Mmdetection: Core Adumpted - when training on my own dataset (VOC format)

[CrazyStoneonRoad]

When I start to train faster-rcnn-r50 on my own dataset, there's a problem I cannot figure out.

I defined my own data set type and initialized it.

from .registry import DATASETS
from .xml_style import XMLDataset


@DATASETS.register_module
class HRRSDDataset(XMLDataset):

    CLASSES = ('ship', 'bridge', 'ground track field', 'storage tank',
               'basketball court', 'tennis court', 'airplane', 'baseball diamond',
               'harbor', 'vehicle', 'crossroad', 'T junction', 'parking lot')

    def __init__(self, **kwargs):
        super(HRRSDDataset, self).__init__(**kwargs)
        self.year = 2007

My configurations:

# model settings
model = dict(
    type='FasterRCNN',
    pretrained='torchvision://resnet50',
    backbone=dict(
        type='ResNet',
        depth=50,
        num_stages=4,
        out_indices=(0, 1, 2, 3),
        frozen_stages=1,
        style='pytorch'),
    neck=dict(
        type='FPN',
        in_channels=[256, 512, 1024, 2048],
        out_channels=256,
        num_outs=5),
    rpn_head=dict(
        type='RPNHead',
        in_channels=256,
        feat_channels=256,
        anchor_scales=[8],
        anchor_ratios=[0.5, 1.0, 2.0],
        anchor_strides=[4, 8, 16, 32, 64],
        target_means=[.0, .0, .0, .0],
        target_stds=[1.0, 1.0, 1.0, 1.0],
        loss_cls=dict(
            type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0),
        loss_bbox=dict(type='SmoothL1Loss', beta=1.0 / 9.0, loss_weight=1.0)),
    bbox_roi_extractor=dict(
        type='SingleRoIExtractor',
        roi_layer=dict(type='RoIAlign', out_size=7, sample_num=2),
        out_channels=256,
        featmap_strides=[4, 8, 16, 32]),
    bbox_head=dict(
        type='SharedFCBBoxHead',
        num_fcs=2,
        in_channels=256,
        fc_out_channels=1024,
        roi_feat_size=7,
        num_classes=14,
        target_means=[0., 0., 0., 0.],
        target_stds=[0.1, 0.1, 0.2, 0.2],
        reg_class_agnostic=False,
        loss_cls=dict(
            type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
        loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)))
# model training and testing settings
train_cfg = dict(
    rpn=dict(
        assigner=dict(
            type='MaxIoUAssigner',
            pos_iou_thr=0.7,
            neg_iou_thr=0.3,
            min_pos_iou=0.3,
            ignore_iof_thr=-1),
        sampler=dict(
            type='RandomSampler',
            num=256,
            pos_fraction=0.5,
            neg_pos_ub=-1,
            add_gt_as_proposals=False),
        allowed_border=0,
        pos_weight=-1,
        debug=False),
    rpn_proposal=dict(
        nms_across_levels=False,
        nms_pre=2000,
        nms_post=2000,
        max_num=2000,
        nms_thr=0.7,
        min_bbox_size=0),
    rcnn=dict(
        assigner=dict(
            type='MaxIoUAssigner',
            pos_iou_thr=0.5,
            neg_iou_thr=0.5,
            min_pos_iou=0.5,
            ignore_iof_thr=-1),
        sampler=dict(
            type='RandomSampler',
            num=512,
            pos_fraction=0.25,
            neg_pos_ub=-1,
            add_gt_as_proposals=True),
        pos_weight=-1,
        debug=False))
test_cfg = dict(
    rpn=dict(
        nms_across_levels=False,
        nms_pre=1000,
        nms_post=1000,
        max_num=1000,
        nms_thr=0.7,
        min_bbox_size=0),
    rcnn=dict(
        score_thr=0.05, nms=dict(type='nms', iou_thr=0.5), max_per_img=100)
    # soft-nms is also supported for rcnn testing
    # e.g., nms=dict(type='soft_nms', iou_thr=0.5, min_score=0.05)
)
# dataset settings
dataset_type = 'HRRSDDataset'
data_root = 'data/VOCdevkit/HRRSD/'
img_norm_cfg = dict(
    mean=[91.55, 92.79, 85.09], std=[54.28, 52.65, 51.91], to_rgb=True)
train_pipeline = [
    dict(type='LoadImageFromFile'),
    dict(type='LoadAnnotations', with_bbox=True),
    dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
    dict(type='RandomFlip', flip_ratio=0.5),
    dict(type='Normalize', **img_norm_cfg),
    dict(type='Pad', size_divisor=32),
    dict(type='DefaultFormatBundle'),
    dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
]
test_pipeline = [
    dict(type='LoadImageFromFile'),
    dict(
        type='MultiScaleFlipAug',
        img_scale=(1333, 800),
        flip=False,
        transforms=[
            dict(type='Resize', keep_ratio=True),
            dict(type='RandomFlip'),
            dict(type='Normalize', **img_norm_cfg),
            dict(type='Pad', size_divisor=32),
            dict(type='ImageToTensor', keys=['img']),
            dict(type='Collect', keys=['img']),
        ])
]
data = dict(
    imgs_per_gpu=2,
    workers_per_gpu=2,
    train=dict(
        type=dataset_type,
        ann_file=data_root + 'ImageSets/Main/train.txt',
        img_prefix=data_root,
        pipeline=train_pipeline),
    val=dict(
        type=dataset_type,
        ann_file=data_root + 'ImageSets/Main/val.txt',
        img_prefix=data_root,
        pipeline=test_pipeline),
    test=dict(
        type=dataset_type,
        ann_file=data_root + 'ImageSets/Main/val.txt',
        img_prefix=data_root,
        pipeline=test_pipeline))
# optimizer
optimizer = dict(type='SGD', lr=0.02, momentum=0.9, weight_decay=0.0001)
optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
# learning policy
lr_config = dict(
    policy='step',
    warmup='linear',
    warmup_iters=500,
    warmup_ratio=1.0 / 3,
    step=[8, 11])
checkpoint_config = dict(interval=1)
# yapf:disable
log_config = dict(
    interval=50,
    hooks=[
        dict(type='TextLoggerHook'),
        # dict(type='TensorboardLoggerHook')
    ])
# yapf:enable
# runtime settings
total_epochs = 12
dist_params = dict(backend='nccl')
log_level = 'INFO'
work_dir = './work_dirs/faster_rcnn_r50_fpn_1x--hrrsd'
load_from = None # 'checkpoints/faster_rcnn_r50_fpn_1x_20181010-3d1b3351.pth' #
resume_from = None
workflow = [('train', 1)]

