update code

mmdet/configs/_base_/models/mask_rcnn_r50_fpn.py

+# Copyright (c) OpenMMLab. All rights reserved.
+from mmcv.ops import RoIAlign, nms
+from mmengine.model.weight_init import PretrainedInit
+from torch.nn import BatchNorm2d
+
+from mmdet.models.backbones.resnet import ResNet
+from mmdet.models.data_preprocessors.data_preprocessor import \
+    DetDataPreprocessor
+from mmdet.models.dense_heads.rpn_head import RPNHead
+from mmdet.models.detectors.mask_rcnn import MaskRCNN
+from mmdet.models.losses.cross_entropy_loss import CrossEntropyLoss
+from mmdet.models.losses.smooth_l1_loss import L1Loss
+from mmdet.models.necks.fpn import FPN
+from mmdet.models.roi_heads.bbox_heads.convfc_bbox_head import \
+    Shared2FCBBoxHead
+from mmdet.models.roi_heads.mask_heads.fcn_mask_head import FCNMaskHead
+from mmdet.models.roi_heads.roi_extractors.single_level_roi_extractor import \
+    SingleRoIExtractor
+from mmdet.models.roi_heads.standard_roi_head import StandardRoIHead
+from mmdet.models.task_modules.assigners.max_iou_assigner import MaxIoUAssigner
+from mmdet.models.task_modules.coders.delta_xywh_bbox_coder import \
+    DeltaXYWHBBoxCoder
+from mmdet.models.task_modules.prior_generators.anchor_generator import \
+    AnchorGenerator
+from mmdet.models.task_modules.samplers.random_sampler import RandomSampler
+
+# model settings
+model = dict(
+    type=MaskRCNN,
+    data_preprocessor=dict(
+        type=DetDataPreprocessor,
+        mean=[123.675, 116.28, 103.53],
+        std=[58.395, 57.12, 57.375],
+        bgr_to_rgb=True,
+        pad_mask=True,
+        pad_size_divisor=32),
+    backbone=dict(
+        type=ResNet,
+        depth=50,
+        num_stages=4,
+        out_indices=(0, 1, 2, 3),
+        frozen_stages=1,
+        norm_cfg=dict(type=BatchNorm2d, requires_grad=True),
+        norm_eval=True,
+        style='pytorch',
+        init_cfg=dict(
+            type=PretrainedInit, checkpoint='torchvision://resnet50')),
+    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_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=L1Loss, loss_weight=1.0)),
+    roi_head=dict(
+        type=StandardRoIHead,
+        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]),
+        bbox_head=dict(
+            type=Shared2FCBBoxHead,
+            in_channels=256,
+            fc_out_channels=1024,
+            roi_feat_size=7,
+            num_classes=80,
+            bbox_coder=dict(
+                type=DeltaXYWHBBoxCoder,
+                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=L1Loss, loss_weight=1.0)),
+        mask_roi_extractor=dict(
+            type=SingleRoIExtractor,
+            roi_layer=dict(type=RoIAlign, output_size=14, sampling_ratio=0),
+            out_channels=256,
+            featmap_strides=[4, 8, 16, 32]),
+        mask_head=dict(
+            type=FCNMaskHead,
+            num_convs=4,
+            in_channels=256,
+            conv_out_channels=256,
+            num_classes=80,
+            loss_mask=dict(
+                type=CrossEntropyLoss, use_mask=True, 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,
+                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=-1,
+            pos_weight=-1,
+            debug=False),
+        rpn_proposal=dict(
+            nms_pre=2000,
+            max_per_img=1000,
+            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=True,
+                ignore_iof_thr=-1),
+            sampler=dict(
+                type=RandomSampler,
+                num=512,
+                pos_fraction=0.25,
+                neg_pos_ub=-1,
+                add_gt_as_proposals=True),
+            mask_size=28,
+            pos_weight=-1,
+            debug=False)),
+    test_cfg=dict(
+        rpn=dict(
+            nms_pre=1000,
+            max_per_img=1000,
+            nms=dict(type=nms, iou_threshold=0.7),
+            min_bbox_size=0),
+        rcnn=dict(
+            score_thr=0.05,
+            nms=dict(type=nms, iou_threshold=0.5),
+            max_per_img=100,
+            mask_thr_binary=0.5)))

