, 160, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
          (OSA3_1_1/norm): FrozenBatchNorm2d(num_features=160, eps=1e-05)
          (OSA3_1_1/relu): ReLU(inplace=True)
        )
        (2): Sequential(
          (OSA3_1_2/conv): Conv2d(160, 160, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
          (OSA3_1_2/norm): FrozenBatchNorm2d(num_features=160, eps=1e-05)
          (OSA3_1_2/relu): ReLU(inplace=True)
        )
      )
      (concat): Sequential(
        (OSA3_1_concat/conv): Conv2d(736, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)
        (OSA3_1_concat/norm): FrozenBatchNorm2d(num_features=512, eps=1e-05)
        (OSA3_1_concat/relu): ReLU(inplace=True)
      )
      (ese): eSEModule(
        (avg_pool): AdaptiveAvgPool2d(output_size=1)
        (fc): Conv2d(512, 512, kernel_size=(1, 1), stride=(1, 1))
        (hsigmoid): Hsigmoid()
      )
    )
  )
  (stage4): _OSA_stage(
    (Pooling): MaxPool2d(kernel_size=3, stride=2, padding=0, dilation=1, ceil_mode=True)
    (OSA4_1): _OSA_module(
      (layers): ModuleList(
        (0): Sequential(
          (OSA4_1_0/conv): Conv2d(512, 192, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
          (OSA4_1_0/norm): FrozenBatchNorm2d(num_features=192, eps=1e-05)
          (OSA4_1_0/relu): ReLU(inplace=True)
        )
        (1): Sequential(
          (OSA4_1_1/conv): Conv2d(192, 192, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
          (OSA4_1_1/norm): FrozenBatchNorm2d(num_features=192, eps=1e-05)
          (OSA4_1_1/relu): ReLU(inplace=True)
        )
        (2): Sequential(
          (OSA4_1_2/conv): Conv2d(192, 192, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
          (OSA4_1_2/norm): FrozenBatchNorm2d(num_features=192, eps=1e-05)
          (OSA4_1_2/relu): ReLU(inplace=True)
        )
      )
      (concat): Sequential(
        (OSA4_1_concat/conv): Conv2d(1088, 768, kernel_size=(1, 1), stride=(1, 1), bias=False)
        (OSA4_1_concat/norm): FrozenBatchNorm2d(num_features=768, eps=1e-05)
        (OSA4_1_concat/relu): ReLU(inplace=True)
      )
      (ese): eSEModule(
        (avg_pool): AdaptiveAvgPool2d(output_size=1)
        (fc): Conv2d(768, 768, kernel_size=(1, 1), stride=(1, 1))
        (hsigmoid): Hsigmoid()
      )
    )
  )
  (stage5): _OSA_stage(
    (Pooling): MaxPool2d(kernel_size=3, stride=2, padding=0, dilation=1, ceil_mode=True)
    (OSA5_1): _OSA_module(
      (layers): ModuleList(
        (0): Sequential(
          (OSA5_1_0/conv): Conv2d(768, 224, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
          (OSA5_1_0/norm): FrozenBatchNorm2d(num_features=224, eps=1e-05)
          (OSA5_1_0/relu): ReLU(inplace=True)
        )
        (1): Sequential(
          (OSA5_1_1/conv): Conv2d(224, 224, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
          (OSA5_1_1/norm): FrozenBatchNorm2d(num_features=224, eps=1e-05)
          (OSA5_1_1/relu): ReLU(inplace=True)
        )
        (2): Sequential(
          (OSA5_1_2/conv): Conv2d(224, 224, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
          (OSA5_1_2/norm): FrozenBatchNorm2d(num_features=224, eps=1e-05)
          (OSA5_1_2/relu): ReLU(inplace=True)
        )
      )
      (concat): Sequential(
        (OSA5_1_concat/conv): Conv2d(1440, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)
        (OSA5_1_concat/norm): FrozenBatchNorm2d(num_features=1024, eps=1e-05)
        (OSA5_1_concat/relu): ReLU(inplace=True)
      )
      (ese): eSEModule(
        (avg_pool): AdaptiveAvgPool2d(output_size=1)
        (fc): Conv2d(1024, 1024, kernel_size=(1, 1), stride=(1, 1))
        (hsigmoid): Hsigmoid()
      )
    )
  )
)

