youngwanlee/centermask

[CVPR 2020] CenterMask : Real-Time Anchor-Free Instance Segmentation

👋  Hi there! I'm Youngwan, a senior researcher at ETRI and Ph.D student in Graduate school of AI at KAIST, where I'm advised by Prof. Sung Ju Hwang in the Machine Learning and Artificial Intelligence (MLAI) lab.

My research interest is how computers understand the world, including efficient 2D/3D neural network design, object detection, instance segmentation, semantic segmentation, and video classification. 🖥️🌏

Representative publications and Codes

See Google scholar for full list.

• RC-MAE: Exploring the Role of Mean Teachers in Self-supervised Masked Auto-Encoders, ICLR 2023.
• MPViT : Multi-Path Vision Transformer for Dense Prediction, CVPR 2022.
• CenterMask : Real-Time Anchor-Free Instance Segmentation, CVPR 2020.
• 2D convolutional neural network : VoVNet
• 3D convolutional neural network : VoV3D

About me

• 📝 I enjoy teaching talking what I know learn, so I am giving lectures on AI as an AI Facilitator at ETRI AI Academy.
  
• 🌏🌱🌲🌊 ⛰️ I love to appreciate the beautiful nature.
  
• 🎾 🏀 I enjoy playing tennis and basket ball.
  
• 📫 How to reach me: [email protected] | [email protected]

💪 Skills

Platforms & Languages

centermask's Issues

训练的权重文件不存在

当我下载训练权重时，网址内容是空的

BBox coordinates and Mask

CPU version not implemented - ml_nms

Hey!

I was wondering if there is any ETA on a possible CPU implementation of the ml_nms.cu for forward-passes on the CPU? Would be nice to be able to run the demo on CPU.

Kindly,
Christoffer

Intel MKL FATAL ERROR: Cannot load libmkl_core.so.

Hi,

When I run following command
python setup.py build_ext install

as given in installation guideline, I get the error given below:

Intel MKL FATAL ERROR: Cannot load libmkl_core.so.

Any idea how to resolve it?

Evaluation example not working

I tried to run the single-gpu evaluation & test batch size 1 example.

The link for wget (https://www.dropbox.com/s/2enqxenccz4xy6l/centermask-lite-R-50-ms-bs32-1x.pth) is not available.

So I modified the example as follows:

wget https://www.dropbox.com/s/alifk31z3roife1/centermask-lite-V-19-eSE-ms-bs16-4x.pth?dl=1 -O centermask-lite-V-19-eSE-ms-bs16-4x.pth

python tools/test_net.py --config-file "configs/centermask/centermask_V_19_eSE_FPN_lite_res600_ms_bs16_4x.yaml" TEST.IMS_PER_BATCH 1 MODEL.WEIGHT centermask-lite-V-19-eSE-ms-bs16-4x.pth

I get IndexError: list index out of range caused by selected_polygons.append(self.polygons[i])

I have downloaded and extracted coco_annotations_minival.tgz into datasets/coco/annotations/, and val2014 into datasets/coco/val2014.

Is there something more I need, in order to run this demonstrator?

Here is the full output:

BATCH 1 MODEL.WEIGHT centermask-lite-V-19-eSE-ms-bs16-4x.pth
2020-03-22 19:04:53,347 maskrcnn_benchmark INFO: Using 1 GPUs
2020-03-22 19:04:53,348 maskrcnn_benchmark INFO: DATALOADER:
  ASPECT_RATIO_GROUPING: True
  NUM_WORKERS: 4
  SIZE_DIVISIBILITY: 32
DATASETS:
  TEST: ('coco_2014_minival',)
  TRAIN: ('coco_2014_train', 'coco_2014_valminusminival')
INPUT:
  MAX_SIZE_TEST: 1000
  MAX_SIZE_TRAIN: 1000
  MIN_SIZE_RANGE_TRAIN: (580, 600)
  MIN_SIZE_TEST: 600
  MIN_SIZE_TRAIN: (800,)
  PIXEL_MEAN: [102.9801, 115.9465, 122.7717]
  PIXEL_STD: [1.0, 1.0, 1.0]
  TO_BGR255: True
MODEL:
  BACKBONE:
    CONV_BODY: V-19-eSE-FPN-RETINANET
    FREEZE_CONV_BODY_AT: 0
    USE_GN: False
  CLS_AGNOSTIC_BBOX_REG: False
  DEVICE: cuda
  FBNET:
    ARCH: default
    ARCH_DEF: 
    BN_TYPE: bn
    DET_HEAD_BLOCKS: []
    DET_HEAD_LAST_SCALE: 1.0
    DET_HEAD_STRIDE: 0
    DW_CONV_SKIP_BN: True
    DW_CONV_SKIP_RELU: True
    KPTS_HEAD_BLOCKS: []
    KPTS_HEAD_LAST_SCALE: 0.0
    KPTS_HEAD_STRIDE: 0
    MASK_HEAD_BLOCKS: []
    MASK_HEAD_LAST_SCALE: 0.0
    MASK_HEAD_STRIDE: 0
    RPN_BN_TYPE: 
    RPN_HEAD_BLOCKS: 0
    SCALE_FACTOR: 1.0
    WIDTH_DIVISOR: 1
  FCOS:
    CENTER_SAMPLE: True
    DENSE_POINTS: 1
    FPN_STRIDES: [8, 16, 32, 64, 128]
    HEAD: FCOSHead
    INFERENCE_TH: 0.03
    LOC_LOSS_TYPE: giou
    LOSS_ALPHA: 0.25
    LOSS_GAMMA: 2.0
    NMS_TH: 0.6
    NUM_CLASSES: 81
    NUM_CONVS: 2
    POST_NMS_TOP_N_TRAIN: 100
    POS_RADIUS: 1.5
    PRE_NMS_TOP_N: 1000
    PRIOR_PROB: 0.01
    RESIDUAL_CONNECTION: False
    TARGET_ASSIGN: [[-1, 64], [64, 128], [128, 256], [256, 512], [512, 100000000]]
  FCOS_MASK: True
  FCOS_ON: True
  FPN:
    USE_GN: False
    USE_RELU: False
  GROUP_NORM:
    DIM_PER_GP: -1
    EPSILON: 1e-05
    NUM_GROUPS: 32
  HRNET:
    FPN:
      CONV_STRIDE: 2
      OUT_CHANNEL: 256
      TYPE: HRFPN
  KEYPOINT_ON: False
  MASKIOU_LOSS_WEIGHT: 1.0
  MASKIOU_ON: True
  MASK_ON: True
  META_ARCHITECTURE: GeneralizedRCNN
  NECK:
    ACTIVATION: False
    IN_CHANNELS: (32, 64, 128, 256)
    NUM_OUTS: 5
    OUT_CHANNELS: 256
    POOLING:  AVG
    SHARING_CONV: False
  RESNETS:
    BACKBONE_OUT_CHANNELS: 1024
    NUM_GROUPS: 1
    RES2_OUT_CHANNELS: 256
    RES5_DILATION: 1
    STEM_FUNC: StemWithFixedBatchNorm
    STEM_OUT_CHANNELS: 64
    STRIDE_IN_1X1: True
    TRANS_FUNC: BottleneckWithFixedBatchNorm
    WIDTH_PER_GROUP: 64
  RETINANET:
    ANCHOR_SIZES: (32, 64, 128, 256, 512)
    ANCHOR_STRIDES: (8, 16, 32, 64, 128)
    ASPECT_RATIOS: (0.5, 1.0, 2.0)
    BBOX_REG_BETA: 0.11
    BBOX_REG_WEIGHT: 4.0
    BG_IOU_THRESHOLD: 0.4
    FG_IOU_THRESHOLD: 0.5
    INFERENCE_TH: 0.05
    LOSS_ALPHA: 0.25
    LOSS_GAMMA: 2.0
    NMS_TH: 0.4
    NUM_CLASSES: 81
    NUM_CONVS: 4
    OCTAVE: 2.0
    PRE_NMS_TOP_N: 1000
    PRIOR_PROB: 0.01
    SCALES_PER_OCTAVE: 3
    STRADDLE_THRESH: 0
    USE_C5: False
  RETINANET_ON: False
  ROI_BOX_HEAD:
    CONV_HEAD_DIM: 256
    DILATION: 1
    FEATURE_EXTRACTOR: ResNet50Conv5ROIFeatureExtractor
    MLP_HEAD_DIM: 1024
    NUM_CLASSES: 81
    NUM_STACKED_CONVS: 4
    POOLER_RESOLUTION: 14
    POOLER_SAMPLING_RATIO: 0
    POOLER_SCALES: (0.0625,)
    PREDICTOR: FastRCNNPredictor
    USE_GN: False
  ROI_HEADS:
    BATCH_SIZE_PER_IMAGE: 512
    BBOX_REG_WEIGHTS: (10.0, 10.0, 5.0, 5.0)
    BG_IOU_THRESHOLD: 0.5
    DETECTIONS_PER_IMG: 100
    FG_IOU_THRESHOLD: 0.5
    NMS: 0.5
    POSITIVE_FRACTION: 0.25
    SCORE_THRESH: 0.05
    USE_FPN: True
  ROI_KEYPOINT_HEAD:
    CONV_LAYERS: (512, 512, 512, 512, 512, 512, 512, 512)
    FEATURE_EXTRACTOR: KeypointRCNNFeatureExtractor
    MLP_HEAD_DIM: 1024
    NUM_CLASSES: 17
    POOLER_RESOLUTION: 14
    POOLER_SAMPLING_RATIO: 0
    POOLER_SCALES: (0.0625,)
    PREDICTOR: KeypointRCNNPredictor
    RESOLUTION: 14
    SHARE_BOX_FEATURE_EXTRACTOR: True
  ROI_MASKIOU_HEAD:
    CONV_LAYERS: (128, 128)
    FEATURE_EXTRACTOR: MaskIoUFeatureExtractor
    PREDICTOR: MaskIoUPredictor
  ROI_MASK_HEAD:
    CONV_LAYERS: (128, 128)
    DILATION: 1
    FEATURE_EXTRACTOR: MaskRCNNFPNSpatialAttentionFeatureExtractor
    LEVEL_MAP_FUNCTION: CenterMaskLevelMapFunc
    MLP_HEAD_DIM: 1024
    POOLER_RESOLUTION: 14
    POOLER_SAMPLING_RATIO: 2
    POOLER_SCALES: (0.125, 0.0625, 0.03125)
    POSTPROCESS_MASKS: False
    POSTPROCESS_MASKS_THRESHOLD: 0.5
    PREDICTOR: MaskRCNNC4Predictor
    RESOLUTION: 28
    SHARE_BOX_FEATURE_EXTRACTOR: False
    USE_GN: False
  RPN:
    ANCHOR_SIZES: (32, 64, 128, 256, 512)
    ANCHOR_STRIDE: (16,)
    ASPECT_RATIOS: (0.5, 1.0, 2.0)
    BATCH_SIZE_PER_IMAGE: 256
    BG_IOU_THRESHOLD: 0.3
    FG_IOU_THRESHOLD: 0.7
    FPN_POST_NMS_TOP_N_TEST: 2000
    FPN_POST_NMS_TOP_N_TRAIN: 2000
    MIN_SIZE: 0
    NMS_THRESH: 0.7
    POSITIVE_FRACTION: 0.5
    POST_NMS_TOP_N_TEST: 1000
    POST_NMS_TOP_N_TRAIN: 2000
    PRE_NMS_TOP_N_TEST: 6000
    PRE_NMS_TOP_N_TRAIN: 12000
    RPN_HEAD: SingleConvRPNHead
    STRADDLE_THRESH: 0
    USE_FPN: False
  RPN_ONLY: True
  USE_SYNCBN: False
  VOVNET:
    BACKBONE_OUT_CHANNELS: 128
    OUT_CHANNELS: 256
    PRETRAINED: 
    USE_GN: False
  WEIGHT: centermask-lite-V-19-eSE-ms-bs16-4x.pth
OUTPUT_DIR: checkpoints/CenterMask-Lite-V-19-eSE-FPN-ress600-ms-b16-4x
PATHS_CATALOG: /home/panda/install/CenterMask/maskrcnn_benchmark/config/paths_catalog.py
SOLVER:
  BASE_LR: 0.01
  BIAS_LR_FACTOR: 2
  CHECKPOINT_PERIOD: 10000
  GAMMA: 0.1
  IMS_PER_BATCH: 16
  MAX_ITER: 360000
  MOMENTUM: 0.9
  STEPS: (300000, 340000)
  TEST_PERIOD: 0
  WARMUP_FACTOR: 0.3333333333333333
  WARMUP_ITERS: 500
  WARMUP_METHOD: constant
  WEIGHT_DECAY: 0.0001
  WEIGHT_DECAY_BIAS: 0
TEST:
  DETECTIONS_PER_IMG: 50
  EXPECTED_RESULTS: []
  EXPECTED_RESULTS_SIGMA_TOL: 4
  IMS_PER_BATCH: 1
2020-03-22 19:04:53,351 maskrcnn_benchmark INFO: Collecting env info (might take some time)
2020-03-22 19:05:03,397 maskrcnn_benchmark INFO: 
PyTorch version: 1.4.0
Is debug build: No
CUDA used to build PyTorch: 10.0