mmdet/configs/_base_/models/retinanet_r50_fpn.py

+# Copyright (c) OpenMMLab. All rights reserved.
+from mmcv.ops import nms
+from torch.nn import BatchNorm2d
+
+from mmdet.models import (FPN, DetDataPreprocessor, FocalLoss, L1Loss, ResNet,
+                          RetinaHead, RetinaNet)
+from mmdet.models.task_modules import (AnchorGenerator, DeltaXYWHBBoxCoder,
+                                       MaxIoUAssigner, PseudoSampler)
+
+# model settings
+model = dict(
+    type=RetinaNet,
+    data_preprocessor=dict(
+        type=DetDataPreprocessor,
+        mean=[123.675, 116.28, 103.53],
+        std=[58.395, 57.12, 57.375],
+        bgr_to_rgb=True,
+        pad_size_divisor=32),
+    backbone=dict(
+        type=ResNet,
+        depth=50,
+        num_stages=4,
+        out_indices=(0, 1, 2, 3),
+        frozen_stages=1,
+        norm_cfg=dict(type=BatchNorm2d, 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,
+        start_level=1,
+        add_extra_convs='on_input',
+        num_outs=5),
+    bbox_head=dict(
+        type=RetinaHead,
+        num_classes=80,
+        in_channels=256,
+        stacked_convs=4,
+        feat_channels=256,
+        anchor_generator=dict(
+            type=AnchorGenerator,
+            octave_base_scale=4,
+            scales_per_octave=3,
+            ratios=[0.5, 1.0, 2.0],
+            strides=[8, 16, 32, 64, 128]),
+        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=FocalLoss,
+            use_sigmoid=True,
+            gamma=2.0,
+            alpha=0.25,
+            loss_weight=1.0),
+        loss_bbox=dict(type=L1Loss, loss_weight=1.0)),
+    # model training and testing settings
+    train_cfg=dict(
+        assigner=dict(
+            type=MaxIoUAssigner,
+            pos_iou_thr=0.5,
+            neg_iou_thr=0.4,
+            min_pos_iou=0,
+            ignore_iof_thr=-1),
+        sampler=dict(
+            type=PseudoSampler),  # Focal loss should use PseudoSampler
+        allowed_border=-1,
+        pos_weight=-1,
+        debug=False),
+    test_cfg=dict(
+        nms_pre=1000,
+        min_bbox_size=0,
+        score_thr=0.05,
+        nms=dict(type=nms, iou_threshold=0.5),
+        max_per_img=100))

mmdet/configs/_base_/schedules/schedule_1x.py

+# Copyright (c) OpenMMLab. All rights reserved.
+from mmengine.optim.optimizer.optimizer_wrapper import OptimWrapper
+from mmengine.optim.scheduler.lr_scheduler import LinearLR, MultiStepLR
+from mmengine.runner.loops import EpochBasedTrainLoop, TestLoop, ValLoop
+from torch.optim.sgd import SGD
+
+# training schedule for 1x
+train_cfg = dict(type=EpochBasedTrainLoop, max_epochs=12, val_interval=1)
+val_cfg = dict(type=ValLoop)
+test_cfg = dict(type=TestLoop)
+
+# learning rate
+param_scheduler = [
+    dict(type=LinearLR, start_factor=0.001, by_epoch=False, begin=0, end=500),
+    dict(
+        type=MultiStepLR,
+        begin=0,
+        end=12,
+        by_epoch=True,
+        milestones=[8, 11],
+        gamma=0.1)
+]
+
+# optimizer
+optim_wrapper = dict(
+    type=OptimWrapper,
+    optimizer=dict(type=SGD, lr=0.02, momentum=0.9, weight_decay=0.0001))
+
+# Default setting for scaling LR automatically
+#   - `enable` means enable scaling LR automatically
+#       or not by default.
+#   - `base_batch_size` = (8 GPUs) x (2 samples per GPU).
+auto_scale_lr = dict(enable=False, base_batch_size=16)

