旋转框目标检测mmrotate v1.0.0rc1 之RTMDet训练DOTA的官方问题解析整理（四）

1、Batchsize和学习率问题

1、关于rotated_rtmdet_l-coco_pretrain-3x-dota_ms.py配置文件的batchsize和学习率设置

问题：

回答：

2、不同batchsize下s2anet的mAP #59

问题：

回答：

3、关于lr和batchsize的问题 #645

2、如何进行多尺度测试？ #201

3、为什么相同的物体分类分数相差很大？ #455

问题：

回答：

_base_ = ['./roi_trans_r50_fpn_1x_dota_le90.py']

data_root = 'datasets/split_ms_dotav1/'
angle_version = 'le90'
img_norm_cfg = dict(
    mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
train_pipeline = [
    dict(type='LoadImageFromFile'),
    dict(type='LoadAnnotations', with_bbox=True),
    dict(type='RResize', img_scale=(1024, 1024)),
    dict(
        type='RRandomFlip',
        flip_ratio=[0.25, 0.25, 0.25],
        direction=['horizontal', 'vertical', 'diagonal'],
        version=angle_version),
    dict(
        type='PolyRandomRotate',
        rotate_ratio=0.5,
        angles_range=180,
        auto_bound=False,
        rect_classes=[9, 11],
        version=angle_version),
    dict(type='Normalize', **img_norm_cfg),
    dict(type='Pad', size_divisor=32),
    dict(type='DefaultFormatBundle'),
    dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels'])
]
data = dict(
    train=dict(
        pipeline=train_pipeline,
        ann_file=data_root + 'trainval/annfiles/',
        img_prefix=data_root + 'trainval/images/'),
    val=dict(
        ann_file=data_root + 'trainval/annfiles/',
        img_prefix=data_root + 'trainval/images/'),
    test=dict(
        ann_file=data_root + 'test/images/',
        img_prefix=data_root + 'test/images/'))

model = dict(train_cfg=dict(rpn=dict(assigner=dict(gpu_assign_thr=200))))

4、检测细长物体的困难 #384

问题：

回答：

def gen_single_level_base_anchors(self,
                                      base_size,
                                      scales,
                                      ratios,
                                      center=None):
        """Generate base anchors of a single level.
        Args:
            base_size (int | float): Basic size of an anchor.
            scales (torch.Tensor): Scales of the anchor.
            ratios (torch.Tensor): The ratio between between the height
                and width of anchors in a single level.
            center (tuple[float], optional): The center of the base anchor
                related to a single feature grid. Defaults to None.
        Returns:
            torch.Tensor: Anchors in a single-level feature maps.
        """
        w = base_size
        h = base_size
        if center is None:
            x_center = self.center_offset * w
            y_center = self.center_offset * h
        else:
            x_center, y_center = center

        h_ratios = torch.sqrt(ratios)
        w_ratios = 1 / h_ratios
        if self.scale_major:
            ws = (w * w_ratios[:, None] * scales[None, :]).view(-1)
            hs = (h * h_ratios[:, None] * scales[None, :]).view(-1)
        else:
            ws = (w * scales[:, None] * w_ratios[None, :]).view(-1)
            hs = (h * scales[:, None] * h_ratios[None, :]).view(-1)

        # use float anchor and the anchor's center is aligned with the
        # pixel center
        base_anchors = [
            x_center - 0.5 * ws, y_center - 0.5 * hs, x_center + 0.5 * ws,
            y_center + 0.5 * hs
        ]
        base_anchors = torch.stack(base_anchors, dim=-1)

        return base_anchors

5、如何训练宽高比大的物体 #285

6、请问这个项目中通道的ms+rr和论文中常说的多维度测试和多维度训练有什么区别

问题：

回答：

7、如何在mmrotate中绘制特征图

问题：

回答：

你好@AllieLan，您可以尝试使用https://github.com/open-mmlab/mmyolo/blob/main/demo/featmap_vis_demo.py