OS: Ubuntu 18.04.3 LTS
GCC version: (Ubuntu/Linaro 7.5.0-3ubuntu1~18.04) 7.5.0
CMake version: version 3.10.2

Python version: 3.6
Is CUDA available: Yes
CUDA runtime version: 10.0.326
GPU models and configuration: Could not collect
Nvidia driver version: Could not collect
cuDNN version: /usr/lib/aarch64-linux-gnu/libcudnn.so.7.6.3

Versions of relevant libraries:
[pip3] numpy==1.18.2
[pip3] torch==1.4.0
[pip3] torchvision==0.5.0a0+85b8fbf
[conda] Could not collect
        Pillow (7.0.0)
2020-03-22 19:05:12,051 maskrcnn_benchmark.utils.checkpoint INFO: Loading checkpoint from centermask-lite-V-19-eSE-ms-bs16-4x.pth
2020-03-22 19:05:13,820 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stage2.OSA2_1.concat.OSA2_1_concat/conv.weight       loaded from backbone.body.stage2.OSA2_1.concat.OSA2_1_concat/conv.weight       of shape (256, 512, 1, 1)
2020-03-22 19:05:13,820 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stage2.OSA2_1.concat.OSA2_1_concat/norm.bias         loaded from backbone.body.stage2.OSA2_1.concat.OSA2_1_concat/norm.bias         of shape (256,)
2020-03-22 19:05:13,821 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stage2.OSA2_1.concat.OSA2_1_concat/norm.running_mean loaded from backbone.body.stage2.OSA2_1.concat.OSA2_1_concat/norm.running_mean of shape (256,)
2020-03-22 19:05:13,821 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stage2.OSA2_1.concat.OSA2_1_concat/norm.running_var  loaded from backbone.body.stage2.OSA2_1.concat.OSA2_1_concat/norm.running_var  of shape (256,)
2020-03-22 19:05:13,821 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stage2.OSA2_1.concat.OSA2_1_concat/norm.weight       loaded from backbone.body.stage2.OSA2_1.concat.OSA2_1_concat/norm.weight       of shape (256,)
2020-03-22 19:05:13,821 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stage2.OSA2_1.ese.fc.bias                            loaded from backbone.body.stage2.OSA2_1.ese.fc.bias                            of shape (256,)
2020-03-22 19:05:13,821 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stage2.OSA2_1.ese.fc.weight                          loaded from backbone.body.stage2.OSA2_1.ese.fc.weight                          of shape (256, 256, 1, 1)
2020-03-22 19:05:13,822 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stage2.OSA2_1.layers.0.OSA2_1_0/conv.weight          loaded from backbone.body.stage2.OSA2_1.layers.0.OSA2_1_0/conv.weight          of shape (128, 128, 3, 3)
2020-03-22 19:05:13,822 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stage2.OSA2_1.layers.0.OSA2_1_0/norm.bias            loaded from backbone.body.stage2.OSA2_1.layers.0.OSA2_1_0/norm.bias            of shape (128,)
2020-03-22 19:05:13,822 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stage2.OSA2_1.layers.0.OSA2_1_0/norm.running_mean    loaded from backbone.body.stage2.OSA2_1.layers.0.OSA2_1_0/norm.running_mean    of shape (128,)
2020-03-22 19:05:13,822 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stage2.OSA2_1.layers.0.OSA2_1_0/norm.running_var     loaded from backbone.body.stage2.OSA2_1.layers.0.OSA2_1_0/norm.running_var     of shape (128,)
2020-03-22 19:05:13,822 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stage2.OSA2_1.layers.0.OSA2_1_0/norm.weight          loaded from backbone.body.stage2.OSA2_1.layers.0.OSA2_1_0/norm.weight          of shape (128,)
2020-03-22 19:05:13,822 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stage2.OSA2_1.layers.1.OSA2_1_1/conv.weight          loaded from backbone.body.stage2.OSA2_1.layers.1.OSA2_1_1/conv.weight          of shape (128, 128, 3, 3)
2020-03-22 19:05:13,823 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stage2.OSA2_1.layers.1.OSA2_1_1/norm.bias            loaded from backbone.body.stage2.OSA2_1.layers.1.OSA2_1_1/norm.bias            of shape (128,)
2020-03-22 19:05:13,823 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stage2.OSA2_1.layers.1.OSA2_1_1/norm.running_mean    loaded from backbone.body.stage2.OSA2_1.layers.1.OSA2_1_1/norm.running_mean    of shape (128,)
2020-03-22 19:05:13,823 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stage2.OSA2_1.layers.1.OSA2_1_1/norm.running_var     loaded from backbone.body.stage2.OSA2_1.layers.1.OSA2_1_1/norm.running_var     of shape (128,)
2020-03-22 19:05:13,823 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stage2.OSA2_1.layers.1.OSA2_1_1/norm.weight          loaded from backbone.body.stage2.OSA2_1.layers.1.OSA2_1_1/norm.weight          of shape (128,)
2020-03-22 19:05:13,823 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stage2.OSA2_1.layers.2.OSA2_1_2/conv.weight          loaded from backbone.body.stage2.OSA2_1.layers.2.OSA2_1_2/conv.weight          of shape (128, 128, 3, 3)
2020-03-22 19:05:13,823 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stage2.OSA2_1.layers.2.OSA2_1_2/norm.bias            loaded from backbone.body.stage2.OSA2_1.layers.2.OSA2_1_2/norm.bias            of shape (128,)
2020-03-22 19:05:13,824 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stage2.OSA2_1.layers.2.OSA2_1_2/norm.running_mean    loaded from backbone.body.stage2.OSA2_1.layers.2.OSA2_1_2/norm.running_mean    of shape (128,)
2020-03-22 19:05:13,824 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stage2.OSA2_1.layers.2.OSA2_1_2/norm.running_var     loaded from backbone.body.stage2.OSA2_1.layers.2.OSA2_1_2/norm.running_var     of shape (128,)
2020-03-22 19:05:13,824 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stage2.OSA2_1.layers.2.OSA2_1_2/norm.weight          loaded from backbone.body.stage2.OSA2_1.layers.2.OSA2_1_2/norm.weight          of shape (128,)
2020-03-22 19:05:13,824 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stage3.OSA3_1.concat.OSA3_1_concat/conv.weight       loaded from backbone.body.stage3.OSA3_1.concat.OSA3_1_concat/conv.weight       of shape (512, 736, 1, 1)
2020-03-22 19:05:13,824 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stage3.OSA3_1.concat.OSA3_1_concat/norm.bias         loaded from backbone.body.stage3.OSA3_1.concat.OSA3_1_concat/norm.bias         of shape (512,)
2020-03-22 19:05:13,824 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stage3.OSA3_1.concat.OSA3_1_concat/norm.running_mean loaded from backbone.body.stage3.OSA3_1.concat.OSA3_1_concat/norm.running_mean of shape (512,)
2020-03-22 19:05:13,825 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stage3.OSA3_1.concat.OSA3_1_concat/norm.running_var  loaded from backbone.body.stage3.OSA3_1.concat.OSA3_1_concat/norm.running_var  of shape (512,)
2020-03-22 19:05:13,825 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stage3.OSA3_1.concat.OSA3_1_concat/norm.weight       loaded from backbone.body.stage3.OSA3_1.concat.OSA3_1_concat/norm.weight       of shape (512,)
2020-03-22 19:05:13,825 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stage3.OSA3_1.ese.fc.bias                            loaded from backbone.body.stage3.OSA3_1.ese.fc.bias                            of shape (512,)
2020-03-22 19:05:13,825 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stage3.OSA3_1.ese.fc.weight                          loaded from backbone.body.stage3.OSA3_1.ese.fc.weight                          of shape (512, 512, 1, 1)
2020-03-22 19:05:13,825 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stage3.OSA3_1.layers.0.OSA3_1_0/conv.weight          loaded from backbone.body.stage3.OSA3_1.layers.0.OSA3_1_0/conv.weight          of shape (160, 256, 3, 3)
2020-03-22 19:05:13,825 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stage3.OSA3_1.layers.0.OSA3_1_0/norm.bias            loaded from backbone.body.stage3.OSA3_1.layers.0.OSA3_1_0/norm.bias            of shape (160,)
2020-03-22 19:05:13,826 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stage3.OSA3_1.layers.0.OSA3_1_0/norm.running_mean    loaded from backbone.body.stage3.OSA3_1.layers.0.OSA3_1_0/norm.running_mean    of shape (160,)
2020-03-22 19:05:13,826 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stage3.OSA3_1.layers.0.OSA3_1_0/norm.running_var     loaded from backbone.body.stage3.OSA3_1.layers.0.OSA3_1_0/norm.running_var     of shape (160,)
2020-03-22 19:05:13,826 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stage3.OSA3_1.layers.0.OSA3_1_0/norm.weight          loaded from backbone.body.stage3.OSA3_1.layers.0.OSA3_1_0/norm.weight          of shape (160,)
2020-03-22 19:05:13,826 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stage3.OSA3_1.layers.1.OSA3_1_1/conv.weight          loaded from backbone.body.stage3.OSA3_1.layers.1.OSA3_1_1/conv.weight          of shape (160, 160, 3, 3)
2020-03-22 19:05:13,826 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stage3.OSA3_1.layers.1.OSA3_1_1/norm.bias            loaded from backbone.body.stage3.OSA3_1.layers.1.OSA3_1_1/norm.bias            of shape (160,)
2020-03-22 19:05:13,827 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stage3.OSA3_1.layers.1.OSA3_1_1/norm.running_mean    loaded from backbone.body.stage3.OSA3_1.layers.1.OSA3_1_1/norm.running_mean    of shape (160,)
2020-03-22 19:05:13,827 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stage3.OSA3_1.layers.1.OSA3_1_1/norm.running_var     loaded from backbone.body.stage3.OSA3_1.layers.1.OSA3_1_1/norm.running_var     of shape (160,)
2020-03-22 19:05:13,827 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stage3.OSA3_1.layers.1.OSA3_1_1/norm.weight          loaded from backbone.body.stage3.OSA3_1.layers.1.OSA3_1_1/norm.weight          of shape (160,)
2020-03-22 19:05:13,827 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stage3.OSA3_1.layers.2.OSA3_1_2/conv.weight          loaded from backbone.body.stage3.OSA3_1.layers.2.OSA3_1_2/conv.weight          of shape (160, 160, 3, 3)
2020-03-22 19:05:13,827 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stage3.OSA3_1.layers.2.OSA3_1_2/norm.bias            loaded from backbone.body.stage3.OSA3_1.layers.2.OSA3_1_2/norm.bias            of shape (160,)
2020-03-22 19:05:13,828 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stage3.OSA3_1.layers.2.OSA3_1_2/norm.running_mean    loaded from backbone.body.stage3.OSA3_1.layers.2.OSA3_1_2/norm.running_mean    of shape (160,)
2020-03-22 19:05:13,828 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stage3.OSA3_1.layers.2.OSA3_1_2/norm.running_var     loaded from backbone.body.stage3.OSA3_1.layers.2.OSA3_1_2/norm.running_var     of shape (160,)
2020-03-22 19:05:13,828 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stage3.OSA3_1.layers.2.OSA3_1_2/norm.weight          loaded from backbone.body.stage3.OSA3_1.layers.2.OSA3_1_2/norm.weight          of shape (160,)
2020-03-22 19:05:13,828 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stage4.OSA4_1.concat.OSA4_1_concat/conv.weight       loaded from backbone.body.stage4.OSA4_1.concat.OSA4_1_concat/conv.weight       of shape (768, 1088, 1, 1)
2020-03-22 19:05:13,828 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stage4.OSA4_1.concat.OSA4_1_concat/norm.bias         loaded from backbone.body.stage4.OSA4_1.concat.OSA4_1_concat/norm.bias         of shape (768,)
2020-03-22 19:05:13,829 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stage4.OSA4_1.concat.OSA4_1_concat/norm.running_mean loaded from backbone.body.stage4.OSA4_1.concat.OSA4_1_concat/norm.running_mean of shape (768,)
2020-03-22 19:05:13,829 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stage4.OSA4_1.concat.OSA4_1_concat/norm.running_var  loaded from backbone.body.stage4.OSA4_1.concat.OSA4_1_concat/norm.running_var  of shape (768,)
2020-03-22 19:05:13,829 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stage4.OSA4_1.concat.OSA4_1_concat/norm.weight       loaded from backbone.body.stage4.OSA4_1.concat.OSA4_1_concat/norm.weight       of shape (768,)
2020-03-22 19:05:13,829 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stage4.OSA4_1.ese.fc.bias                            loaded from backbone.body.stage4.OSA4_1.ese.fc.bias                            of shape (768,)
2020-03-22 19:05:13,829 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stage4.OSA4_1.ese.fc.weight                          loaded from backbone.body.stage4.OSA4_1.ese.fc.weight                          of shape (768, 768, 1, 1)
2020-03-22 19:05:13,830 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stage4.OSA4_1.layers.0.OSA4_1_0/conv.weight          loaded from backbone.body.stage4.OSA4_1.layers.0.OSA4_1_0/conv.weight          of shape (192, 512, 3, 3)
2020-03-22 19:05:13,830 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stage4.OSA4_1.layers.0.OSA4_1_0/norm.bias            loaded from backbone.body.stage4.OSA4_1.layers.0.OSA4_1_0/norm.bias            of shape (192,)
2020-03-22 19:05:13,830 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stage4.OSA4_1.layers.0.OSA4_1_0/norm.running_mean    loaded from backbone.body.stage4.OSA4_1.layers.0.OSA4_1_0/norm.running_mean    of shape (192,)
2020-03-22 19:05:13,830 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stage4.OSA4_1.layers.0.OSA4_1_0/norm.running_var     loaded from backbone.body.stage4.OSA4_1.layers.0.OSA4_1_0/norm.running_var     of shape (192,)
2020-03-22 19:05:13,830 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stage4.OSA4_1.layers.0.OSA4_1_0/norm.weight          loaded from backbone.body.stage4.OSA4_1.layers.0.OSA4_1_0/norm.weight          of shape (192,)
2020-03-22 19:05:13,831 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stage4.OSA4_1.layers.1.OSA4_1_1/conv.weight          loaded from backbone.body.stage4.OSA4_1.layers.1.OSA4_1_1/conv.weight          of shape (192, 192, 3, 3)
2020-03-22 19:05:13,831 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stage4.OSA4_1.layers.1.OSA4_1_1/norm.bias            loaded from backbone.body.stage4.OSA4_1.layers.1.OSA4_1_1/norm.bias            of shape (192,)
2020-03-22 19:05:13,831 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stage4.OSA4_1.layers.1.OSA4_1_1/norm.running_mean    loaded from backbone.body.stage4.OSA4_1.layers.1.OSA4_1_1/norm.running_mean    of shape (192,)