mmdet/configs/_base_/schedules/schedule_2x.py

+# Copyright (c) OpenMMLab. All rights reserved.
+from mmengine.optim.optimizer.optimizer_wrapper import OptimWrapper
+from mmengine.optim.scheduler.lr_scheduler import LinearLR, MultiStepLR
+from mmengine.runner.loops import EpochBasedTrainLoop, TestLoop, ValLoop
+from torch.optim.sgd import SGD
+
+# training schedule for 1x
+train_cfg = dict(type=EpochBasedTrainLoop, max_epochs=24, val_interval=1)
+val_cfg = dict(type=ValLoop)
+test_cfg = dict(type=TestLoop)
+
+# learning rate
+param_scheduler = [
+    dict(type=LinearLR, start_factor=0.001, by_epoch=False, begin=0, end=500),
+    dict(
+        type=MultiStepLR,
+        begin=0,
+        end=24,
+        by_epoch=True,
+        milestones=[16, 22],
+        gamma=0.1)
+]
+
+# optimizer
+optim_wrapper = dict(
+    type=OptimWrapper,
+    optimizer=dict(type=SGD, lr=0.02, momentum=0.9, weight_decay=0.0001))
+
+# Default setting for scaling LR automatically
+#   - `enable` means enable scaling LR automatically
+#       or not by default.
+#   - `base_batch_size` = (8 GPUs) x (2 samples per GPU).
+auto_scale_lr = dict(enable=False, base_batch_size=16)

mmdet/configs/cascade_rcnn/cascade_mask_rcnn_r50_fpn_1x_coco.py

+# Copyright (c) OpenMMLab. All rights reserved.
+
+# Please refer to https://mmengine.readthedocs.io/en/latest/advanced_tutorials/config.html#a-pure-python-style-configuration-file-beta for more details. # noqa
+# mmcv >= 2.0.1
+# mmengine >= 0.8.0
+
+from mmengine.config import read_base
+
+with read_base():
+    from .._base_.datasets.coco_instance import *
+    from .._base_.default_runtime import *
+    from .._base_.models.cascade_mask_rcnn_r50_fpn import *
+    from .._base_.schedules.schedule_1x import *

mmdet/configs/cascade_rcnn/cascade_rcnn_r50_fpn_1x_coco.py

+# Copyright (c) OpenMMLab. All rights reserved.
+
+# Please refer to https://mmengine.readthedocs.io/en/latest/advanced_tutorials/config.html#a-pure-python-style-configuration-file-beta for more details. # noqa
+# mmcv >= 2.0.1
+# mmengine >= 0.8.0
+
+from mmengine.config import read_base
+
+with read_base():
+    from .._base_.datasets.coco_detection import *
+    from .._base_.default_runtime import *
+    from .._base_.models.cascade_rcnn_r50_fpn import *
+    from .._base_.schedules.schedule_1x import *