8、oriented reppoints 支持 filter_empty_gt=False 的训练

问题：

回答：

9、[1.x] RTMDet-R (tiny) 内存不足的 CUDA，具有 24GB VRAM 和 batch_size=1

问题：

回答:

10、如何在自己的数据集上测试大场景图片？

问题：

回答：

你好@TheGreatTreatsby, 你可以试试https://github.com/CAPTAIN-WHU/DOTA_devkit

11、使用 DOTA V1.0 数据集时 CFA 重新分配过程中的张量不匹配错误

问题：

回答：

12、如何改变旋转框的定义范围(如何更改旋转框的定义范围)

问题：

回答：

13、如何获得精度和F1分数

问题：

回答：

在我的例子中，我修改了 eval_map.py 和我的 custumdataset.py

通过在 def eval_rbbox_map 中创建额外的变量来计算

cls_all_tp = np.sum(tp) cls_all_fp = np.sum(fp)

参考我项目的代码

我的项目

14、关于 R3Det 中的随机种子 #464

问题：

回答：

15、Question about random seed. #291

16、HRSC2016 数据集性能重新实现 #202

问题：

回答：

_base_ = [
    '../_base_/datasets/hrsc.py', '../_base_/schedules/schedule_3x.py',
    '../_base_/default_runtime.py'
]

angle_version = 'le90'
model = dict(
    type='ReDet',
    backbone=dict(
        type='ReResNet',
        depth=50,
        num_stages=4,
        out_indices=(0, 1, 2, 3),
        frozen_stages=1,
        style='pytorch',
        pretrained='./work_dirs/re_resnet50_c8_batch256-25b16846.pth'),
    neck=dict(
        type='ReFPN',
        in_channels=[256, 512, 1024, 2048],
        out_channels=256,
        num_outs=5),
    rpn_head=dict(
        type='RotatedRPNHead',
        in_channels=256,
        feat_channels=256,
        version=angle_version,
        anchor_generator=dict(
            type='AnchorGenerator',
            scales=[8],
            ratios=[0.5, 1.0, 2.0],
            strides=[4, 8, 16, 32, 64]),
        bbox_coder=dict(
            type='DeltaXYWHBBoxCoder',
            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)),
    roi_head=dict(
        type='RoITransRoIHead',
        version=angle_version,
        num_stages=2,
        stage_loss_weights=[1, 1],
        bbox_roi_extractor=[
            dict(
                type='SingleRoIExtractor',
                roi_layer=dict(
                    type='RoIAlign', output_size=7, sampling_ratio=0),
                out_channels=256,
                featmap_strides=[4, 8, 16, 32]),
            dict(
                type='RotatedSingleRoIExtractor',
                roi_layer=dict(
                    type='RiRoIAlignRotated',
                    out_size=7,
                    num_samples=2,
                    num_orientations=8,
                    clockwise=True),
                out_channels=256,
                featmap_strides=[4, 8, 16, 32]),
        ],
        bbox_head=[
            dict(
                type='RotatedShared2FCBBoxHead',
                in_channels=256,
                fc_out_channels=1024,
                roi_feat_size=7,
                num_classes=1,
                bbox_coder=dict(
                    type='DeltaXYWHAHBBoxCoder',
                    angle_range=angle_version,
                    norm_factor=2,
                    edge_swap=True,
                    target_means=[0., 0., 0., 0., 0.],
                    target_stds=[0.1, 0.1, 0.2, 0.2, 0.1]),
                reg_class_agnostic=True,
                loss_cls=dict(
                    type='CrossEntropyLoss',
                    use_sigmoid=False,
                    loss_weight=1.0),
                loss_bbox=dict(type='SmoothL1Loss', beta=1.0,
                               loss_weight=1.0)),
            dict(
                type='RotatedShared2FCBBoxHead',
                in_channels=256,
                fc_out_channels=1024,
                roi_feat_size=7,
                num_classes=1,
                bbox_coder=dict(
                    type='DeltaXYWHAOBBoxCoder',
                    angle_range=angle_version,
                    norm_factor=None,
                    edge_swap=True,
                    proj_xy=True,
                    target_means=[0., 0., 0., 0., 0.],
                    target_stds=[0.05, 0.05, 0.1, 0.1, 0.05]),
                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))
        ]),
    train_cfg=dict(
        rpn=dict(
            assigner=dict(
                type='MaxIoUAssigner',
                pos_iou_thr=0.7,
                neg_iou_thr=0.3,
                min_pos_iou=0.3,
                match_low_quality=True,
                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_pre=2000,
            max_per_img=2000,
            nms=dict(type='nms', iou_threshold=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,
                    match_low_quality=False,
                    ignore_iof_thr=-1,
                    iou_calculator=dict(type='BboxOverlaps2D')),
                sampler=dict(
                    type='RandomSampler',
                    num=512,
                    pos_fraction=0.25,
                    neg_pos_ub=-1,
                    add_gt_as_proposals=True),
                pos_weight=-1,
                debug=False),
            dict(
                assigner=dict(
                    type='MaxIoUAssigner',
                    pos_iou_thr=0.5,
                    neg_iou_thr=0.5,
                    min_pos_iou=0.5,
                    match_low_quality=False,
                    ignore_iof_thr=-1,
                    iou_calculator=dict(type='RBboxOverlaps2D')),
                sampler=dict(
                    type='RRandomSampler',
                    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_pre=2000,
            max_per_img=2000,
            nms=dict(type='nms', iou_threshold=0.7),
            min_bbox_size=0),
        rcnn=dict(
            nms_pre=2000,
            min_bbox_size=0,
            score_thr=0.05,
            nms=dict(iou_thr=0.1),
            max_per_img=2000)))