2020-03-22 19:05:13,831 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stage4.OSA4_1.layers.1.OSA4_1_1/norm.running_var     loaded from backbone.body.stage4.OSA4_1.layers.1.OSA4_1_1/norm.running_var     of shape (192,)
2020-03-22 19:05:13,831 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stage4.OSA4_1.layers.1.OSA4_1_1/norm.weight          loaded from backbone.body.stage4.OSA4_1.layers.1.OSA4_1_1/norm.weight          of shape (192,)
2020-03-22 19:05:13,832 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stage4.OSA4_1.layers.2.OSA4_1_2/conv.weight          loaded from backbone.body.stage4.OSA4_1.layers.2.OSA4_1_2/conv.weight          of shape (192, 192, 3, 3)
2020-03-22 19:05:13,832 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stage4.OSA4_1.layers.2.OSA4_1_2/norm.bias            loaded from backbone.body.stage4.OSA4_1.layers.2.OSA4_1_2/norm.bias            of shape (192,)
2020-03-22 19:05:13,832 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stage4.OSA4_1.layers.2.OSA4_1_2/norm.running_mean    loaded from backbone.body.stage4.OSA4_1.layers.2.OSA4_1_2/norm.running_mean    of shape (192,)
2020-03-22 19:05:13,832 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stage4.OSA4_1.layers.2.OSA4_1_2/norm.running_var     loaded from backbone.body.stage4.OSA4_1.layers.2.OSA4_1_2/norm.running_var     of shape (192,)
2020-03-22 19:05:13,832 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stage4.OSA4_1.layers.2.OSA4_1_2/norm.weight          loaded from backbone.body.stage4.OSA4_1.layers.2.OSA4_1_2/norm.weight          of shape (192,)
2020-03-22 19:05:13,832 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stage5.OSA5_1.concat.OSA5_1_concat/conv.weight       loaded from backbone.body.stage5.OSA5_1.concat.OSA5_1_concat/conv.weight       of shape (1024, 1440, 1, 1)
2020-03-22 19:05:13,833 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stage5.OSA5_1.concat.OSA5_1_concat/norm.bias         loaded from backbone.body.stage5.OSA5_1.concat.OSA5_1_concat/norm.bias         of shape (1024,)
2020-03-22 19:05:13,833 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stage5.OSA5_1.concat.OSA5_1_concat/norm.running_mean loaded from backbone.body.stage5.OSA5_1.concat.OSA5_1_concat/norm.running_mean of shape (1024,)
2020-03-22 19:05:13,833 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stage5.OSA5_1.concat.OSA5_1_concat/norm.running_var  loaded from backbone.body.stage5.OSA5_1.concat.OSA5_1_concat/norm.running_var  of shape (1024,)
2020-03-22 19:05:13,833 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stage5.OSA5_1.concat.OSA5_1_concat/norm.weight       loaded from backbone.body.stage5.OSA5_1.concat.OSA5_1_concat/norm.weight       of shape (1024,)
2020-03-22 19:05:13,833 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stage5.OSA5_1.ese.fc.bias                            loaded from backbone.body.stage5.OSA5_1.ese.fc.bias                            of shape (1024,)
2020-03-22 19:05:13,833 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stage5.OSA5_1.ese.fc.weight                          loaded from backbone.body.stage5.OSA5_1.ese.fc.weight                          of shape (1024, 1024, 1, 1)
2020-03-22 19:05:13,834 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stage5.OSA5_1.layers.0.OSA5_1_0/conv.weight          loaded from backbone.body.stage5.OSA5_1.layers.0.OSA5_1_0/conv.weight          of shape (224, 768, 3, 3)
2020-03-22 19:05:13,834 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stage5.OSA5_1.layers.0.OSA5_1_0/norm.bias            loaded from backbone.body.stage5.OSA5_1.layers.0.OSA5_1_0/norm.bias            of shape (224,)
2020-03-22 19:05:13,834 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stage5.OSA5_1.layers.0.OSA5_1_0/norm.running_mean    loaded from backbone.body.stage5.OSA5_1.layers.0.OSA5_1_0/norm.running_mean    of shape (224,)
2020-03-22 19:05:13,834 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stage5.OSA5_1.layers.0.OSA5_1_0/norm.running_var     loaded from backbone.body.stage5.OSA5_1.layers.0.OSA5_1_0/norm.running_var     of shape (224,)
2020-03-22 19:05:13,834 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stage5.OSA5_1.layers.0.OSA5_1_0/norm.weight          loaded from backbone.body.stage5.OSA5_1.layers.0.OSA5_1_0/norm.weight          of shape (224,)
2020-03-22 19:05:13,834 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stage5.OSA5_1.layers.1.OSA5_1_1/conv.weight          loaded from backbone.body.stage5.OSA5_1.layers.1.OSA5_1_1/conv.weight          of shape (224, 224, 3, 3)
2020-03-22 19:05:13,835 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stage5.OSA5_1.layers.1.OSA5_1_1/norm.bias            loaded from backbone.body.stage5.OSA5_1.layers.1.OSA5_1_1/norm.bias            of shape (224,)
2020-03-22 19:05:13,835 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stage5.OSA5_1.layers.1.OSA5_1_1/norm.running_mean    loaded from backbone.body.stage5.OSA5_1.layers.1.OSA5_1_1/norm.running_mean    of shape (224,)
2020-03-22 19:05:13,835 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stage5.OSA5_1.layers.1.OSA5_1_1/norm.running_var     loaded from backbone.body.stage5.OSA5_1.layers.1.OSA5_1_1/norm.running_var     of shape (224,)
2020-03-22 19:05:13,835 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stage5.OSA5_1.layers.1.OSA5_1_1/norm.weight          loaded from backbone.body.stage5.OSA5_1.layers.1.OSA5_1_1/norm.weight          of shape (224,)
2020-03-22 19:05:13,835 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stage5.OSA5_1.layers.2.OSA5_1_2/conv.weight          loaded from backbone.body.stage5.OSA5_1.layers.2.OSA5_1_2/conv.weight          of shape (224, 224, 3, 3)
2020-03-22 19:05:13,835 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stage5.OSA5_1.layers.2.OSA5_1_2/norm.bias            loaded from backbone.body.stage5.OSA5_1.layers.2.OSA5_1_2/norm.bias            of shape (224,)
2020-03-22 19:05:13,836 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stage5.OSA5_1.layers.2.OSA5_1_2/norm.running_mean    loaded from backbone.body.stage5.OSA5_1.layers.2.OSA5_1_2/norm.running_mean    of shape (224,)
2020-03-22 19:05:13,836 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stage5.OSA5_1.layers.2.OSA5_1_2/norm.running_var     loaded from backbone.body.stage5.OSA5_1.layers.2.OSA5_1_2/norm.running_var     of shape (224,)
2020-03-22 19:05:13,836 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stage5.OSA5_1.layers.2.OSA5_1_2/norm.weight          loaded from backbone.body.stage5.OSA5_1.layers.2.OSA5_1_2/norm.weight          of shape (224,)
2020-03-22 19:05:13,836 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stem.stem_1/conv.weight                              loaded from backbone.body.stem.stem_1/conv.weight                              of shape (64, 3, 3, 3)
2020-03-22 19:05:13,836 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stem.stem_1/norm.bias                                loaded from backbone.body.stem.stem_1/norm.bias                                of shape (64,)
2020-03-22 19:05:13,836 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stem.stem_1/norm.running_mean                        loaded from backbone.body.stem.stem_1/norm.running_mean                        of shape (64,)
2020-03-22 19:05:13,837 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stem.stem_1/norm.running_var                         loaded from backbone.body.stem.stem_1/norm.running_var                         of shape (64,)
2020-03-22 19:05:13,837 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stem.stem_1/norm.weight                              loaded from backbone.body.stem.stem_1/norm.weight                              of shape (64,)
2020-03-22 19:05:13,837 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stem.stem_2/conv.weight                              loaded from backbone.body.stem.stem_2/conv.weight                              of shape (64, 64, 3, 3)
2020-03-22 19:05:13,837 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stem.stem_2/norm.bias                                loaded from backbone.body.stem.stem_2/norm.bias                                of shape (64,)
2020-03-22 19:05:13,837 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stem.stem_2/norm.running_mean                        loaded from backbone.body.stem.stem_2/norm.running_mean                        of shape (64,)
2020-03-22 19:05:13,837 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stem.stem_2/norm.running_var                         loaded from backbone.body.stem.stem_2/norm.running_var                         of shape (64,)
2020-03-22 19:05:13,838 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stem.stem_2/norm.weight                              loaded from backbone.body.stem.stem_2/norm.weight                              of shape (64,)
2020-03-22 19:05:13,838 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stem.stem_3/conv.weight                              loaded from backbone.body.stem.stem_3/conv.weight                              of shape (128, 64, 3, 3)
2020-03-22 19:05:13,838 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stem.stem_3/norm.bias                                loaded from backbone.body.stem.stem_3/norm.bias                                of shape (128,)
2020-03-22 19:05:13,838 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stem.stem_3/norm.running_mean                        loaded from backbone.body.stem.stem_3/norm.running_mean                        of shape (128,)
2020-03-22 19:05:13,838 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stem.stem_3/norm.running_var                         loaded from backbone.body.stem.stem_3/norm.running_var                         of shape (128,)
2020-03-22 19:05:13,838 maskrcnn_benchmark.utils.model_serialization INFO: backbone.body.stem.stem_3/norm.weight                              loaded from backbone.body.stem.stem_3/norm.weight                              of shape (128,)
2020-03-22 19:05:13,839 maskrcnn_benchmark.utils.model_serialization INFO: backbone.fpn.fpn_inner2.bias                                       loaded from backbone.fpn.fpn_inner2.bias                                       of shape (128,)
2020-03-22 19:05:13,839 maskrcnn_benchmark.utils.model_serialization INFO: backbone.fpn.fpn_inner2.weight                                     loaded from backbone.fpn.fpn_inner2.weight                                     of shape (128, 512, 1, 1)
2020-03-22 19:05:13,839 maskrcnn_benchmark.utils.model_serialization INFO: backbone.fpn.fpn_inner3.bias                                       loaded from backbone.fpn.fpn_inner3.bias                                       of shape (128,)
2020-03-22 19:05:13,839 maskrcnn_benchmark.utils.model_serialization INFO: backbone.fpn.fpn_inner3.weight                                     loaded from backbone.fpn.fpn_inner3.weight                                     of shape (128, 768, 1, 1)
2020-03-22 19:05:13,839 maskrcnn_benchmark.utils.model_serialization INFO: backbone.fpn.fpn_inner4.bias                                       loaded from backbone.fpn.fpn_inner4.bias                                       of shape (128,)
2020-03-22 19:05:13,839 maskrcnn_benchmark.utils.model_serialization INFO: backbone.fpn.fpn_inner4.weight                                     loaded from backbone.fpn.fpn_inner4.weight                                     of shape (128, 1024, 1, 1)
2020-03-22 19:05:13,840 maskrcnn_benchmark.utils.model_serialization INFO: backbone.fpn.fpn_layer2.bias                                       loaded from backbone.fpn.fpn_layer2.bias                                       of shape (128,)
2020-03-22 19:05:13,840 maskrcnn_benchmark.utils.model_serialization INFO: backbone.fpn.fpn_layer2.weight                                     loaded from backbone.fpn.fpn_layer2.weight                                     of shape (128, 128, 3, 3)
2020-03-22 19:05:13,840 maskrcnn_benchmark.utils.model_serialization INFO: backbone.fpn.fpn_layer3.bias                                       loaded from backbone.fpn.fpn_layer3.bias                                       of shape (128,)
2020-03-22 19:05:13,840 maskrcnn_benchmark.utils.model_serialization INFO: backbone.fpn.fpn_layer3.weight                                     loaded from backbone.fpn.fpn_layer3.weight                                     of shape (128, 128, 3, 3)
2020-03-22 19:05:13,840 maskrcnn_benchmark.utils.model_serialization INFO: backbone.fpn.fpn_layer4.bias                                       loaded from backbone.fpn.fpn_layer4.bias                                       of shape (128,)
2020-03-22 19:05:13,840 maskrcnn_benchmark.utils.model_serialization INFO: backbone.fpn.fpn_layer4.weight                                     loaded from backbone.fpn.fpn_layer4.weight                                     of shape (128, 128, 3, 3)
2020-03-22 19:05:13,841 maskrcnn_benchmark.utils.model_serialization INFO: backbone.fpn.top_blocks.p6.bias                                    loaded from backbone.fpn.top_blocks.p6.bias                                    of shape (128,)
2020-03-22 19:05:13,841 maskrcnn_benchmark.utils.model_serialization INFO: backbone.fpn.top_blocks.p6.weight                                  loaded from backbone.fpn.top_blocks.p6.weight                                  of shape (128, 128, 3, 3)
2020-03-22 19:05:13,841 maskrcnn_benchmark.utils.model_serialization INFO: backbone.fpn.top_blocks.p7.bias                                    loaded from backbone.fpn.top_blocks.p7.bias                                    of shape (128,)
2020-03-22 19:05:13,841 maskrcnn_benchmark.utils.model_serialization INFO: backbone.fpn.top_blocks.p7.weight                                  loaded from backbone.fpn.top_blocks.p7.weight                                  of shape (128, 128, 3, 3)