mmdet/configs/common/lsj_100e_coco_detection.py

+# Copyright (c) OpenMMLab. All rights reserved.
+
+# Please refer to https://mmengine.readthedocs.io/en/latest/advanced_tutorials/config.html#a-pure-python-style-configuration-file-beta for more details. # noqa
+# mmcv >= 2.0.1
+# mmengine >= 0.8.0
+
+from mmengine.config import read_base
+
+with read_base():
+    from .._base_.default_runtime import *
+
+from mmengine.dataset.sampler import DefaultSampler
+from mmengine.optim import OptimWrapper
+from mmengine.optim.scheduler.lr_scheduler import LinearLR, MultiStepLR
+from mmengine.runner.loops import EpochBasedTrainLoop, TestLoop, ValLoop
+from torch.optim import SGD
+
+from mmdet.datasets import CocoDataset, RepeatDataset
+from mmdet.datasets.transforms.formatting import PackDetInputs
+from mmdet.datasets.transforms.loading import (FilterAnnotations,
+                                               LoadAnnotations,
+                                               LoadImageFromFile)
+from mmdet.datasets.transforms.transforms import (CachedMixUp, CachedMosaic,
+                                                  Pad, RandomCrop, RandomFlip,
+                                                  RandomResize, Resize)
+from mmdet.evaluation import CocoMetric
+
+# dataset settings
+dataset_type = CocoDataset
+data_root = 'data/coco/'
+image_size = (1024, 1024)
+
+backend_args = None
+
+train_pipeline = [
+    dict(type=LoadImageFromFile, backend_args=backend_args),
+    dict(type=LoadAnnotations, with_bbox=True, with_mask=True),
+    dict(
+        type=RandomResize,
+        scale=image_size,
+        ratio_range=(0.1, 2.0),
+        keep_ratio=True),
+    dict(
+        type=RandomCrop,
+        crop_type='absolute_range',
+        crop_size=image_size,
+        recompute_bbox=True,
+        allow_negative_crop=True),
+    dict(type=FilterAnnotations, min_gt_bbox_wh=(1e-2, 1e-2)),
+    dict(type=RandomFlip, prob=0.5),
+    dict(type=PackDetInputs)
+]
+test_pipeline = [
+    dict(type=LoadImageFromFile, backend_args=backend_args),
+    dict(type=Resize, scale=(1333, 800), keep_ratio=True),
+    dict(type=LoadAnnotations, with_bbox=True),
+    dict(
+        type=PackDetInputs,
+        meta_keys=('img_id', 'img_path', 'ori_shape', 'img_shape',
+                   'scale_factor'))
+]
+
+# Use RepeatDataset to speed up training
+train_dataloader = dict(
+    batch_size=2,
+    num_workers=2,
+    persistent_workers=True,
+    sampler=dict(type=DefaultSampler, shuffle=True),
+    dataset=dict(
+        type=RepeatDataset,
+        times=4,  # simply change this from 2 to 16 for 50e - 400e training.
+        dataset=dict(
+            type=dataset_type,
+            data_root=data_root,
+            ann_file='annotations/instances_train2017.json',
+            data_prefix=dict(img='train2017/'),
+            filter_cfg=dict(filter_empty_gt=True, min_size=32),
+            pipeline=train_pipeline,
+            backend_args=backend_args)))
+val_dataloader = dict(
+    batch_size=1,
+    num_workers=2,
+    persistent_workers=True,
+    drop_last=False,
+    sampler=dict(type=DefaultSampler, shuffle=False),
+    dataset=dict(
+        type=dataset_type,
+        data_root=data_root,
+        ann_file='annotations/instances_val2017.json',
+        data_prefix=dict(img='val2017/'),
+        test_mode=True,
+        pipeline=test_pipeline,
+        backend_args=backend_args))
+test_dataloader = val_dataloader
+
+val_evaluator = dict(
+    type=CocoMetric,
+    ann_file=data_root + 'annotations/instances_val2017.json',
+    metric=['bbox', 'segm'],
+    format_only=False,
+    backend_args=backend_args)
+test_evaluator = val_evaluator
+
+max_epochs = 25
+
+train_cfg = dict(
+    type=EpochBasedTrainLoop, max_epochs=max_epochs, val_interval=5)
+val_cfg = dict(type=ValLoop)
+test_cfg = dict(type=TestLoop)
+
+# optimizer assumes bs=64
+optim_wrapper = dict(
+    type=OptimWrapper,
+    optimizer=dict(type=SGD, lr=0.1, momentum=0.9, weight_decay=0.00004))
+
+# learning rate
+param_scheduler = [
+    dict(type=LinearLR, start_factor=0.067, by_epoch=False, begin=0, end=500),
+    dict(
+        type=MultiStepLR,
+        begin=0,
+        end=max_epochs,
+        by_epoch=True,
+        milestones=[22, 24],
+        gamma=0.1)
+]
+
+# only keep latest 2 checkpoints
+default_hooks.update(dict(checkpoint=dict(max_keep_ckpts=2)))
+
+# NOTE: `auto_scale_lr` is for automatically scaling LR,
+# USER SHOULD NOT CHANGE ITS VALUES.
+# base_batch_size = (32 GPUs) x (2 samples per GPU)
+auto_scale_lr = dict(base_batch_size=64)