img_norm_cfg = dict(
    mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
train_pipeline = [
    dict(type='LoadImageFromFile'),
    dict(type='LoadAnnotations', with_bbox=True),
    dict(type='RResize', img_scale=(800, 512)),
    dict(type='RRandomFlip', 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=(800, 512),
        flip=False,
        transforms=[
            dict(type='RResize'),
            dict(type='Normalize', **img_norm_cfg),
            dict(type='Pad', size_divisor=32),
            dict(type='DefaultFormatBundle'),
            dict(type='Collect', keys=['img'])
        ])
]

dataset_type = 'HRSCDataset'
data_root = '/data/dataset_share/HRSC2016/HRSC2016/'
data = dict(
    samples_per_gpu=2,
    workers_per_gpu=2,
    train=dict(
        type=dataset_type,
        classwise=False,
        ann_file=data_root + 'ImageSets/trainval.txt',
        ann_subdir=data_root + 'FullDataSet/Annotations/',
        img_subdir=data_root + 'FullDataSet/AllImages/',
        pipeline=train_pipeline),
    val=dict(
        type=dataset_type,
        classwise=False,
        ann_file=data_root + 'ImageSets/test.txt',
        ann_subdir=data_root + 'FullDataSet/Annotations/',
        img_subdir=data_root + 'FullDataSet/AllImages/',
        pipeline=test_pipeline),
    test=dict(
        type=dataset_type,
        classwise=False,
        ann_file=data_root + 'ImageSets/test.txt',
        ann_subdir=data_root + 'FullDataSet/Annotations/',
        img_subdir=data_root + 'FullDataSet/AllImages/',
        pipeline=test_pipeline))

evaluation = dict(interval=12, metric='mAP')
optimizer = dict(lr=0.01)

#原因是问题者的Target_stds和官方不一致，同时学习率也不一致导致的，官方也是使用单张GPU进行模型训练的。

问题：

回答：

17、单个类别训练报错

问题：

回答：

18、HRSC2016 的 classwise 设置为 True 时，在评估时出现“IndexError: tuple index out of range”。 #182

