This is my conf.

dataset_type = 'DOTADataset' data_root = '/media/giant/Data/WorkSpace/hw/mmrotate-main/data/split_1024_dota1_0/' 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='le90'), dict( type='Normalize', mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True), 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=(1024, 1024), flip=False, transforms=[ dict(type='RResize'), dict( type='Normalize', mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True), dict(type='Pad', size_divisor=32), dict(type='DefaultFormatBundle'), dict(type='Collect', keys=['img']) ]) ] data = dict( samples_per_gpu=2, workers_per_gpu=2, train=dict( type='DOTADataset', ann_file= '/media/giant/Data/WorkSpace/hw/mmrotate-main/data/split_1024_dota1_0/trainval/annfiles/', img_prefix= '/media/giant/Data/WorkSpace/hw/mmrotate-main/data/split_1024_dota1_0/trainval/images/', 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='le90'), dict( type='Normalize', mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True), dict(type='Pad', size_divisor=32), dict(type='DefaultFormatBundle'), dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']) ], version='le90'), val=dict( type='DOTADataset', ann_file= '/media/giant/Data/WorkSpace/hw/mmrotate-main/data/split_1024_dota1_0/trainval/annfiles/', img_prefix= '/media/giant/Data/WorkSpace/hw/mmrotate-main/data/split_1024_dota1_0/trainval/images/', pipeline=[ dict(type='LoadImageFromFile'), dict( type='MultiScaleFlipAug', img_scale=(1024, 1024), flip=False, transforms=[ dict(type='RResize'), dict( type='Normalize', mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True), dict(type='Pad', size_divisor=32), dict(type='DefaultFormatBundle'), dict(type='Collect', keys=['img']) ]) ], version='le90'), test=dict( type='DOTADataset', ann_file= '/media/giant/Data/WorkSpace/hw/mmrotate-main/data/split_1024_dota1_0/test/images/', img_prefix= '/media/giant/Data/WorkSpace/hw/mmrotate-main/data/split_1024_dota1_0/test/images/', pipeline=[ dict(type='LoadImageFromFile'), dict( type='MultiScaleFlipAug', img_scale=(1024, 1024), flip=False, transforms=[ dict(type='RResize'), dict( type='Normalize', mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True), dict(type='Pad', size_divisor=32), dict(type='DefaultFormatBundle'), dict(type='Collect', keys=['img']) ]) ], version='le90')) evaluation = dict(interval=1, metric='mAP') optimizer = dict(type='SGD', lr=0.005, momentum=0.9, weight_decay=0.0001) optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2)) lr_config = dict( policy='step', warmup='linear', warmup_iters=500, warmup_ratio=0.3333333333333333, step=[8, 11]) runner = dict(type='EpochBasedRunner', max_epochs=12) checkpoint_config = dict(interval=12) log_config = dict(interval=50, hooks=[dict(type='TextLoggerHook')]) dist_params = dict(backend='nccl') log_level = 'INFO' load_from = None resume_from = None workflow = [('train', 1)] opencv_num_threads = 0 mp_start_method = 'fork' angle_version = 'le90' model = dict( type='OrientedRCNN', backbone=dict( type='ResNet', depth=50, num_stages=4, out_indices=(0, 1, 2, 3), frozen_stages=1, norm_cfg=dict(type='BN', requires_grad=True), norm_eval=True, style='pytorch', pretrained= '/media/giant/Data/WorkSpace/hw/mmrotate-main/pre_model/oriented_rcnn_r50_fpn_1x_dota_le90-6d2b2ce0.pth' ), neck=dict( type='FPN', in_channels=[256, 512, 1024, 2048], out_channels=256, num_outs=5), rpn_head=dict( type='OrientedRPNHead', in_channels=256, feat_channels=256, version='le90', anchor_generator=dict( type='AnchorGenerator', scales=[8], ratios=[0.5, 1.0, 2.0], strides=[4, 8, 16, 32, 64]), bbox_coder=dict( type='MidpointOffsetCoder', angle_range='le90', target_means=[0.0, 0.0, 0.0, 0.0, 0.0, 0.0], target_stds=[1.0, 1.0, 1.0, 1.0, 0.5, 0.5]), loss_cls=dict( type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0), loss_bbox=dict( type='SmoothL1Loss', beta=0.1111111111111111, loss_weight=1.0)), roi_head=dict( type='OrientedStandardRoIHead', bbox_roi_extractor=dict( type='RotatedSingleRoIExtractor', roi_layer=dict( type='RoIAlignRotated', out_size=7, sample_num=2, 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=15, bbox_coder=dict( type='DeltaXYWHAOBBoxCoder', angle_range='le90', norm_factor=None, edge_swap=True, proj_xy=True, target_means=(0.0, 0.0, 0.0, 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))), 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=0.8), 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, iou_calculator=dict(type='RBboxOverlaps2D'), ignore_iof_thr=-1), 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=0.8), 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))) work_dir = '/media/giant/Data/WorkSpace/hw/mmrotate-main/work_dirs/oriented_RCNN_3' auto_resume = False gpu_ids = range(0, 1)

