r3det-oc_r50_fpn_1x_dota.py

_base_ = [
    '../_base_/datasets/dota.py', '../_base_/schedules/schedule_1x.py',
    '../_base_/default_runtime.py'
]
angle_version = 'oc'

model = dict(
    type='RefineSingleStageDetector',
    data_preprocessor=dict(
        type='mmdet.DetDataPreprocessor',
        mean=[123.675, 116.28, 103.53],
        std=[58.395, 57.12, 57.375],
        bgr_to_rgb=True,
        pad_size_divisor=32,
        boxtype2tensor=False),
    backbone=dict(
        type='mmdet.ResNet',
        depth=50,
        num_stages=4,
        out_indices=(0, 1, 2, 3),
        frozen_stages=1,
        zero_init_residual=False,
        norm_cfg=dict(type='BN', requires_grad=True),
        norm_eval=True,
        style='pytorch',
        init_cfg=dict(type='Pretrained', checkpoint='torchvision://resnet50')),
    neck=dict(
        type='mmdet.FPN',
        in_channels=[256, 512, 1024, 2048],
        out_channels=256,
        start_level=1,
        add_extra_convs='on_input',
        num_outs=5),
    bbox_head_init=dict(
        type='R3Head',
        num_classes=15,
        in_channels=256,
        stacked_convs=4,
        feat_channels=256,
        anchor_generator=dict(
            type='FakeRotatedAnchorGenerator',
            angle_version=angle_version,
            octave_base_scale=4,
            scales_per_octave=3,
            ratios=[1.0, 0.5, 2.0],
            strides=[8, 16, 32, 64, 128]),
        bbox_coder=dict(
            type='DeltaXYWHTRBBoxCoder',
            angle_version=angle_version,
            norm_factor=None,
            edge_swap=False,
            proj_xy=False,
            target_means=(.0, .0, .0, .0, .0),
            target_stds=(1.0, 1.0, 1.0, 1.0, 1.0),
            use_box_type=False),
        loss_cls=dict(
            type='mmdet.FocalLoss',
            use_sigmoid=True,
            gamma=2.0,
            alpha=0.25,
            loss_weight=1.0),
        loss_bbox=dict(type='mmdet.SmoothL1Loss', beta=0.11, loss_weight=1.0)),
    bbox_head_refine=[
        dict(
            type='R3RefineHead',
            num_classes=15,
            in_channels=256,
            stacked_convs=4,
            feat_channels=256,
            frm_cfg=dict(
                type='FRM', feat_channels=256, strides=[8, 16, 32, 64, 128]),
            anchor_generator=dict(
                type='PseudoRotatedAnchorGenerator',
                strides=[8, 16, 32, 64, 128]),
            bbox_coder=dict(
                type='DeltaXYWHTRBBoxCoder',
                angle_version=angle_version,
                norm_factor=None,
                edge_swap=False,
                proj_xy=False,
                target_means=(0.0, 0.0, 0.0, 0.0, 0.0),
                target_stds=(1.0, 1.0, 1.0, 1.0, 1.0)),
            loss_cls=dict(
                type='mmdet.FocalLoss',
                use_sigmoid=True,
                gamma=2.0,
                alpha=0.25,
                loss_weight=1.0),
            loss_bbox=dict(
                type='mmdet.SmoothL1Loss', beta=0.11, loss_weight=1.0))
    ],
    train_cfg=dict(
        init=dict(
            assigner=dict(
                type='mmdet.MaxIoUAssigner',
                pos_iou_thr=0.5,
                neg_iou_thr=0.4,
                min_pos_iou=0,
                ignore_iof_thr=-1,
                iou_calculator=dict(type='RBboxOverlaps2D')),
            allowed_border=-1,
            pos_weight=-1,
            debug=False),
        refine=[
            dict(
                assigner=dict(
                    type='mmdet.MaxIoUAssigner',
                    pos_iou_thr=0.6,
                    neg_iou_thr=0.5,
                    min_pos_iou=0,
                    ignore_iof_thr=-1,
                    iou_calculator=dict(type='RBboxOverlaps2D')),
                allowed_border=-1,
                pos_weight=-1,
                debug=False)
        ],
        stage_loss_weights=[1.0]),
    test_cfg=dict(
        nms_pre=2000,
        min_bbox_size=0,
        score_thr=0.05,
        nms=dict(type='nms_rotated', iou_threshold=0.1),
        max_per_img=2000))