)
(proposal_generator): FCOS(
(iou_loss): IOULoss()
(fcos_head): FCOSHead(
(cls_tower): Sequential(
(0): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
(1): GroupNorm(32, 128, eps=1e-05, affine=True)
(2): ReLU()
(3): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
(4): GroupNorm(32, 128, eps=1e-05, affine=True)
(5): ReLU()
)
(bbox_tower): Sequential(
(0): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
(1): GroupNorm(32, 128, eps=1e-05, affine=True)
(2): ReLU()
(3): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
(4): GroupNorm(32, 128, eps=1e-05, affine=True)
(5): ReLU()
)
(share_tower): Sequential()
(cls_logits): Conv2d(128, 80, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
(bbox_pred): Conv2d(128, 4, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
(ctrness): Conv2d(128, 1, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
(scales): ModuleList(
(0): Scale()
(1): Scale()
(2): Scale()
(3): Scale()
(4): Scale()
)
)
)
(roi_heads): CenterROIHeads(
(mask_pooler): ROIPooler(
(level_poolers): ModuleList(
(0): ROIAlign(output_size=(14, 14), spatial_scale=0.125, sampling_ratio=0, aligned=True)
(1): ROIAlign(output_size=(14, 14), spatial_scale=0.0625, sampling_ratio=0, aligned=True)
(2): ROIAlign(output_size=(14, 14), spatial_scale=0.03125, sampling_ratio=0, aligned=True)
)
)
(mask_head): SpatialAttentionMaskHead(
(mask_fcn1): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
(mask_fcn2): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
(spatialAtt): SpatialAttention(
(conv): Conv2d(2, 1, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(sigmoid): Sigmoid()
)
(deconv): ConvTranspose2d(128, 128, kernel_size=(2, 2), stride=(2, 2))
(predictor): Conv2d(128, 80, kernel_size=(1, 1), stride=(1, 1))
)
(maskiou_head): MaskIoUHead(
(maskiou_fcn1): Conv2d(129, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
(maskiou_fcn2): Conv2d(128, 128, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1))
(maskiou_fc1): Linear(in_features=6272, out_features=1024, bias=True)
(maskiou_fc2): Linear(in_features=1024, out_features=1024, bias=True)
(maskiou): Linear(in_features=1024, out_features=80, bias=True)
(pooling): MaxPool2d(kernel_size=[2, 2], stride=[2, 2], padding=[0, 0], dilation=[1, 1], ceil_mode=False)
)
)
)
[12/08 18:47:13 d2.data.dataset_mapper]: [DatasetMapper] Augmentations used in training: [ResizeShortestEdge(short_edge_length=(580, 600), max_size=1000, sample_style='choice'), RandomFlip()]
[12/08 18:47:55 d2.data.datasets.coco]: Loading datasets/coco/annotations/instances_train2017.json takes 41.12 seconds.
[12/08 18:47:56 d2.data.datasets.coco]: Loaded 118287 images in COCO format from datasets/coco/annotations/instances_train2017.json
[12/08 18:48:20 d2.data.build]: Removed 1021 images with no usable annotations. 117266 images left.
[12/08 18:48:33 d2.data.build]: Using training sampler TrainingSampler
[12/08 18:48:34 d2.data.common]: Serializing 117266 elements to byte tensors and concatenating them all ...
[12/08 18:48:45 d2.data.common]: Serialized dataset takes 451.21 MiB
/jmx/detectron2/detectron2/structures/masks.py:345: UserWarning: This overload of nonzero is deprecated:
nonzero()
Consider using one of the following signatures instead:
nonzero(, bool as_tuple) (Triggered internally at /pytorch/torch/csrc/utils/python_arg_parser.cpp:766.)
item = item.nonzero().squeeze(1).cpu().numpy().tolist()
[12/08 18:48:54 fvcore.common.checkpoint]: Loading checkpoint from /jmx/centermask2/models/vovnet19_ese_detectron2.pth
/jmx/detectron2/detectron2/structures/masks.py:345: UserWarning: This overload of nonzero is deprecated:
nonzero()
Consider using one of the following signatures instead:
nonzero(, bool as_tuple) (Triggered internally at /pytorch/torch/csrc/utils/python_arg_parser.cpp:766.)