Dota datasets used ss_trainval.json,ss_test.json

python tools/test.py /media/giant/Data/WorkSpace/hw/mmrotate-main/configs/oriented_rcnn/oriented_rcnn_r50_fpn_1x_dota_le90.py /media/giant/Data/WorkSpace/hw/mmrotate-main/work_dirs/oriented_RCNN_3/latest.pth --format-only --eval-options submission_dir=/media/giant/Data/WorkSpace/hw/mmrotate-main/work_dirs/oriented_RCNN_3/Task1_results

my test result is: This is your evaluation result for task 1 (VOC metrics):

mAP: 0.4646640734613539 ap of each class: plane:0.7987955222037281, baseball-diamond:0.3772997959320968, bridge:0.2193527129605339, ground-track-field:0.14461461362651976, small-vehicle:0.684794349188321, large-vehicle:0.7142836492183425, ship:0.7630970486078997, tennis-court:0.8339892706765937, basketball-court:0.2951130331270134, storage-tank:0.6383442164219478, soccer-ball-field:0.09673222235202107, roundabout:0.18280349757502767, harbor:0.4440369802128685, swimming-pool:0.5310300747688381, helicopter:0.24567411504855508 COCO style result:

AP50: 0.4646640734613539 AP75: 0.23914854912215705 mAP: 0.25241695775786577

I would be grateful if you could reply to me，Thank the authors for providing such an excellent rotating target detection framework.

Suggest a potential alternative/fix

No response

You can upload the log file in the training process and use python mmrotate/utils/collect_env.py uploads your training environment details file.

您可以在训练过程中上传日志文件，并使用 python mmrotate/utils/collect_env.py 上传您的训练环境详细信息文件。 I will try after this trian，i used datasets flipped horizontally and vertically, .thanks for you reply.

In my dota_v1.0 dataset split folder:

du -hd1 /data/split_ss_dota/
6.8G    /data/split_ss_dota/val
14G /data/split_ss_dota/test
21G /data/split_ss_dota/train
28G /data/split_ss_dota/trainval
70G /data/split_ss_dota/

for reference.

You can check whether your dataset is consistent with mine, or whether it is cut in a reasonable way.

在我的 dota_v1.0 数据集拆分文件夹中：
du -hd1 /data/split_ss_dota/
6.8G  /data/split_ss_dota/val
14G   /data/split_ss_dota/test
21G   /data/split_ss_dota/train
28G   /data/split_ss_dota/trainval
70G   /data/split_ss_dota/
以供参考。

您可以检查您的数据集是否与我的一致，或者是否以合理的方式切割。

It seems the same

您可以在训练过程中上传日志文件，并使用 python mmrotate/utils/collect_env.py 上传您的训练环境详细信息文件。

/home/giant/miniconda3/envs/mmrt/bin/python /media/giant/Data/WorkSpace/hw/mmrotate-main/mmrotate/utils/collect_env.py sys.platform: linux Python: 3.7.11 (default, Jul 27 2021, 14:32:16) [GCC 7.5.0] CUDA available: True GPU 0: NVIDIA GeForce RTX 2080 Ti CUDA_HOME: /usr/local/cuda NVCC: Cuda compilation tools, release 10.2, V10.2.8 GCC: gcc (Ubuntu 7.5.0-6ubuntu2) 7.5.0 PyTorch: 1.7.0 PyTorch compiling details: PyTorch built with:

GCC 7.3
C++ Version: 201402
Intel(R) oneAPI Math Kernel Library Version 2021.4-Product Build 20210904 for Intel(R) 64 architecture applications
Intel(R) MKL-DNN v1.6.0 (Git Hash 5ef631a030a6f73131c77892041042805a06064f)
OpenMP 201511 (a.k.a. OpenMP 4.5)
NNPACK is enabled
CPU capability usage: AVX2
CUDA Runtime 10.1
NVCC architecture flags: -gencode;arch=compute_37,code=sm_37;-gencode;arch=compute_50,code=sm_50;-gencode;arch=compute_60,code=sm_60;-gencode;arch=compute_61,code=sm_61;-gencode;arch=compute_70,code=sm_70;-gencode;arch=compute_75,code=sm_75;-gencode;arch=compute_37,code=compute_37
CuDNN 7.6.3
Magma 2.5.2
Build settings: BLAS=MKL, BUILD_TYPE=Release, CXX_FLAGS= -Wno-deprecated -fvisibility-inlines-hidden -DUSE_PTHREADPOOL -fopenmp -DNDEBUG -DUSE_FBGEMM -DUSE_QNNPACK -DUSE_PYTORCH_QNNPACK -DUSE_XNNPACK -DUSE_VULKAN_WRAPPER -O2 -fPIC -Wno-narrowing -Wall -Wextra -Werror=return-type -Wno-missing-field-initializers -Wno-type-limits -Wno-array-bounds -Wno-unknown-pragmas -Wno-sign-compare -Wno-unused-parameter -Wno-unused-variable -Wno-unused-function -Wno-unused-result -Wno-unused-local-typedefs -Wno-strict-overflow -Wno-strict-aliasing -Wno-error=deprecated-declarations -Wno-stringop-overflow -Wno-psabi -Wno-error=pedantic -Wno-error=redundant-decls -Wno-error=old-style-cast -fdiagnostics-color=always -faligned-new -Wno-unused-but-set-variable -Wno-maybe-uninitialized -fno-math-errno -fno-trapping-math -Werror=format -Wno-stringop-overflow, PERF_WITH_AVX=1, PERF_WITH_AVX2=1, PERF_WITH_AVX512=1, USE_CUDA=ON, USE_EXCEPTION_PTR=1, USE_GFLAGS=OFF, USE_GLOG=OFF, USE_MKL=ON, USE_MKLDNN=ON, USE_MPI=OFF, USE_NCCL=ON, USE_NNPACK=ON, USE_OPENMP=ON,

TorchVision: 0.8.0 OpenCV: 4.5.5 MMCV: 1.5.3 MMCV Compiler: GCC 7.3 MMCV CUDA Compiler: 10.1 MMRotate: 0.3.2+ fatal: not a git repository (or any parent up to mount point /media/giant) Stopping at filesystem boundary (GIT_DISCOVERY_ACROSS_FILESYSTEM not set).

您可以在训练过程中上传环境日志文件，并使用 python mmrotate/utils/collect_env.py 上传您的训练详细信息文件。

20221101_184251.log

Try init_cfg=dict(type='Pretrained', checkpoint='torchvision://resnet50') ?

Full configs and envs for ref.

2022-11-06 11:35:09,500 - mmrotate - INFO - Environment info:
------------------------------------------------------------
sys.platform: linux
Python: 3.8.13 (default, Oct 21 2022, 23:50:54) [GCC 11.2.0]
CUDA available: True
GPU 0,1,2,3,4,5,6,7,8,9: GeForce RTX 2080 Ti
CUDA_HOME: /home/cuda101
NVCC: Cuda compilation tools, release 10.1, V10.1.24
GCC: gcc (Ubuntu 5.4.0-6ubuntu1~16.04.12) 5.4.0 20160609
PyTorch: 1.8.0
PyTorch compiling details: PyTorch built with:
  - GCC 7.3
  - C++ Version: 201402
  - Intel(R) oneAPI Math Kernel Library Version 2021.4-Product Build 20210904 for Intel(R) 64 architecture applications
  - Intel(R) MKL-DNN v1.7.0 (Git Hash 7aed236906b1f7a05c0917e5257a1af05e9ff683)
  - OpenMP 201511 (a.k.a. OpenMP 4.5)
  - NNPACK is enabled
  - CPU capability usage: AVX2
  - CUDA Runtime 10.1
  - NVCC architecture flags: -gencode;arch=compute_37,code=sm_37;-gencode;arch=compute_50,code=sm_50;-gencode;arch=compute_60,code=sm_60;-gencode;arch=compute_61,code=sm_61;-gencode;arch=compute_70,code=sm_70;-gencode;arch=compute_75,code=sm_75;-gencode;arch=compute_37,code=compute_37
  - CuDNN 7.6.3
  - Magma 2.5.2
  - Build settings: BLAS_INFO=mkl, BUILD_TYPE=Release, CUDA_VERSION=10.1, CUDNN_VERSION=7.6.3, CXX_COMPILER=/opt/rh/devtoolset-7/root/usr/bin/c++, CXX_FLAGS= -Wno-deprecated -fvisibility-inlines-hidden -DUSE_PTHREADPOOL -fopenmp -DNDEBUG -DUSE_KINETO -DUSE_FBGEMM -DUSE_QNNPACK -DUSE_PYTORCH_QNNPACK -DUSE_XNNPACK -O2 -fPIC -Wno-narrowing -Wall -Wextra -Werror=return-type -Wno-missing-field-initializers -Wno-type-limits -Wno-array-bounds -Wno-unknown-pragmas -Wno-sign-compare -Wno-unused-parameter -Wno-unused-variable -Wno-unused-function -Wno-unused-result -Wno-unused-local-typedefs -Wno-strict-overflow -Wno-strict-aliasing -Wno-error=deprecated-declarations -Wno-stringop-overflow -Wno-psabi -Wno-error=pedantic -Wno-error=redundant-decls -Wno-error=old-style-cast -fdiagnostics-color=always -faligned-new -Wno-unused-but-set-variable -Wno-maybe-uninitialized -fno-math-errno -fno-trapping-math -Werror=format -Wno-stringop-overflow, LAPACK_INFO=mkl, PERF_WITH_AVX=1, PERF_WITH_AVX2=1, PERF_WITH_AVX512=1, TORCH_VERSION=1.8.0, USE_CUDA=ON, USE_CUDNN=ON, USE_EXCEPTION_PTR=1, USE_GFLAGS=OFF, USE_GLOG=OFF, USE_MKL=ON, USE_MKLDNN=ON, USE_MPI=OFF, USE_NCCL=ON, USE_NNPACK=ON, USE_OPENMP=ON, 