19、尝试结合 swin-Transform 和 s2anet #217

问题：

_base_ = ['./s2anet_r50_fpn_1x_dota_le135.py']

pretrained = 'https://github.com/SwinTransformer/storage/releases/download/v1.0.0/swin_tiny_patch4_window7_224.pth'  # noqa

model = dict(
    backbone=dict(
        _delete_=True,
        type='SwinTransformer',
        embed_dims=96,
        depths=[2, 2, 6, 2],
        num_heads=[3, 6, 12, 24],
        window_size=7,
        mlp_ratio=4,
        qkv_bias=True,
        qk_scale=None,
        drop_rate=0.,
        attn_drop_rate=0.,
        drop_path_rate=0.2,
        patch_norm=True,
        out_indices=(0, 1, 2, 3),
        with_cp=False,
        convert_weights=True,
        init_cfg=dict(type='Pretrained', checkpoint=pretrained)),
    neck=dict(
        _delete_=True,
        type='FPN',
        in_channels=[96, 192, 384, 768],
        out_channels=256,
        num_outs=5))

optimizer = dict(
    _delete_=True,
    type='AdamW',
    lr=0.0001,
    betas=(0.9, 0.999),
    weight_decay=0.05,
    paramwise_cfg=dict(
        custom_keys={
            'absolute_pos_embed': dict(decay_mult=0.),
            'relative_position_bias_table': dict(decay_mult=0.),
            'norm': dict(decay_mult=0.)
        }))

20、数据标签过多次显示会爆炸、预测后不出指标结果；当数据标签过多时，显存会爆，预测后指标结果不显示； #333

21、一个对象在 oriented-reppoints 中有两个预测类 #426

22、loss降不下来 #330

23、当我使用 rmosaic 时如何可视化 #686

你好@QAQTATQAQTAT，您可以使用demo/image_demo.py可视化 rmosaic 的输出。rmosaic的使用方法可以参考https://github.com/open-mmlab/mmrotate/blob/dev-1.x/configs/rotated_rtmdet/rotated_rtmdet_tiny-300e-aug-hrsc.py 。

Mosaic( img_scale=(1024, 1024))-> Resize(scale=(2048, 2048))->RandomCrop(crop_size=(1024, 1024))

train_pipeline = [
    dict(
        type='mmdet.LoadImageFromFile',
        file_client_args={
    
    {_base_.file_client_args}}),
    dict(type='mmdet.LoadAnnotations', with_bbox=True, box_type='qbox'),
    dict(type='ConvertBoxType', box_type_mapping=dict(gt_bboxes='rbox')),
    dict(
        type='mmdet.CachedMosaic',
        img_scale=(800, 800),
        pad_val=114.0,
        max_cached_images=20,
        random_pop=False),
    dict(
        type='mmdet.RandomResize',
        resize_type='mmdet.Resize',
        scale=(1600, 1600),
        ratio_range=(0.5, 2.0),
        keep_ratio=True),
    dict(type='RandomRotate', prob=0.5, angle_range=180),
    dict(type='mmdet.RandomCrop', crop_size=(800, 800)),
    dict(type='mmdet.YOLOXHSVRandomAug'),
    dict(
        type='mmdet.RandomFlip',
        prob=0.75,
        direction=['horizontal', 'vertical', 'diagonal']),
    dict(type='mmdet.Pad', size=(800, 800), pad_val=dict(img=(114, 114, 114))),
    dict(
        type='mmdet.CachedMixUp',
        img_scale=(800, 800),
        ratio_range=(1.0, 1.0),
        max_cached_images=10,
        random_pop=False,
        pad_val=(114, 114, 114),
        prob=0.5),
    dict(type='mmdet.PackDetInputs')
]

24、[WIP] 在 TRR360D 中支持 RR360（旋转矩形 360）检测 #731

https://github.com/open-mmlab/mmrotate/pull/731

25、Oriented RCNN 不支持 iou loss？ #649