2020-03-22 19:05:13,841 maskrcnn_benchmark.utils.model_serialization INFO: roi_heads.mask.feature_extractor.mask_fcn1.bias                    loaded from roi_heads.mask.feature_extractor.mask_fcn1.bias                    of shape (128,)
2020-03-22 19:05:13,841 maskrcnn_benchmark.utils.model_serialization INFO: roi_heads.mask.feature_extractor.mask_fcn1.weight                  loaded from roi_heads.mask.feature_extractor.mask_fcn1.weight                  of shape (128, 128, 3, 3)
2020-03-22 19:05:13,842 maskrcnn_benchmark.utils.model_serialization INFO: roi_heads.mask.feature_extractor.mask_fcn2.bias                    loaded from roi_heads.mask.feature_extractor.mask_fcn2.bias                    of shape (128,)
2020-03-22 19:05:13,842 maskrcnn_benchmark.utils.model_serialization INFO: roi_heads.mask.feature_extractor.mask_fcn2.weight                  loaded from roi_heads.mask.feature_extractor.mask_fcn2.weight                  of shape (128, 128, 3, 3)
2020-03-22 19:05:13,842 maskrcnn_benchmark.utils.model_serialization INFO: roi_heads.mask.feature_extractor.spatialAtt.conv.weight            loaded from roi_heads.mask.feature_extractor.spatialAtt.conv.weight            of shape (1, 2, 3, 3)
2020-03-22 19:05:13,842 maskrcnn_benchmark.utils.model_serialization INFO: roi_heads.mask.predictor.conv5_mask.bias                           loaded from roi_heads.mask.predictor.conv5_mask.bias                           of shape (128,)
2020-03-22 19:05:13,842 maskrcnn_benchmark.utils.model_serialization INFO: roi_heads.mask.predictor.conv5_mask.weight                         loaded from roi_heads.mask.predictor.conv5_mask.weight                         of shape (128, 128, 2, 2)
2020-03-22 19:05:13,842 maskrcnn_benchmark.utils.model_serialization INFO: roi_heads.mask.predictor.mask_fcn_logits.bias                      loaded from roi_heads.mask.predictor.mask_fcn_logits.bias                      of shape (81,)
2020-03-22 19:05:13,843 maskrcnn_benchmark.utils.model_serialization INFO: roi_heads.mask.predictor.mask_fcn_logits.weight                    loaded from roi_heads.mask.predictor.mask_fcn_logits.weight                    of shape (81, 128, 1, 1)
2020-03-22 19:05:13,843 maskrcnn_benchmark.utils.model_serialization INFO: roi_heads.maskiou.feature_extractor.maskiou_fc1.bias               loaded from roi_heads.maskiou.feature_extractor.maskiou_fc1.bias               of shape (1024,)
2020-03-22 19:05:13,843 maskrcnn_benchmark.utils.model_serialization INFO: roi_heads.maskiou.feature_extractor.maskiou_fc1.weight             loaded from roi_heads.maskiou.feature_extractor.maskiou_fc1.weight             of shape (1024, 6272)
2020-03-22 19:05:13,843 maskrcnn_benchmark.utils.model_serialization INFO: roi_heads.maskiou.feature_extractor.maskiou_fc2.bias               loaded from roi_heads.maskiou.feature_extractor.maskiou_fc2.bias               of shape (1024,)
2020-03-22 19:05:13,843 maskrcnn_benchmark.utils.model_serialization INFO: roi_heads.maskiou.feature_extractor.maskiou_fc2.weight             loaded from roi_heads.maskiou.feature_extractor.maskiou_fc2.weight             of shape (1024, 1024)
2020-03-22 19:05:13,843 maskrcnn_benchmark.utils.model_serialization INFO: roi_heads.maskiou.feature_extractor.maskiou_fcn1.bias              loaded from roi_heads.maskiou.feature_extractor.maskiou_fcn1.bias              of shape (128,)
2020-03-22 19:05:13,843 maskrcnn_benchmark.utils.model_serialization INFO: roi_heads.maskiou.feature_extractor.maskiou_fcn1.weight            loaded from roi_heads.maskiou.feature_extractor.maskiou_fcn1.weight            of shape (128, 129, 3, 3)
2020-03-22 19:05:13,844 maskrcnn_benchmark.utils.model_serialization INFO: roi_heads.maskiou.feature_extractor.maskiou_fcn2.bias              loaded from roi_heads.maskiou.feature_extractor.maskiou_fcn2.bias              of shape (128,)
2020-03-22 19:05:13,844 maskrcnn_benchmark.utils.model_serialization INFO: roi_heads.maskiou.feature_extractor.maskiou_fcn2.weight            loaded from roi_heads.maskiou.feature_extractor.maskiou_fcn2.weight            of shape (128, 128, 3, 3)
2020-03-22 19:05:13,844 maskrcnn_benchmark.utils.model_serialization INFO: roi_heads.maskiou.predictor.maskiou.bias                           loaded from roi_heads.maskiou.predictor.maskiou.bias                           of shape (81,)
2020-03-22 19:05:13,844 maskrcnn_benchmark.utils.model_serialization INFO: roi_heads.maskiou.predictor.maskiou.weight                         loaded from roi_heads.maskiou.predictor.maskiou.weight                         of shape (81, 1024)
2020-03-22 19:05:13,844 maskrcnn_benchmark.utils.model_serialization INFO: rpn.head.bbox_pred.bias                                            loaded from rpn.head.bbox_pred.bias                                            of shape (4,)
2020-03-22 19:05:13,844 maskrcnn_benchmark.utils.model_serialization INFO: rpn.head.bbox_pred.weight                                          loaded from rpn.head.bbox_pred.weight                                          of shape (4, 128, 3, 3)
2020-03-22 19:05:13,845 maskrcnn_benchmark.utils.model_serialization INFO: rpn.head.bbox_tower.0.bias                                         loaded from rpn.head.bbox_tower.0.bias                                         of shape (128,)
2020-03-22 19:05:13,845 maskrcnn_benchmark.utils.model_serialization INFO: rpn.head.bbox_tower.0.weight                                       loaded from rpn.head.bbox_tower.0.weight                                       of shape (128, 128, 3, 3)
2020-03-22 19:05:13,845 maskrcnn_benchmark.utils.model_serialization INFO: rpn.head.bbox_tower.1.bias                                         loaded from rpn.head.bbox_tower.1.bias                                         of shape (128,)
2020-03-22 19:05:13,845 maskrcnn_benchmark.utils.model_serialization INFO: rpn.head.bbox_tower.1.weight                                       loaded from rpn.head.bbox_tower.1.weight                                       of shape (128,)
2020-03-22 19:05:13,845 maskrcnn_benchmark.utils.model_serialization INFO: rpn.head.bbox_tower.3.bias                                         loaded from rpn.head.bbox_tower.3.bias                                         of shape (128,)
2020-03-22 19:05:13,845 maskrcnn_benchmark.utils.model_serialization INFO: rpn.head.bbox_tower.3.weight                                       loaded from rpn.head.bbox_tower.3.weight                                       of shape (128, 128, 3, 3)
2020-03-22 19:05:13,846 maskrcnn_benchmark.utils.model_serialization INFO: rpn.head.bbox_tower.4.bias                                         loaded from rpn.head.bbox_tower.4.bias                                         of shape (128,)
2020-03-22 19:05:13,846 maskrcnn_benchmark.utils.model_serialization INFO: rpn.head.bbox_tower.4.weight                                       loaded from rpn.head.bbox_tower.4.weight                                       of shape (128,)
2020-03-22 19:05:13,846 maskrcnn_benchmark.utils.model_serialization INFO: rpn.head.centerness.bias                                           loaded from rpn.head.centerness.bias                                           of shape (1,)
2020-03-22 19:05:13,846 maskrcnn_benchmark.utils.model_serialization INFO: rpn.head.centerness.weight                                         loaded from rpn.head.centerness.weight                                         of shape (1, 128, 3, 3)
2020-03-22 19:05:13,846 maskrcnn_benchmark.utils.model_serialization INFO: rpn.head.cls_logits.bias                                           loaded from rpn.head.cls_logits.bias                                           of shape (80,)
2020-03-22 19:05:13,846 maskrcnn_benchmark.utils.model_serialization INFO: rpn.head.cls_logits.weight                                         loaded from rpn.head.cls_logits.weight                                         of shape (80, 128, 3, 3)
2020-03-22 19:05:13,847 maskrcnn_benchmark.utils.model_serialization INFO: rpn.head.cls_tower.0.bias                                          loaded from rpn.head.cls_tower.0.bias                                          of shape (128,)
2020-03-22 19:05:13,847 maskrcnn_benchmark.utils.model_serialization INFO: rpn.head.cls_tower.0.weight                                        loaded from rpn.head.cls_tower.0.weight                                        of shape (128, 128, 3, 3)
2020-03-22 19:05:13,847 maskrcnn_benchmark.utils.model_serialization INFO: rpn.head.cls_tower.1.bias                                          loaded from rpn.head.cls_tower.1.bias                                          of shape (128,)
2020-03-22 19:05:13,847 maskrcnn_benchmark.utils.model_serialization INFO: rpn.head.cls_tower.1.weight                                        loaded from rpn.head.cls_tower.1.weight                                        of shape (128,)
2020-03-22 19:05:13,847 maskrcnn_benchmark.utils.model_serialization INFO: rpn.head.cls_tower.3.bias                                          loaded from rpn.head.cls_tower.3.bias                                          of shape (128,)
2020-03-22 19:05:13,847 maskrcnn_benchmark.utils.model_serialization INFO: rpn.head.cls_tower.3.weight                                        loaded from rpn.head.cls_tower.3.weight                                        of shape (128, 128, 3, 3)
2020-03-22 19:05:13,848 maskrcnn_benchmark.utils.model_serialization INFO: rpn.head.cls_tower.4.bias                                          loaded from rpn.head.cls_tower.4.bias                                          of shape (128,)
2020-03-22 19:05:13,848 maskrcnn_benchmark.utils.model_serialization INFO: rpn.head.cls_tower.4.weight                                        loaded from rpn.head.cls_tower.4.weight                                        of shape (128,)
2020-03-22 19:05:13,848 maskrcnn_benchmark.utils.model_serialization INFO: rpn.head.scales.0.scale                                            loaded from rpn.head.scales.0.scale                                            of shape (1,)
2020-03-22 19:05:13,848 maskrcnn_benchmark.utils.model_serialization INFO: rpn.head.scales.1.scale                                            loaded from rpn.head.scales.1.scale                                            of shape (1,)
2020-03-22 19:05:13,848 maskrcnn_benchmark.utils.model_serialization INFO: rpn.head.scales.2.scale                                            loaded from rpn.head.scales.2.scale                                            of shape (1,)
2020-03-22 19:05:13,848 maskrcnn_benchmark.utils.model_serialization INFO: rpn.head.scales.3.scale                                            loaded from rpn.head.scales.3.scale                                            of shape (1,)
2020-03-22 19:05:13,849 maskrcnn_benchmark.utils.model_serialization INFO: rpn.head.scales.4.scale                                            loaded from rpn.head.scales.4.scale                                            of shape (1,)
loading annotations into memory...
Done (t=2.02s)
creating index...
index created!
2020-03-22 19:05:17,232 maskrcnn_benchmark.inference INFO: Start evaluation on coco_2014_minival dataset(5000 images).
  0%|                                                                                                                       | 1/5000 [00:08<11:50:39,  8.53s/it]Traceback (most recent call last):
  File "tools/test_net.py", line 97, in <module>
    main()
  File "tools/test_net.py", line 91, in main
    output_folder=output_folder,
  File "/home/panda/install/CenterMask/maskrcnn_benchmark/engine/inference.py", line 79, in inference
    predictions = compute_on_dataset(model, data_loader, device, inference_timer)
  File "/home/panda/install/CenterMask/maskrcnn_benchmark/engine/inference.py", line 20, in compute_on_dataset
    for _, batch in enumerate(tqdm(data_loader)):
  File "/home/panda/.local/lib/python3.6/site-packages/tqdm/std.py", line 1108, in __iter__
    for obj in iterable:
  File "/home/panda/.local/lib/python3.6/site-packages/torch/utils/data/dataloader.py", line 345, in __next__
    data = self._next_data()
  File "/home/panda/.local/lib/python3.6/site-packages/torch/utils/data/dataloader.py", line 838, in _next_data
    return self._process_data(data)
  File "/home/panda/.local/lib/python3.6/site-packages/torch/utils/data/dataloader.py", line 881, in _process_data
    data.reraise()
  File "/home/panda/.local/lib/python3.6/site-packages/torch/_utils.py", line 394, in reraise
    raise self.exc_type(msg)
IndexError: Caught IndexError in DataLoader worker process 1.
Original Traceback (most recent call last):
  File "/home/panda/.local/lib/python3.6/site-packages/torch/utils/data/_utils/worker.py", line 178, in _worker_loop
    data = fetcher.fetch(index)
  File "/home/panda/.local/lib/python3.6/site-packages/torch/utils/data/_utils/fetch.py", line 44, in fetch
    data = [self.dataset[idx] for idx in possibly_batched_index]
  File "/home/panda/.local/lib/python3.6/site-packages/torch/utils/data/_utils/fetch.py", line 44, in <listcomp>
    data = [self.dataset[idx] for idx in possibly_batched_index]
  File "/home/panda/install/CenterMask/maskrcnn_benchmark/data/datasets/coco.py", line 91, in __getitem__
    target = target.clip_to_image(remove_empty=True)
  File "/home/panda/install/CenterMask/maskrcnn_benchmark/structures/bounding_box.py", line 224, in clip_to_image
    return self[keep]
  File "/home/panda/install/CenterMask/maskrcnn_benchmark/structures/bounding_box.py", line 209, in __getitem__
    bbox.add_field(k, v[item])
  File "/home/panda/install/CenterMask/maskrcnn_benchmark/structures/segmentation_mask.py", line 513, in __getitem__
    selected_instances = self.instances.__getitem__(item)
  File "/home/panda/install/CenterMask/maskrcnn_benchmark/structures/segmentation_mask.py", line 422, in __getitem__
    selected_polygons.append(self.polygons[i])
IndexError: list index out of range