mmdet/configs/common/lsj_100e_coco_instance.py

+# Copyright (c) OpenMMLab. All rights reserved.
+
+# Please refer to https://mmengine.readthedocs.io/en/latest/advanced_tutorials/config.html#a-pure-python-style-configuration-file-beta for more details. # noqa
+# mmcv >= 2.0.1
+# mmengine >= 0.8.0
+
+from mmengine.config import read_base
+
+with read_base():
+    from .._base_.default_runtime import *
+
+from mmengine.dataset.sampler import DefaultSampler
+from mmengine.optim import OptimWrapper
+from mmengine.optim.scheduler.lr_scheduler import LinearLR, MultiStepLR
+from mmengine.runner.loops import EpochBasedTrainLoop, TestLoop, ValLoop
+from torch.optim import SGD
+
+from mmdet.datasets import CocoDataset, RepeatDataset
+from mmdet.datasets.transforms.formatting import PackDetInputs
+from mmdet.datasets.transforms.loading import (FilterAnnotations,
+                                               LoadAnnotations,
+                                               LoadImageFromFile)
+from mmdet.datasets.transforms.transforms import (CachedMixUp, CachedMosaic,
+                                                  Pad, RandomCrop, RandomFlip,
+                                                  RandomResize, Resize)
+from mmdet.evaluation import CocoMetric
+
+# dataset settings
+dataset_type = CocoDataset
+data_root = 'data/coco/'
+image_size = (1024, 1024)
+
+backend_args = None
+
+train_pipeline = [
+    dict(type=LoadImageFromFile, backend_args=backend_args),
+    dict(type=LoadAnnotations, with_bbox=True, with_mask=True),
+    dict(
+        type=RandomResize,
+        scale=image_size,
+        ratio_range=(0.1, 2.0),
+        keep_ratio=True),
+    dict(
+        type=RandomCrop,
+        crop_type='absolute_range',
+        crop_size=image_size,
+        recompute_bbox=True,
+        allow_negative_crop=True),
+    dict(type=FilterAnnotations, min_gt_bbox_wh=(1e-2, 1e-2)),
+    dict(type=RandomFlip, prob=0.5),
+    dict(type=PackDetInputs)
+]
+test_pipeline = [
+    dict(type=LoadImageFromFile, backend_args=backend_args),
+    dict(type=Resize, scale=(1333, 800), keep_ratio=True),
+    dict(type=LoadAnnotations, with_bbox=True, with_mask=True),
+    dict(
+        type=PackDetInputs,
+        meta_keys=('img_id', 'img_path', 'ori_shape', 'img_shape',
+                   'scale_factor'))
+]
+
+# Use RepeatDataset to speed up training
+train_dataloader = dict(
+    batch_size=2,
+    num_workers=2,
+    persistent_workers=True,
+    sampler=dict(type=DefaultSampler, shuffle=True),
+    dataset=dict(
+        type=RepeatDataset,
+        times=4,  # simply change this from 2 to 16 for 50e - 400e training.
+        dataset=dict(
+            type=dataset_type,
+            data_root=data_root,
+            ann_file='annotations/instances_train2017.json',
+            data_prefix=dict(img='train2017/'),
+            filter_cfg=dict(filter_empty_gt=True, min_size=32),
+            pipeline=train_pipeline,
+            backend_args=backend_args)))
+val_dataloader = dict(
+    batch_size=1,
+    num_workers=2,
+    persistent_workers=True,
+    drop_last=False,
+    sampler=dict(type=DefaultSampler, shuffle=False),
+    dataset=dict(
+        type=dataset_type,
+        data_root=data_root,
+        ann_file='annotations/instances_val2017.json',
+        data_prefix=dict(img='val2017/'),
+        test_mode=True,
+        pipeline=test_pipeline,
+        backend_args=backend_args))
+test_dataloader = val_dataloader
+
+val_evaluator = dict(
+    type=CocoMetric,
+    ann_file=data_root + 'annotations/instances_val2017.json',
+    metric=['bbox', 'segm'],
+    format_only=False,
+    backend_args=backend_args)
+test_evaluator = val_evaluator
+
+max_epochs = 25
+
+train_cfg = dict(
+    type=EpochBasedTrainLoop, max_epochs=max_epochs, val_interval=5)
+val_cfg = dict(type=ValLoop)
+test_cfg = dict(type=TestLoop)
+
+# optimizer assumes bs=64
+optim_wrapper = dict(
+    type=OptimWrapper,
+    optimizer=dict(type=SGD, lr=0.1, momentum=0.9, weight_decay=0.00004))
+
+# learning rate
+param_scheduler = [
+    dict(type=LinearLR, start_factor=0.067, by_epoch=False, begin=0, end=500),
+    dict(
+        type=MultiStepLR,
+        begin=0,
+        end=max_epochs,
+        by_epoch=True,
+        milestones=[22, 24],
+        gamma=0.1)
+]
+
+# only keep latest 2 checkpoints
+default_hooks.update(dict(checkpoint=dict(max_keep_ckpts=2)))
+
+# NOTE: `auto_scale_lr` is for automatically scaling LR,
+# USER SHOULD NOT CHANGE ITS VALUES.
+# base_batch_size = (32 GPUs) x (2 samples per GPU)
+auto_scale_lr = dict(base_batch_size=64)