TorchVision: 0.9.0
OpenCV: 4.6.0
MMCV: 1.7.0
MMCV Compiler: GCC 7.3
MMCV CUDA Compiler: 10.1
MMRotate: 0.3.3+4bbef14
------------------------------------------------------------

2022-11-06 11:35:09,825 - mmrotate - INFO - Distributed training: False
2022-11-06 11:35:10,270 - mmrotate - INFO - Config:
dataset_type = 'DOTADataset'
data_root = 'data/split_ss_dota/'
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='le90'),
    dict(
        type='Normalize',
        mean=[123.675, 116.28, 103.53],
        std=[58.395, 57.12, 57.375],
        to_rgb=True),
    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=(1024, 1024),
        flip=False,
        transforms=[
            dict(type='RResize'),
            dict(
                type='Normalize',
                mean=[123.675, 116.28, 103.53],
                std=[58.395, 57.12, 57.375],
                to_rgb=True),
            dict(type='Pad', size_divisor=32),
            dict(type='DefaultFormatBundle'),
            dict(type='Collect', keys=['img'])
        ])
]
data = dict(
    samples_per_gpu=2,
    workers_per_gpu=2,
    train=dict(
        type='DOTADataset',
        ann_file='data/split_ss_dota/trainval/annfiles/',
        img_prefix='data/split_ss_dota/trainval/images/',
        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='le90'),
            dict(
                type='Normalize',
                mean=[123.675, 116.28, 103.53],
                std=[58.395, 57.12, 57.375],
                to_rgb=True),
            dict(type='Pad', size_divisor=32),
            dict(type='DefaultFormatBundle'),
            dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels'])
        ],
        version='le90'),
    val=dict(
        type='DOTADataset',
        ann_file='data/split_ss_dota/trainval/annfiles/',
        img_prefix='data/split_ss_dota/trainval/images/',
        pipeline=[
            dict(type='LoadImageFromFile'),
            dict(
                type='MultiScaleFlipAug',
                img_scale=(1024, 1024),
                flip=False,
                transforms=[
                    dict(type='RResize'),
                    dict(
                        type='Normalize',
                        mean=[123.675, 116.28, 103.53],
                        std=[58.395, 57.12, 57.375],
                        to_rgb=True),
                    dict(type='Pad', size_divisor=32),
                    dict(type='DefaultFormatBundle'),
                    dict(type='Collect', keys=['img'])
                ])
        ],
        version='le90'),
    test=dict(
        type='DOTADataset',
        ann_file='data/split_ss_dota/test/images/',
        img_prefix='data/split_ss_dota/test/images/',
        pipeline=[
            dict(type='LoadImageFromFile'),
            dict(
                type='MultiScaleFlipAug',
                img_scale=(1024, 1024),
                flip=False,
                transforms=[
                    dict(type='RResize'),
                    dict(
                        type='Normalize',
                        mean=[123.675, 116.28, 103.53],
                        std=[58.395, 57.12, 57.375],
                        to_rgb=True),
                    dict(type='Pad', size_divisor=32),
                    dict(type='DefaultFormatBundle'),
                    dict(type='Collect', keys=['img'])
                ])
        ],
        version='le90'))
evaluation = dict(interval=1, metric='mAP')
optimizer = dict(type='SGD', lr=0.005, momentum=0.9, weight_decay=0.0001)
optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
lr_config = dict(
    policy='step',
    warmup='linear',
    warmup_iters=500,
    warmup_ratio=0.3333333333333333,
    step=[8, 11])
runner = dict(type='EpochBasedRunner', max_epochs=12)
checkpoint_config = dict(interval=1)
log_config = dict(interval=50, hooks=[dict(type='TextLoggerHook')])