Qusetions about result in papers

In the Ablation study section of your paper, what is the experiment configuration of Feature selection?
I find the best result is 34.6 AP in Table 2, but Table 1 shows the best result is 34.7. Is there any difference between them？

loss nan

When I try to train on COCO 2014 dataset, I get loss nan, but not always, sometimes it's going ok.

cmd line:

python tools/train_net.py --config-file "configs/centermask/centermask_V_39_eSE_FPN_lite_res600_ms_bs16_4x.yaml" SOLVER.IMS_PER_BATCH 4 SOLVER.TEST_PERIOD 1000 DATASETS.TRAIN "('coco_2014_train',)" DATASETS.TEST "('coco_2014_minival',)"

env:

OS: Microsoft Windows 10 Pro
Python version: 3.7
Is CUDA available: Yes
CUDA runtime version: 10.0.130
[pip3] numpy==1.18.1
[pip3] torch==1.1.0
[pip3] torchvision==0.3.0
        Pillow (6.1.0)

Bad results from beginning:

2020-02-09 23:27:24,918 maskrcnn_benchmark.trainer INFO: eta: 7 days, 22:00:10  iter: 20  loss: nan (nan)  loss_mask: nan (nan)  loss_maskiou: nan (nan)  loss_cls: 21.4419 (18.3350)  loss_reg: nan (nan)  loss_centerness: nan (nan)  time: 0.3928 (1.9001)  data: 0.0029 (1.4490)  lr: 0.003333  max mem: 4277

2020-02-09 23:37:35,529 maskrcnn_benchmark.trainer INFO: eta: 8 days, 1:01:12  iter: 20  loss: 23542148363337776622290525596155904.0000 (nan)  loss_mask: 0.6979 (424800877876977736876032.0000)  loss_maskiou: 23542148363337776622290525596155904.0000 (nan)  loss_cls: 1.1945 (11.2513)  loss_reg: 0.9993 (nan)  loss_centerness: 0.7402 (nan)  time: 0.4166 (1.9303)  data: 0.0025 (1.4344)  lr: 0.003333  max mem: 4481

2020-02-09 23:38:44,647 maskrcnn_benchmark.trainer INFO: eta: 7 days, 21:55:19  iter: 20  loss: nan (nan)  loss_mask: 1.5184 (6998281.7163)  loss_maskiou: nan (nan)  loss_cls: 21.3143 (16.3219)  loss_reg: nan (nan)  loss_centerness: nan (nan)  time: 0.3993 (1.8993)  data: 0.0030 (1.4402)  lr: 0.003333  max mem: 4468

2020-02-09 23:42:07,362 maskrcnn_benchmark.trainer INFO: eta: 8 days, 17:18:19  iter: 20  loss: 3.6971 (nan)  loss_mask: 0.6961 (15.4345)  loss_maskiou: 0.2314 (nan)  loss_cls: 1.0910 (6.1945)  loss_reg: 0.9994 (nan)  loss_centerness: 0.6840 (nan)  time: 0.6344 (2.0932)  data: 0.0040 (1.4359)  lr: 0.003333  max mem: 4492
2020-02-09 23:42:17,270 maskrcnn_benchmark.trainer INFO: eta: 5 days, 9:24:52  iter: 40  loss: nan (nan)  loss_mask: 1.8518 (8.6920)  loss_maskiou: nan (nan)  loss_cls: 21.4820 (13.8231)  loss_reg: nan (nan)  loss_centerness: nan (nan)  time: 0.4881 (1.2943)  data: 0.0045 (0.7206)  lr: 0.003333  max mem: 4492

2020-02-09 23:52:21,182 maskrcnn_benchmark.trainer INFO: eta: 8 days, 7:18:50  iter: 20  loss: nan (nan)  loss_mask: nan (nan)  loss_maskiou: nan (nan)  loss_cls: 21.3143 (16.3264)  loss_reg: nan (nan)  loss_centerness: nan (nan)  time: 0.5104 (1.9933)  data: 0.0035 (1.4069)  lr: 0.003333  max mem: 4513
2020-02-09 23:52:31,129 maskrcnn_benchmark.trainer INFO: eta: 5 days, 4:30:52  iter: 40  loss: nan (nan)  loss_mask: nan (nan)  loss_maskiou: nan (nan)  loss_cls: 21.3671 (18.8594)  loss_reg: nan (nan)  loss_centerness: nan (nan)  time: 0.4871 (1.2453)  data: 0.0040 (0.7057)  lr: 0.003333  max mem: 4513