mmdet/configs/common/lsj_200e_coco_detection.py

+# Copyright (c) OpenMMLab. All rights reserved.
+
+# Please refer to https://mmengine.readthedocs.io/en/latest/advanced_tutorials/config.html#a-pure-python-style-configuration-file-beta for more details. # noqa
+# mmcv >= 2.0.1
+# mmengine >= 0.8.0
+
+from mmengine.config import read_base
+
+with read_base():
+    from .lsj_100e_coco_detection import *
+
+# 8x25=200e
+train_dataloader.update(dict(dataset=dict(times=8)))
+
+# learning rate
+param_scheduler = [
+    dict(type=LinearLR, start_factor=0.067, by_epoch=False, begin=0, end=1000),
+    dict(
+        type=MultiStepLR,
+        begin=0,
+        end=25,
+        by_epoch=True,
+        milestones=[22, 24],
+        gamma=0.1)
+]

mmdet/configs/common/lsj_200e_coco_instance.py

+# Copyright (c) OpenMMLab. All rights reserved.
+
+# Please refer to https://mmengine.readthedocs.io/en/latest/advanced_tutorials/config.html#a-pure-python-style-configuration-file-beta for more details. # noqa
+# mmcv >= 2.0.1
+# mmengine >= 0.8.0
+
+from mmengine.config import read_base
+
+with read_base():
+    from .lsj_100e_coco_instance import *
+
+# 8x25=200e
+train_dataloader.update(dict(dataset=dict(times=8)))
+
+# learning rate
+param_scheduler = [
+    dict(type=LinearLR, start_factor=0.067, by_epoch=False, begin=0, end=1000),
+    dict(
+        type=MultiStepLR,
+        begin=0,
+        end=25,
+        by_epoch=True,
+        milestones=[22, 24],
+        gamma=0.1)
+]