Finally when I try to start training, I got:

(open-mmlab) [user3@~/CODE/mmdetection]$python tools/train.py configs/collin/faster_rcnn_r50_fpn_1x--hrrsd.py
2019-09-28 17:41:35,810 - INFO - Distributed training: False
2019-09-28 17:41:36,405 - INFO - load model from: torchvision://resnet50
2019-09-28 17:41:36,893 - WARNING - The model and loaded state dict do not match exactly

unexpected key in source state_dict: fc.weight, fc.bias

2019-09-28 17:41:43,913 - INFO - Start running, host: user3@optimal6, work_dir: /home/user3/CODE/mmdetection/work_dirs/faster_rcnn_r50_fpn_1x--hrrsd
2019-09-28 17:41:43,913 - INFO - workflow: [('train', 1)], max: 12 epochs
段错误 （核心已转储）

[CrazyStoneonRoad]

I tried to debug with my files, found out the error happened in:
mmdet/ops/nms/nms_wrapper.py

ln 43          inds = nms_cuda.nms( dets_th, iou_thr)

[CrazyStoneonRoad]

This problem seems to be a compiling error of .cpp files or .cu file under mmdet/ops/nms/src/.

[CrazyStoneonRoad]

OS & Env：
Python 3.7.3
GCC 7.3.0
PyTorch 1.2.0 （install with pip）
CUDA 10.1
gcc 4.8.5
g++ 4.9
ubuntu 14.04.6

[CrazyStoneonRoad]

This error is because of gcc version.
I finally upgraded gcc to 5.4 and g++ to 5.5.
After that, I removed the old mmdetection folder, and unzip a new one, then python setup.py develop.
Then I repeated what have been done before, and it started running.

[b03505036]

hi im facing the same problem,

[b03505036]

it did nothing after an hour

[b03505036]

gcc --version = 5.4

[b03505036]

pytorch = 1.1
cuda = 9.0

[CrazyStoneonRoad]

I'm sorry, but I don't think we have the same problem.

My running returned Segment error (core dumped) after several seconds.

But I still suggest you to check your g++ installation. (not gcc ! )