item = item.nonzero().squeeze(1).cpu().numpy().tolist()
/jmx/detectron2/detectron2/structures/masks.py:345: UserWarning: This overload of nonzero is deprecated:
nonzero()
Consider using one of the following signatures instead:
nonzero(, bool as_tuple) (Triggered internally at /pytorch/torch/csrc/utils/python_arg_parser.cpp:766.)
item = item.nonzero().squeeze(1).cpu().numpy().tolist()
[12/08 18:48:54 fvcore.common.checkpoint]: Some model parameters or buffers are not found in the checkpoint:
roi_heads.mask_head.mask_fcn2.{bias, weight}
proposal_generator.fcos_head.cls_tower.1.{weight, bias}
roi_heads.maskiou_head.maskiou.{weight, bias}
proposal_generator.fcos_head.ctrness.{weight, bias}
roi_heads.maskiou_head.maskiou_fc1.{weight, bias}
backbone.fpn_lateral3.{weight, bias}
roi_heads.mask_head.spatialAtt.conv.weight
backbone.top_block.p6.{bias, weight}
roi_heads.maskiou_head.maskiou_fcn1.{weight, bias}
proposal_generator.fcos_head.cls_logits.{weight, bias}
backbone.top_block.p7.{bias, weight}
proposal_generator.fcos_head.bbox_tower.1.{weight, bias}
proposal_generator.fcos_head.scales.1.scale
proposal_generator.fcos_head.bbox_tower.4.{bias, weight}
proposal_generator.fcos_head.bbox_pred.{bias, weight}
proposal_generator.fcos_head.scales.2.scale
backbone.fpn_lateral5.{bias, weight}
roi_heads.maskiou_head.maskiou_fcn2.{bias, weight}
backbone.fpn_lateral4.{bias, weight}
roi_heads.mask_head.deconv.{weight, bias}
roi_heads.mask_head.mask_fcn1.{bias, weight}
proposal_generator.fcos_head.scales.3.scale
roi_heads.mask_head.predictor.{weight, bias}
backbone.fpn_output3.{weight, bias}
proposal_generator.fcos_head.cls_tower.3.{bias, weight}
proposal_generator.fcos_head.scales.0.scale
proposal_generator.fcos_head.bbox_tower.0.{weight, bias}
backbone.fpn_output5.{bias, weight}
proposal_generator.fcos_head.cls_tower.0.{weight, bias}
proposal_generator.fcos_head.cls_tower.4.{bias, weight}
roi_heads.maskiou_head.maskiou_fc2.{bias, weight}
backbone.fpn_output4.{bias, weight}
proposal_generator.fcos_head.bbox_tower.3.{bias, weight}
proposal_generator.fcos_head.scales.4.scale
[12/08 18:48:54 fvcore.common.checkpoint]: The checkpoint state_dict contains keys that are not used by the model:
backbone.bottom_up.stem.stem_1/norm.num_batches_tracked
backbone.bottom_up.stem.stem_2/norm.num_batches_tracked
backbone.bottom_up.stem.stem_3/norm.num_batches_tracked
backbone.bottom_up.stage2.OSA2_1.layers.0.OSA2_1_0/norm.num_batches_tracked
backbone.bottom_up.stage2.OSA2_1.layers.1.OSA2_1_1/norm.num_batches_tracked
backbone.bottom_up.stage2.OSA2_1.layers.2.OSA2_1_2/norm.num_batches_tracked
backbone.bottom_up.stage2.OSA2_1.concat.OSA2_1_concat/norm.num_batches_tracked
backbone.bottom_up.stage3.OSA3_1.layers.0.OSA3_1_0/norm.num_batches_tracked
backbone.bottom_up.stage3.OSA3_1.layers.1.OSA3_1_1/norm.num_batches_tracked
backbone.bottom_up.stage3.OSA3_1.layers.2.OSA3_1_2/norm.num_batches_tracked
backbone.bottom_up.stage3.OSA3_1.concat.OSA3_1_concat/norm.num_batches_tracked
backbone.bottom_up.stage4.OSA4_1.layers.0.OSA4_1_0/norm.num_batches_tracked
backbone.bottom_up.stage4.OSA4_1.layers.1.OSA4_1_1/norm.num_batches_tracked
backbone.bottom_up.stage4.OSA4_1.layers.2.OSA4_1_2/norm.num_batches_tracked
backbone.bottom_up.stage4.OSA4_1.concat.OSA4_1_concat/norm.num_batches_tracked
backbone.bottom_up.stage5.OSA5_1.layers.0.OSA5_1_0/norm.num_batches_tracked
backbone.bottom_up.stage5.OSA5_1.layers.1.OSA5_1_1/norm.num_batches_tracked
backbone.bottom_up.stage5.OSA5_1.layers.2.OSA5_1_2/norm.num_batches_tracked
backbone.bottom_up.stage5.OSA5_1.concat.OSA5_1_concat/norm.num_batches_tracked
/jmx/detectron2/detectron2/structures/masks.py:345: UserWarning: This overload of nonzero is deprecated:
nonzero()
Consider using one of the following signatures instead:
nonzero(, bool as_tuple) (Triggered internally at /pytorch/torch/csrc/utils/python_arg_parser.cpp:766.)
item = item.nonzero().squeeze(1).cpu().numpy().tolist()
[W TensorIterator.cpp:918] Warning: Mixed memory format inputs detected while calling the operator. The operator will output contiguous tensor even if some of the inputs are in channels_last format. (function operator())
[W TensorIterator.cpp:918] Warning: Mixed memory format inputs detected while calling the operator. The operator will output contiguous tensor even if some of the inputs are in channels_last format. (function operator())
[W TensorIterator.cpp:918] Warning: Mixed memory format inputs detected while calling the operator. The operator will output contiguous tensor even if some of the inputs are in channels_last format. (function operator())
[W TensorIterator.cpp:918] Warning: Mixed memory format inputs detected while calling the operator. The operator will output contiguous tensor even if some of the inputs are in channels_last format. (function operator())
/root/anaconda3/envs/detectron2/lib/python3.6/site-packages/torch/optim/lr_scheduler.py:123: UserWarning: Detected call of lr_scheduler.step() before optimizer.step(). In PyTorch 1.1.0 and later, you should call them in the opposite order: optimizer.step() before lr_scheduler.step(). Failure to do this will result in PyTorch skipping the first value of the learning rate schedule. See more details at https://pytorch.org/docs/stable/optim.html#how-to-adjust-learning-rate
"https://pytorch.org/docs/stable/optim.html#how-to-adjust-learning-rate", UserWarning)
/root/anaconda3/envs/detectron2/lib/python3.6/site-packages/torch/optim/lr_scheduler.py:123: UserWarning: Detected call of lr_scheduler.step() before optimizer.step(). In PyTorch 1.1.0 and later, you should call them in the opposite order: optimizer.step() before lr_scheduler.step(). Failure to do this will result in PyTorch skipping the first value of the learning rate schedule. See more details at https://pytorch.org/docs/stable/optim.html#how-to-adjust-learning-rate
"https://pytorch.org/docs/stable/optim.html#how-to-adjust-learning-rate", UserWarning)
/root/anaconda3/envs/detectron2/lib/python3.6/site-packages/torch/optim/lr_scheduler.py:123: UserWarning: Detected call of lr_scheduler.step() before optimizer.step(). In PyTorch 1.1.0 and later, you should call them in the opposite order: optimizer.step() before lr_scheduler.step(). Failure to do this will result in PyTorch skipping the first value of the learning rate schedule. See more details at https://pytorch.org/docs/stable/optim.html#how-to-adjust-learning-rate
"https://pytorch.org/docs/stable/optim.html#how-to-adjust-learning-rate", UserWarning)
/root/anaconda3/envs/detectron2/lib/python3.6/site-packages/torch/optim/lr_scheduler.py:123: UserWarning: Detected call of lr_scheduler.step() before optimizer.step(). In PyTorch 1.1.0 and later, you should call them in the opposite order: optimizer.step() before lr_scheduler.step(). Failure to do this will result in PyTorch skipping the first value of the learning rate schedule. See more details at https://pytorch.org/docs/stable/optim.html#how-to-adjust-learning-rate
"https://pytorch.org/docs/stable/optim.html#how-to-adjust-learning-rate", UserWarning)
[12/08 18:49:08 d2.utils.events]: eta: 2 days, 12:45:56 iter: 19 total_loss: 3.564 loss_mask: 0.6932 loss_maskiou: 0.06747 loss_fcos_cls: 1.14 loss_fcos_loc: 0.9648 loss_fcos_ctr: 0.6922 time: 0.6086 data_time: 0.2780 lr: 1e-05 max_mem: 2960M
[12/08 18:49:08 d2.utils.events]: eta: 2 days, 13:25:44 iter: 19 total_loss: 3.564 loss_mask: 0.6932 loss_maskiou: 0.06747 loss_fcos_cls: 1.14 loss_fcos_loc: 0.9648 loss_fcos_ctr: 0.6922 time: 0.6118 data_time: 0.2780 lr: 1e-05 max_mem: 2960M