2020-02-09 23:53:46,776 maskrcnn_benchmark.trainer INFO: eta: 8 days, 5:16:40  iter: 20  loss: nan (nan)  loss_mask: 1675440662589417990389760.0000 (2184793125528002094956544.0000)  loss_maskiou: nan (nan)  loss_cls: 21.3885 (17.3715)  loss_reg: nan (nan)  loss_centerness: nan (nan)  time: 0.4871 (1.9729)  data: 0.0039 (1.4136)  lr: 0.003333  max mem: 4217

Ok result:

2020-02-09 23:44:14,344 maskrcnn_benchmark.trainer INFO: eta: 8 days, 17:40:27  iter: 20  loss: 3.5707 (3.6511)  loss_mask: 0.7062 (0.7645)  loss_maskiou: 0.0637 (0.1216)  loss_cls: 1.0763 (1.0718)  loss_reg: 0.9982 (0.9978)  loss_centerness: 0.6932 (0.6954)  time: 0.6289 (2.0969)  data: 0.0035 (1.4190)  lr: 0.003333  max mem: 4548
2020-02-09 23:44:27,658 maskrcnn_benchmark.trainer INFO: eta: 5 days, 18:06:41  iter: 40  loss: 3.3980 (3.5286)  loss_mask: 0.6923 (0.7301)  loss_maskiou: 0.0382 (0.0867)  loss_cls: 0.9761 (1.0231)  loss_reg: 0.9968 (0.9970)  loss_centerness: 0.6855 (0.6916)  time: 0.6582 (1.3813)  data: 0.0040 (0.7119)  lr: 0.003333  max mem: 4578
2020-02-09 23:44:40,926 maskrcnn_benchmark.trainer INFO: eta: 4 days, 18:10:50  iter: 60  loss: 3.3664 (3.4819)  loss_mask: 0.6919 (0.7175)  loss_maskiou: 0.0355 (0.0710)  loss_cls: 0.9938 (1.0140)  loss_reg: 0.9918 (0.9940)  loss_centerness: 0.6751 (0.6854)  time: 0.6557 (1.1420)  data: 0.0040 (0.4762)  lr: 0.003333  max mem: 4728
2020-02-09 23:44:54,655 maskrcnn_benchmark.trainer INFO: eta: 4 days, 6:47:16  iter: 80  loss: 3.2235 (3.4085)  loss_mask: 0.6806 (0.7088)  loss_maskiou: 0.0246 (0.0602)  loss_cls: 0.9886 (1.0100)  loss_reg: 0.8146 (0.9449)  loss_centerness: 0.6809 (0.6846)  time: 0.6860 (1.0281)  data: 0.0049 (0.3585)  lr: 0.003333  max mem: 4728
2020-02-09 23:45:08,560 maskrcnn_benchmark.trainer INFO: eta: 4 days, 0:07:35  iter: 100  loss: 3.0076 (3.3381)  loss_mask: 0.6819 (0.7048)  loss_maskiou: 0.0167 (0.0518)  loss_cls: 0.9510 (1.0000)  loss_reg: 0.6715 (0.8965)  loss_centerness: 0.6853 (0.6849)  time: 0.6875 (0.9615)  data: 0.0045 (0.2879)  lr: 0.003333  max mem: 4728

install error

The time cost is far longer than '20ms' with gpu.

Hi，Thanks for your open source code.
I run the centermask_demo.py with the model "centermask-lite-M-v2-ms-bs32-1x.pth",and config file "centermask_M_v2_FPN_lite_res600_ms_bs32_1x.yaml",
the time cost is 60ms-110ms since the second image :

file 0
COCO_val2014_000000128654 processing...
COCO_val2014_000000128654 inference time: 0.281s
file 1
COCO_val2014_000000007281 processing...
COCO_val2014_000000007281 inference time: 0.072s
file 2
COCO_val2014_000000005477 processing...
COCO_val2014_000000005477 inference time: 0.068s
file 3
COCO_val2014_000000463842 processing...
COCO_val2014_000000463842 inference time: 0.111s
file 4
COCO_val2014_000000053505 processing...
COCO_val2014_000000053505 inference time: 0.075s
file 5
COCO_val2014_000000000885 processing...
COCO_val2014_000000000885 inference time: 0.068s
file 6
COCO_val2014_000000033759 processing...
COCO_val2014_000000033759 inference time: 0.067s
file 7
COCO_val2014_000000050896 processing...
COCO_val2014_000000050896 inference time: 0.119s
file 8
COCO_val2014_000000012639 processing...
COCO_val2014_000000012639 inference time: 0.075s
file 9
COCO_val2014_000000479030 processing...
COCO_val2014_000000479030 inference time: 0.070s
file 10
COCO_val2014_000000001000 processing...
COCO_val2014_000000001000 inference time: 0.076s
file 11
COCO_val2014_000000039769 processing...
COCO_val2014_000000039769 inference time: 0.120s

The test image size set :
INPUT:
MIN_SIZE_RANGE_TRAIN: (580, 600)
MAX_SIZE_TRAIN: 1000
MIN_SIZE_TEST: 600
MAX_SIZE_TEST: 1000

Even I set the MAX_SIZE_TEST to 500, MIN_SIZE_TEST to 300,the time cost is still far longer than “20”ms as you mentioned.

My pytorch version : 1.3.1
CUDA:10.0.1
GPU: P40

How can I improve time efficiency，please？

Where could i download the pretrained models?

Links left here are invalid. Looking forward to your reply.

paper questions

@youngwanLEE Hi Youngwan, Jongyoul, thank for your greate work of CenterMask. However, I have a few questions which confused me a lot, Hope you can comment on it :)

Firstly,  from my understanding, you guys use the result of FCOS's bounding boxes as the input of ROIAlign, Are you using all k=100 predicted results or "highest-scoring boxes" to feed intothe SAG-mask branch for training mask branch?  like how do you define highest-scoring boxes, maybe a hyper-parameter to use top n results, or choose a confidence threshold that greater than m?

Secondly, In the inference stage, I noticed that adding mask head boosts AP from 37.8% to 38.3%, Do you use the original FCOS output to evaluate it or re-predict everything according to the mask-head?  I'm curious is it because the mask-head that refine the original FOCS bbox to make it better or because the mask head helps the LOSS to coverage better than before(if using FCOS to predict bbox).  
I'm only talking about AP_bbox not AP_mask, thank you very much!

paper about VoVNetV2

hi @youngwanLEE
what is the difference between the VoVNetV and VoVNetV2? and have the paper about VoVNetV2?
thanks

Its really difficult to train out a good mAP when training on my own dataset

Problem Summary

Firstly, I trained on standard COCO dataset used in this paper, and got a good mAP. Secondly, I prepared my own dataset according to COCO format and named them as "train2014"、 "val2014"、 "instances_train2014.json"、"instances_val2014.json". Thirdly, since my own dataset contains only one category --'building', so I changed '_C.MODEL.ROI_BOX_HEAD.NUM_CLASSES'、'_C.MODEL. FCOS.NUM_CLASSES'、'_C.MODEL.RETINANET.NUM_CLASSES' in defaults.py from 81 to 2. And then I trained on my own dataset. But got a bad mAP. It is worth mentioning that I have already visualized my own dataset, and my own dataset performed good in maskrcnn. Thus, I want to ask you if the centermask can be used for other datasets, or if I need to modify any other information when training with my own dataset. [I noticed that there is an issue in FCOS which is similar with this problem: https://github.com/tianzhi0549/FCOS/issues/132, but the issue is also not resolved.]

Environment

GPU: 4 titan xp (12GB)
Versions of relevant libraries:
[pip] numpy==1.16.0
[pip] torch==1.0.0.dev20190328
[pip] torchvision==0.2.2
[conda] pytorch-nightly 1.0.0.dev20190328 py3.7_cuda9.0.176_cudnn7.4.2_0
[conda] torchvision 0.2.2 pypi_0 pypi
Pillow (6.2.1)

configs

loss

AP and AR

Average Precision (AP) @[ IoU=0.50:0.95 | area= all | maxDets=100 ] = 0.001
Average Precision (AP) @[ IoU=0.50 | area= all | maxDets=100 ] = 0.004
Average Precision (AP) @[ IoU=0.75 | area= all | maxDets=100 ] = 0.000
Average Precision (AP) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.002
Average Precision (AP) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.001
Average Precision (AP) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.008
Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets= 1 ] = 0.003
Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets= 10 ] = 0.013
Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets=100 ] = 0.018
Average Recall (AR) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.013
Average Recall (AR) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.019
Average Recall (AR) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.091
Maximum f-measures for classes:
[0.04064810445178543]
Score thresholds for classes (used in demos for visualization purposes):
[0.016668733209371567]
Loading and preparing results...
DONE (t=0.19s)
creating index...
index created!
Running per image evaluation...
Evaluate annotation type segm
DONE (t=13.13s).
Accumulating evaluation results...
DONE (t=0.23s).
Average Precision (AP) @[ IoU=0.50:0.95 | area= all | maxDets=100 ] = 0.003
Average Precision (AP) @[ IoU=0.50 | area= all | maxDets=100 ] = 0.005
Average Precision (AP) @[ IoU=0.75 | area= all | maxDets=100 ] = 0.003
Average Precision (AP) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.000
Average Precision (AP) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.002
Average Precision (AP) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.013
Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets= 1 ] = 0.003
Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets= 10 ] = 0.011
Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets=100 ] = 0.013
Average Recall (AR) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.011
Average Recall (AR) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.011
Average Recall (AR) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.086
Maximum f-measures for classes:
[0.02170795306388527]
Score thresholds for classes (used in demos for visualization purposes):
[0.33201679587364197]
2020-01-17 04:11:24,843 maskrcnn_benchmark.inference INFO: OrderedDict([('bbox', OrderedDict([('AP', 0.0010722595928040015), ('AP50', 0.003614397825546365), ('AP75', 0.0003211566688604318), ('APs', 0.0016450320259484557), ('APm', 0.0011716715215730757), ('APl', 0.007575648306486348)])), ('segm', OrderedDict([('AP', 0.0027830979629441433), ('AP50', 0.005436005785051814), ('AP75', 0.002610742186121677), ('APs', 0.00037027883278494873), ('APm', 0.002340194956846525), ('APl', 0.01333755135821637)]))])

Training problem when train centermask_V_39_eSE_FPN_ms_3x.yaml

Hi, when I train centermask_V_39_eSE_FPN_ms_3x.yaml/2x.yaml the losses is None, shown as the following:

2020-01-13 06:49:06,569 maskrcnn_benchmark.trainer INFO: eta: 1 day, 21:25:55 iter: 2740 loss: nan (nan) loss_mask: 0.3441 (0.4154) loss_maskiou: nan (nan) loss_cls: 21.1872 (21.2148) loss_reg: nan (nan) loss_centerness: nan (nan) time: 0.6184 (0.6120) data: 0.0172 (0.0205) lr: 0.010000 max mem: 12121
2020-01-13 06:49:18,936 maskrcnn_benchmark.trainer INFO: eta: 1 day, 21:25:55 iter: 2760 loss: nan (nan) loss_mask: 0.3562 (0.4151) loss_maskiou: nan (nan) loss_cls: 21.2557 (21.2151) loss_reg: nan (nan) loss_centerness: nan (nan) time: 0.6148 (0.6120) data: 0.0181 (0.0205) lr: 0.010000 max mem: 12121

How can I add a LandmarkHead to the net , could you give some advice , please ?

Thanks for your great work, I can train the net with the codes now.
How can I add a LandmarkHead to the net , could you give some advice , please ?

Where did instance norm added?