dist_params = dict(backend='nccl')
log_level = 'INFO'
load_from = None
resume_from = None
workflow = [('train', 1)]
opencv_num_threads = 0
mp_start_method = 'fork'
angle_version = 'le90'
model = dict(
    type='OrientedRCNN',
    backbone=dict(
        type='ResNet',
        depth=50,
        num_stages=4,
        out_indices=(0, 1, 2, 3),
        frozen_stages=1,
        norm_cfg=dict(type='BN', requires_grad=True),
        norm_eval=True,
        style='pytorch',
        init_cfg=dict(type='Pretrained', checkpoint='torchvision://resnet50')),
    neck=dict(
        type='FPN',
        in_channels=[256, 512, 1024, 2048],
        out_channels=256,
        num_outs=5),
    rpn_head=dict(
        type='OrientedRPNHead',
        in_channels=256,
        feat_channels=256,
        version='le90',
        anchor_generator=dict(
            type='AnchorGenerator',
            scales=[8],
            ratios=[0.5, 1.0, 2.0],
            strides=[4, 8, 16, 32, 64]),
        bbox_coder=dict(
            type='MidpointOffsetCoder',
            angle_range='le90',
            target_means=[0.0, 0.0, 0.0, 0.0, 0.0, 0.0],
            target_stds=[1.0, 1.0, 1.0, 1.0, 0.5, 0.5]),
        loss_cls=dict(
            type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0),
        loss_bbox=dict(
            type='SmoothL1Loss', beta=0.1111111111111111, loss_weight=1.0)),
    roi_head=dict(
        type='OrientedStandardRoIHead',
        bbox_roi_extractor=dict(
            type='RotatedSingleRoIExtractor',
            roi_layer=dict(
                type='RoIAlignRotated',
                out_size=7,
                sample_num=2,
                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=15,
            bbox_coder=dict(
                type='DeltaXYWHAOBBoxCoder',
                angle_range='le90',
                norm_factor=None,
                edge_swap=True,
                proj_xy=True,
                target_means=(0.0, 0.0, 0.0, 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))),
    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=0.8),
            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,
                iou_calculator=dict(type='RBboxOverlaps2D'),
                ignore_iof_thr=-1),
            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=0.8),
            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)))
work_dir = './work_dirs/oriented_rcnn_r50_fpn_1x_dota_le90'
auto_resume = False
gpu_ids = range(0, 1)

I found a log file of the previous test.

20221029_222707.log.json 0.5151 0.6180 0.6611 0.6974 0.7238 0.7250 0.7475 0.7670 0.8075 0.8141 0.8196 0.8236

20221029_222707.log

试试 init_cfg=dict(type='Pretrained', checkpoint='torchvision://resnet50') ？

It seems so far，very thank you，This problem has troubled me for a long time.

This is your evaluation result for task 1 (VOC metrics):

mAP: 0.748688746080909 ap of each class: plane:0.8929113390266917, baseball-diamond:0.8170035709474821, bridge:0.5308325728979498, ground-track-field:0.7286759588213109, small-vehicle:0.7901468365737799, large-vehicle:0.8251776388584054, ship:0.8798433759261488, tennis-court:0.9088789489714909, basketball-court:0.8740424248111504, storage-tank:0.8425276194584406, soccer-ball-field:0.6235326863417908, roundabout:0.6222381296635777, harbor:0.7282195800989445, swimming-pool:0.6882721206132796, helicopter:0.47802838820319293 COCO style result:

AP50: 0.748688746080909 AP75: 0.4687732933656175 mAP: 0.44471852462069883

Thanks again for your reply and help.

open-mmlab / mmrotate

[Docs] oriented RCNN recurrence，I can't get good mAP #594

Branch

📚 The doc issue

This is my conf.

Dota datasets used ss_trainval.json,ss_test.json

AP50: 0.4646640734613539 AP75: 0.23914854912215705 mAP: 0.25241695775786577

Suggest a potential alternative/fix