[haderalim]

@jiameixia1202 I mean by "the registered coco dataset" you use coco dataset not custom dataset.

You can debug and check the pre_nms_top_n value in ''centermask/modeling/fcos/fcos_outputs/'

lines are:-
candidate_inds = box_cls > self.pre_nms_thresh
pre_nms_top_n = candidate_inds.view(N, -1).sum(1)

When cfg.SCORE_THRESH_TEST = 0.05 then self.pre_nms_thresh = 0.05. This value may cause of make pre_nms_top_n empty and make zero accuracy.

So, check if pre_nms_top_n is tensor([0, 0] or not.

[jiameixia1202]

I use the registered coco dataset, and I check the pre_ nms_ top_ n is tensor([0, 0] .

It is normal for me to test with the weight provided by you, but when I train myself, the accuracy is 0 (without modifying the network, data and parameters)

[jiameixia1202]

cfg.SCORE_THRESH_TEST = 0.05 then self.pre_nms_thresh = 0.05 How do I modify these two parameters？

[haderalim]

You can modify it through config file ''configs/centermask/centermask_lite_V_19_eSE_FPN_ms_4x.yaml' under MODEL and FCOS. If FCOS not exist, add it and modify the value 0.05 of INFERENCE_TH_TRAIN and INFERENCE_TH_TEST.

Like this:
MODEL:
FCOS:
INFERENCE_TH_TRAIN: 0.05
INFERENCE_TH_TEST: 0.05

[jiameixia1202]

I'm sorry, there may be something wrong with what I just said,
My current settings cfg.SCORE_ THRESH_ TEST = 0.05 then self.pre_ nms_ thresh = 0.05
You said this value may cause of make pre_ nms_ top_ n empty and make zero accuracy.
I wonder why I can avoid this problem by modifying these two values

[Paragjain10]

You can modify it through config file ''configs/centermask/centermask_lite_V_19_eSE_FPN_ms_4x.yaml' under MODEL and FCOS. If FCOS not exist, add it and modify the value 0.05 of INFERENCE_TH_TRAIN and INFERENCE_TH_TEST.

Like this:
MODEL:
FCOS:
INFERENCE_TH_TRAIN: 0.05
INFERENCE_TH_TEST: 0.05

can we change the optimizer to Adam?