HI, I wanna deploy centermask on tensorrt, I have converted model to onnx.

but when convert onnx2trt I got a error:

 In function importInstanceNormalization:
[8] Assertion failed: !isDynamic(tensor_ptr->getDimensions()) && "InstanceNormalization does not support dynamic inputs!"

You may not familiar with tensorrt, but can you tell me where did you add instancenorm in code? I searched a round and I can not locate where is it.

BTW, does instance norm dim is dynamic or not?

Can you share the pretrained model to Google drive?

The training dataset is coco2014?

In your papers, "All models are trained on the train2017 and val2017 are used for ablation studies." But in your code, configs, I found that " TRAIN: ("coco_2014_train", "coco_2014_valminusminival")
TEST: ("coco_2014_minival",)", Does this mean you use coco2014dataset to train and test?

Runtime Error: cuDNN error: CUDNN_STATUS_EXECUTION_FAILED

Hi everyone,
I have a problem to rum the train_net script.
I followed the installations and everything worked fine. I want to train the model with coco dataset But I always get the runtime error. Did anyone has the same problem and can help me out?
Thanks a lot

Bad key "text.kerning_factor" on line 4 in
C:\Users\skugler\AppData\Local\Continuum\anaconda3\envs\env_centermask\lib\site-packages\matplotlib\mpl-data\stylelib_classic_test_patch.mplstyle.
You probably need to get an updated matplotlibrc file from
https://github.com/matplotlib/matplotlib/blob/v3.1.3/matplotlibrc.template
or from the matplotlib source distribution
Traceback (most recent call last):
File "tools/train_net.py", line 196, in
main()
File "tools/train_net.py", line 184, in main
model = train(cfg, args.local_rank, args.distributed)
File "tools/train_net.py", line 88, in train
arguments,
File "c:\users\skugler\desktop\centermask\maskrcnn_benchmark\engine\trainer.py", line 83, in do_train
loss_dict = model(images, targets)
File "C:\Users\skugler\AppData\Local\Continuum\anaconda3\envs\env_centermask\lib\site-packages\torch\nn\modules\module.py", line 493, in call
result = self.forward(*input, **kwargs)
File "c:\users\skugler\desktop\centermask\maskrcnn_benchmark\modeling\detector\generalized_rcnn.py", line 49, in forward
features = self.backbone(images.tensors)
File "C:\Users\skugler\AppData\Local\Continuum\anaconda3\envs\env_centermask\lib\site-packages\torch\nn\modules\module.py", line 493, in call
result = self.forward(*input, **kwargs)
File "C:\Users\skugler\AppData\Local\Continuum\anaconda3\envs\env_centermask\lib\site-packages\torch\nn\modules\container.py", line 92, in forward
input = module(input)
File "C:\Users\skugler\AppData\Local\Continuum\anaconda3\envs\env_centermask\lib\site-packages\torch\nn\modules\module.py", line 493, in call
result = self.forward(*input, **kwargs)
File "c:\users\skugler\desktop\centermask\maskrcnn_benchmark\modeling\backbone\mobilenet.py", line 115, in forward
x = m(x)
File "C:\Users\skugler\AppData\Local\Continuum\anaconda3\envs\env_centermask\lib\site-packages\torch\nn\modules\module.py", line 493, in call
result = self.forward(*input, **kwargs)
File "C:\Users\skugler\AppData\Local\Continuum\anaconda3\envs\env_centermask\lib\site-packages\torch\nn\modules\container.py", line 92, in forward
input = module(input)
File "C:\Users\skugler\AppData\Local\Continuum\anaconda3\envs\env_centermask\lib\site-packages\torch\nn\modules\module.py", line 493, in call
result = self.forward(*input, **kwargs)
File "c:\users\skugler\desktop\centermask\maskrcnn_benchmark\layers\misc.py", line 33, in forward
return super(Conv2d, self).forward(x)
File "C:\Users\skugler\AppData\Local\Continuum\anaconda3\envs\env_centermask\lib\site-packages\torch\nn\modules\conv.py", line 338, in forward
self.padding, self.dilation, self.groups)
RuntimeError: cuDNN error: CUDNN_STATUS_EXECUTION_FAILED

Finetuning with existing model

Hi @youngwanLEE ,

Would you please help me how can I do finetuning on the existed models to get the mask of only one class (person) of coco dataset? I need my model to detect only person, not 80 objects and I don't want to train the model from scratch. I have tried to follow maskrcnn-benchmark to do the finetunning but when I tried to trim the model to get rid of the last layers, I'm getting this error

detectron path: /home/XX/CenterMask/centermask-lite-M-v2-bs16-4x.pth Traceback (most recent call last): File "trimCenterMaskModel.py", line 43, in <module> _d = load_c2_format(cfg, DETECTRON_PATH) File "/home/media4us/PycharmProjects/CenterMask/maskrcnn_benchmark/utils/c2_model_loading.py", line 176, in load_c2_format return C2_FORMAT_LOADER[cfg.MODEL.BACKBONE.CONV_BODY](cfg, f) KeyError: 'MNV2-FPN-RETINANET'

My command line:

python trimCenterMaskModel.py --pretrained_path /home/media4us/PycharmProjects/CenterMask/centermask-lite-M-v2-bs16-4x.pth --save_path /home/media4us/PycharmProjects/CenterMask/centermask_Test --cfg configs/centermask/centermask_M_v2_FPN_lite_res600_ms_bs16_4x.yaml

it is error to download the weight file

error link ？ https://dl.dropbox.com/s/2enqxenccz4xy6l/centermask-lite-R-50-ms-bs32-1x.pth

visualizing loss curve?

hi, I recently started using CenterMask on my own dataset and was wondering what would be the best method to visualize the loss curves? do you take the info from the console text once the code is finished running? or is there a better method?
thank you.

How to choose positive samples from fcos?

During training, FCOS generates 100 detection bboxes. Are these all considered positive samples? Or do we need to calculate the iou to determine whether these bboxes are positive samples in the mask branch training like mask rcnn?

There is an Install error. Is there any version that does not need to be installed

Trim model script for center mask

If anyone has the trim model script for transfer learning CenterMask, can you upload here? I have used maskrcnn script, but I face some problems??

error: command 'D:\\VisualStudio\\VC\\BIN\\x86_amd64\\cl.exe' failed with exit status 2

when i run:
cd CenterMask
python setup.py build develop
I have a problem:
D:\Anaconda3\envs\centermask\lib\site-packages\torch\include\ATen/core/ivalue_inl.h(624): note: see reference to class template instantiation 'c10::ArrayRefc10::IValue' being compiled
error: command 'D:\VisualStudio\VC\BIN\x86_amd64\cl.exe' failed with exit status 2

both my evironment and virtual evironment according your readme.md all have the problom.
my pytorch is 1.4.0
thanks very much your answer

IndexError

I'm trying to train with default config (coco dataset):

python -m torch.distributed.launch --nproc_per_node=1 tools/train_net.py --config-file "configs/centermask/centermask_V_19_eSE_FPN_lite_res600_ms_bs16_4x.yaml"

and I'm getting error right after start training:

2020-01-31 10:25:50,812 maskrcnn_benchmark.trainer INFO: Start training
Traceback (most recent call last):
  File "tools/train_net.py", line 189, in <module>
    main()
  File "tools/train_net.py", line 182, in main
    model = train(cfg, args.local_rank, args.distributed)
  File "tools/train_net.py", line 88, in train
    arguments,
  File "e:\files\tmp\centermask\maskrcnn_benchmark\engine\trainer.py", line 71, in do_train
    for iteration, (images, targets, _) in enumerate(data_loader, start_iter):
  File "d:\env\ce\lib\site-packages\torch\utils\data\dataloader.py", line 345, in __next__
    data = self._next_data()
  File "d:\env\ce\lib\site-packages\torch\utils\data\dataloader.py", line 856, in _next_data
    return self._process_data(data)
  File "d:\env\ce\lib\site-packages\torch\utils\data\dataloader.py", line 881, in _process_data
    data.reraise()
  File "d:\env\ce\lib\site-packages\torch\_utils.py", line 394, in reraise
    raise self.exc_type(msg)
IndexError: Caught IndexError in DataLoader worker process 0.
Original Traceback (most recent call last):
  File "d:\env\ce\lib\site-packages\torch\utils\data\_utils\worker.py", line 178, in _worker_loop
    data = fetcher.fetch(index)
  File "d:\env\ce\lib\site-packages\torch\utils\data\_utils\fetch.py", line 44, in fetch
    data = [self.dataset[idx] for idx in possibly_batched_index]
  File "d:\env\ce\lib\site-packages\torch\utils\data\_utils\fetch.py", line 44, in <listcomp>
    data = [self.dataset[idx] for idx in possibly_batched_index]
  File "d:\env\ce\lib\site-packages\torch\utils\data\dataset.py", line 207, in __getitem__
    return self.datasets[dataset_idx][sample_idx]
  File "e:\files\tmp\centermask\maskrcnn_benchmark\data\datasets\coco.py", line 91, in __getitem__
    target = target.clip_to_image(remove_empty=True)
  File "e:\files\tmp\centermask\maskrcnn_benchmark\structures\bounding_box.py", line 224, in clip_to_image
    return self[keep]
  File "e:\files\tmp\centermask\maskrcnn_benchmark\structures\bounding_box.py", line 209, in __getitem__
    bbox.add_field(k, v[item])
  File "e:\files\tmp\centermask\maskrcnn_benchmark\structures\segmentation_mask.py", line 513, in __getitem__
    selected_instances = self.instances.__getitem__(item)
  File "e:\files\tmp\centermask\maskrcnn_benchmark\structures\segmentation_mask.py", line 422, in __getitem__
    selected_polygons.append(self.polygons[i])
IndexError: list index out of range

when will you open your implementation based on Detectron2?

thank you for your considerable work. by the way, when will you open your source code based on detectron2?

onnx exported?

Any work around on export CenterMask to onnx?-

Cuda out of memory

I have one 2080 titan.I have set the IMG_PER_BATCH=2 and change image size to 512.Why I also met cuda out of memory?

have any one try model.half() and input.half() to speed up the inference?

i try model.half() and input.half() for speeding up the inference, the time decrease from 65ms to 53ms, although, it does not speed up too much, but i find that the model's output is empty, which means it can't detect the objs in the image. it's so weird. any help is appreciated.

Output with shape doesn't match the broadcasting shape

I am performing transfer learning with centermask for my custom dataset. Using the trim_model script, I have reinitialized the following layers:
cls_score = nn.Linear(num_inputs, num_classes)
bbox_pred = nn.Linear(num_inputs, num_bbox_reg_classes * 4)
mask_fcn_logits = nn.Conv2d(dim_reduced, num_classes, 1, 1, 0)
cls_logits = nn.Conv2d(dim_reduced, num_classes-1, 3, 3, 0)
conv5_mask = nn.ConvTranspose2d(dim_reduced, dim_reduced, 2, 2, 0)

newdict['model']['module.roi_heads.box.predictor.cls_score.weight'] = cls_score.weight
newdict['model']['module.roi_heads.box.predictor.cls_score.bias'] = cls_score.bias

newdict['model']['module.roi_heads.maskiou.predictor.maskiou.weight'] = cls_score.weight
newdict['model']['module.roi_heads.maskiou.predictor.maskiou.bias'] = cls_score.bias

newdict['model']['module.roi_heads.box.predictor.bbox_pred.weight'] = bbox_pred.weight
newdict['model']['module.roi_heads.box.predictor.bbox_pred.bias'] = bbox_pred.bias

newdict['model']['module.roi_heads.mask.predictor.mask_fcn_logits.weight'] = mask_fcn_logits.weight
newdict['model']['module.roi_heads.mask.predictor.mask_fcn_logits.bias'] = mask_fcn_logits.bias

newdict['model']['module.rpn.head.cls_logits.weight'] = cls_logits.weight
newdict['model']['module.rpn.head.cls_logits.bias'] = cls_logits.bias

newdict['model']['module.roi_heads.mask.predictor.conv5_mask.weight'] = conv5_mask.weight
newdict['model']['module.roi_heads.mask.predictor.conv5_mask.bias'] = conv5_mask.bias

After doing these, and on training, I am getting the following error:

Traceback (most recent call last):
File "train_net.py", line 189, in
main()
File "train_net.py", line 182, in main
model = train(cfg, args.local_rank, args.distributed)
File "train_net.py", line 88, in train
arguments,
File "F:\Codes\Centermask\CenterMask\maskrcnn_benchmark\engine\trainer.py", line 94, in do_train
optimizer.step()
File "E:\Anaconda3\envs\centermask_trial2\lib\site-packages\torch\optim\sgd.py", line 107, in step
p.data.add_(-group['lr'], d_p)
RuntimeError: output with shape [1, 256, 3, 3] doesn't match the broadcast shape [80, 256, 3, 3]

I am training for single class detection, but don't know how 80 suddenly popped up.
Any help would be appreciated. Thank you

weights file issue

@youngwanLEE
Hello. I want to share a problem, maybe it should corrected, I'm not quite sure :

Does the WeightFile-"centermask-R-101-FPN-ms-3x.pth" provided in GoogleDrive is wrong？

Because I use it in demo/centermask_demo.py and got a very trrible result. Also its size(410MB) is different from the weight-file with a same file name(535MB) provided in the homepage-README.

how to train my own datasets with different num_class

using centermask_V_99_eSE_FPN_ms_3x.yaml and centermask-V2-99-FPN-ms-3x.pth
it occurs this bug:

self.class.name, "\n\t".join(error_msgs)))
RuntimeError: Error(s) in loading state_dict for GeneralizedRCNN:
size mismatch for roi_heads.mask.predictor.mask_fcn_logits.weight: copying a param with shape torch.Size([81, 256, 1, 1]) from checkpoint, theshape in current model is torch.Size([2, 256, 1, 1]).
size mismatch for roi_heads.mask.predictor.mask_fcn_logits.bias: copying a param with shape torch.Size([81]) from checkpoint, the shape in current model is torch.Size([2]).
size mismatch for roi_heads.maskiou.predictor.maskiou.weight: copying a param with shape torch.Size([81, 1024]) from checkpoint, the shape in current model is torch.Size([2, 1024]).
size mismatch for roi_heads.maskiou.predictor.maskiou.bias: copying a param with shape torch.Size([81]) from checkpoint, the shape in current model is torch.Size([2]).

Training on Cityscapes

@youngwanLEE @lsrock1 thanks for sharing your wonderful wanted to knw if you have done any experiements / training on Cityscapes dataset , can your share your results and pretrained modesl ??

some instances don't have mask

I use centermask_demo.py to test my model，i find some instance result don't have mask.

centermask_demo.py: list index out of range

I trained model with custom dataset.
"python tools/test_net.py" works fine, shows results.
But when I try to run centermask_demo.py to visialize results I get error.
Problem is with custom dataset only, pretrained model from site works fine.

cmd line:

py demo\centermask_demo.py --input z:\files\plate\testing\in.mp4 --config-file "configs/centermask/centermask_V_19_eSE_FPN_lite_res600_ms_bs16_4x.yaml" --weights "datasets/model_00100000.pth"

result:

file 0
Traceback (most recent call last):
  File "demo\centermask_demo.py", line 167, in <module>
    main()
  File "demo\centermask_demo.py", line 136, in main
    composite = coco_demo.run_on_opencv_image(img)
  File "e:\files\tmp\CenterMask\demo\predictor.py", line 223, in run_on_opencv_image
    predictions = self.compute_prediction(image)
  File "e:\files\tmp\CenterMask\demo\predictor.py", line 261, in compute_prediction
    prediction = predictions[0]
IndexError: list index out of range

Updates regarding Detectron2

@youngwanLEE
Any updates regarding the detectron2 implementation?

Can you share your pretrained model?

I can't download your model according to your linkd website in README

Have u consider add mask head on centernet rather than FCOS?

what should we do to apply such a mask head on top of centernet? Did you make any experiments on this?

An error in your paper

I notice that channel attention maps' shape is 1 by h by w. But for SE module, it should be C by 1 by 1.

which yaml file is for centermask-V2-99-FPN-ms-3x.pth model

i wanna use centermask-V2-99-FPN-ms-3x.pth model for my own datasets
which yaml should i choose and how modify the file?

segmentation performance is low

Hi，I use the v39_ Lite model you released to test COCO2017 val dataset, the detection accuracy is right but the segmentation accuracy is particularly low. The average value is only about 0.11. Can you tell me what's the reason? Is there any problem in my test? Thank you very much for answering my questions~~

how to use my own dataset?

Excuse me, as title i want to train centermask on my own dataset. i already prepared my own dataset according to COCO format. But i dont know where the code i should change or how to package the data. Can you tell me how to do this work? Thank you.

Evaluation on mask

I use your pre-trained model which is CenterMask-R-101-FPN, but I get a low score.

Evaluate annotation type *bbox*                                                                                                                                                                    [69/1857]
DONE (t=24.09s).
Accumulating evaluation results...
DONE (t=7.39s).
 Average Precision  (AP) @[ IoU=0.50:0.95 | area=   all | maxDets=100 ] = 0.371
 Average Precision  (AP) @[ IoU=0.50      | area=   all | maxDets=100 ] = 0.550
 Average Precision  (AP) @[ IoU=0.75      | area=   all | maxDets=100 ] = 0.402
 Average Precision  (AP) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.204
 Average Precision  (AP) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.405
 Average Precision  (AP) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.487
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets=  1 ] = 0.313
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets= 10 ] = 0.505
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets=100 ] = 0.529
 Average Recall     (AR) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.314
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.575
 Average Recall     (AR) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.693

,

Evaluate annotation type *segm*
DONE (t=31.79s).
Accumulating evaluation results...
DONE (t=7.15s).
Average Precision  (AP) @[ IoU=0.50:0.95 \| area=   all \| maxDets=100 ] = 0.014                                                                                                                    Average Precision  (AP) @[ IoU=0.50      \| area=   all \| maxDets=100 ] = 0.060
Average Precision  (AP) @[ IoU=0.75      \| area=   all \| maxDets=100 ] = 0.002
Average Precision  (AP) @[ IoU=0.50:0.95 \| area= small \| maxDets=100 ] = 0.006
Average Precision  (AP) @[ IoU=0.50:0.95 \| area=medium \| maxDets=100 ] = 0.015
Average Precision  (AP) @[ IoU=0.50:0.95 \| area= large \| maxDets=100 ] = 0.029
Average Recall     (AR) @[ IoU=0.50:0.95 \| area=   all \| maxDets=  1 ] = 0.025
Average Recall     (AR) @[ IoU=0.50:0.95 \| area=   all \| maxDets= 10 ] = 0.039
Average Recall     (AR) @[ IoU=0.50:0.95 \| area=   all \| maxDets=100 ] = 0.041
Average Recall     (AR) @[ IoU=0.50:0.95 \| area= small \| maxDets=100 ] = 0.024
Average Recall     (AR) @[ IoU=0.50:0.95 \| area=medium \| maxDets=100 ] = 0.042
Average Recall     (AR) @[ IoU=0.50:0.95 \| area= large \| maxDets=100 ] = 0.059

And, final log

2020-01-10 12:00:03,442 maskrcnn_benchmark.inference INFO: OrderedDict([('bbox', OrderedDict([('AP', 0.3710008333187123), ('AP50', 0.5504566002409236), ('AP75', 0.40191244330658515), ('APs', 0.20413773326
243875), ('APm', 0.4050982273825964), ('APl', 0.48719014575853675)])), ('segm', OrderedDict([('AP', 0.013625224849318739), ('AP50', 0.06011946156392893), ('AP75', 0.002384730090967514), ('APs', 0.00600034
9337834598), ('APm', 0.015377757756321447), ('APl', 0.028622499625928898)]))])

How can I get the Mask AP score?

Why is the order of "P3-P7" different to other papers?

Training Problem

@youngwanLEE
Thank you for sharing such a great work.
Iam using your Centermask on my dataset.
But the loss is so weird
I have captured it here:

So I changed to coco dataset, and the loss is the nearly the same.
It looks so strange.
Can you have me to figure out what the problem here is ?

ERROR in coco_eval.py when runing test_net.py

@youngwanLEE
When I tried to test the model, there are some touble fllow:
Traceback (most recent call last): File "tools/test_net.py", line 97, in <module> main() File "tools/test_net.py", line 91, in main output_folder=output_folder, File "/home/sadik/CenterMask/maskrcnn_benchmark/engine/inference.py", line 117, in inference **extra_args) File "/home/sadik/CenterMask/maskrcnn_benchmark/data/datasets/evaluation/__init__.py", line 22, in evaluate return coco_evaluation(**args) File "/home/sadik/CenterMask/maskrcnn_benchmark/data/datasets/evaluation/coco/__init__.py", line 20, in coco_evaluation expected_results_sigma_tol=expected_results_sigma_tol, File "/home/sadik/CenterMask/maskrcnn_benchmark/data/datasets/evaluation/coco/coco_eval.py", line 44, in do_coco_evaluation coco_results["bbox"] = prepare_for_coco_detection(predictions, dataset) File "/home/sadik/CenterMask/maskrcnn_benchmark/data/datasets/evaluation/coco/coco_eval.py", line 75, in prepare_for_coco_detection if len(prediction) == 0: TypeError: object of type 'int' has no len()
I tried to convert prediction to list, but it is obviously wrong.Thank for any help

Why Can't I get the mask or boxes?

Hi,

When I run the configs/caffe2/e2e_mask_rcnn_R_50_FPN_1x_caffe2.yaml, I can get the mask and boxes. But when I try to use centermask, configs/centermask/centermask_M_v2_FPN_lite_res600_ms_bs32_1x.yaml, I can get nothing.

Here are my code:

import numpy as np
from maskrcnn_benchmark.config import cfg
from predictor import COCODemo

config_file = "../configs/centermask/centermask_M_v2_FPN_lite_res600_ms_bs32_1x.yaml"

# update the config options with the config file
cfg.merge_from_file(config_file)
# manual override some options
cfg.merge_from_list(["MODEL.DEVICE", "cuda"])

coco_demo = COCODemo(
    cfg,
    confidence_threshold=0.2
)

img = cv2.imread('Test.jpg')
predictions = coco_demo.run_on_opencv_image(color_image)

When I show the predictions, I just get the original image.
Can you help me?

Best,
Mikoy Chinese

_pickle.UnpicklingError: pickle data was truncated

Hi,when i try to run demo/webcam.py i get the error
_pickle.UnpicklingError: pickle data was truncated

Here is the full result:
Traceback (most recent call last):
File "webcam.py", line 80, in
main()
File "webcam.py", line 64, in main
min_image_size=args.min_image_size,
File "D:\Project\CenterMask-master\demo\predictor.py", line 159, in init
_ = checkpointer.load(cfg.MODEL.WEIGHT)
File "d:\project\centermask-master\maskrcnn_benchmark\utils\checkpoint.py", line 62, in load
checkpoint = self._load_file(f)
File "d:\project\centermask-master\maskrcnn_benchmark\utils\checkpoint.py", line 138, in _load_file
return load_c2_format(self.cfg, f)
File "d:\project\centermask-master\maskrcnn_benchmark\utils\c2_model_loading.py", line 175, in load_c2_format
return C2_FORMAT_LOADER[cfg.MODEL.BACKBONE.CONV_BODY](cfg, f)
File "d:\project\centermask-master\maskrcnn_benchmark\utils\c2_model_loading.py", line 165, in load_resnet_c2_format
state_dict = _load_c2_pickled_weights(f)
File "d:\project\centermask-master\maskrcnn_benchmark\utils\c2_model_loading.py", line 136, in _load_c2_pickled_weights
data = pickle.load(f, encoding="latin1")
_pickle.UnpicklingError: pickle data was truncated