mmdet/configs/common/ms_3x_coco.py

+# Copyright (c) OpenMMLab. All rights reserved.
+
+# Please refer to https://mmengine.readthedocs.io/en/latest/advanced_tutorials/config.html#a-pure-python-style-configuration-file-beta for more details. # noqa
+# mmcv >= 2.0.1
+# mmengine >= 0.8.0
+
+from mmengine.config import read_base
+
+with read_base():
+    from .._base_.default_runtime import *
+
+from mmcv.transforms import RandomResize
+from mmengine.dataset import RepeatDataset
+from mmengine.dataset.sampler import DefaultSampler
+from mmengine.optim import OptimWrapper
+from mmengine.optim.scheduler.lr_scheduler import LinearLR, MultiStepLR
+from mmengine.runner.loops import EpochBasedTrainLoop, TestLoop, ValLoop
+from torch.optim import SGD
+
+from mmdet.datasets import AspectRatioBatchSampler, CocoDataset
+from mmdet.datasets.transforms.formatting import PackDetInputs
+from mmdet.datasets.transforms.loading import (LoadAnnotations,
+                                               LoadImageFromFile)
+from mmdet.datasets.transforms.transforms import RandomFlip, Resize
+from mmdet.evaluation import CocoMetric
+
+# dataset settings
+dataset_type = CocoDataset
+data_root = 'data/coco/'
+
+# Example to use different file client
+# Method 1: simply set the data root and let the file I/O module
+# automatically infer from prefix (not support LMDB and Memcache yet)
+
+# data_root = 's3://openmmlab/datasets/detection/coco/'
+
+# Method 2: Use `backend_args`, `file_client_args` in versions before 3.0.0rc6
+# backend_args = dict(
+#     backend='petrel',
+#     path_mapping=dict({
+#         './data/': 's3://openmmlab/datasets/detection/',
+#         'data/': 's3://openmmlab/datasets/detection/'
+#     }))
+backend_args = None
+
+# In mstrain 3x config, img_scale=[(1333, 640), (1333, 800)],
+# multiscale_mode='range'
+train_pipeline = [
+    dict(type=LoadImageFromFile, backend_args=backend_args),
+    dict(type=LoadAnnotations, with_bbox=True),
+    dict(type=RandomResize, scale=[(1333, 640), (1333, 800)], keep_ratio=True),
+    dict(type=RandomFlip, prob=0.5),
+    dict(type=PackDetInputs)
+]
+test_pipeline = [
+    dict(type=LoadImageFromFile, backend_args=backend_args),
+    dict(type=Resize, scale=(1333, 800), keep_ratio=True),
+    dict(type=LoadAnnotations, with_bbox=True),
+    dict(
+        type=PackDetInputs,
+        meta_keys=('img_id', 'img_path', 'ori_shape', 'img_shape',
+                   'scale_factor'))
+]
+train_dataloader = dict(
+    batch_size=2,
+    num_workers=2,
+    persistent_workers=True,
+    pin_memory=True,
+    sampler=dict(type=DefaultSampler, shuffle=True),
+    batch_sampler=dict(type=AspectRatioBatchSampler),
+    dataset=dict(
+        type=RepeatDataset,
+        times=3,
+        dataset=dict(
+            type=dataset_type,
+            data_root=data_root,
+            ann_file='annotations/instances_train2017.json',
+            data_prefix=dict(img='train2017/'),
+            filter_cfg=dict(filter_empty_gt=True, min_size=32),
+            pipeline=train_pipeline,
+            backend_args=backend_args)))
+val_dataloader = dict(
+    batch_size=1,
+    num_workers=2,
+    persistent_workers=True,
+    drop_last=False,
+    sampler=dict(type=DefaultSampler, shuffle=False),
+    dataset=dict(
+        type=dataset_type,
+        data_root=data_root,
+        ann_file='annotations/instances_val2017.json',
+        data_prefix=dict(img='val2017/'),
+        test_mode=True,
+        pipeline=test_pipeline,
+        backend_args=backend_args))
+test_dataloader = val_dataloader
+
+val_evaluator = dict(
+    type=CocoMetric,
+    ann_file=data_root + 'annotations/instances_val2017.json',
+    metric='bbox',
+    backend_args=backend_args)
+test_evaluator = val_evaluator
+
+# training schedule for 3x with `RepeatDataset`
+train_cfg = dict(type=EpochBasedTrainLoop, max_iters=12, val_interval=1)
+val_cfg = dict(type=ValLoop)
+test_cfg = dict(type=TestLoop)
+
+# learning rate
+param_scheduler = [
+    dict(type=LinearLR, start_factor=0.001, by_epoch=False, begin=0, end=500),
+    dict(
+        type=MultiStepLR,
+        begin=0,
+        end=12,
+        by_epoch=False,
+        milestones=[9, 11],
+        gamma=0.1)
+]
+
+# optimizer
+optim_wrapper = dict(
+    type=OptimWrapper,
+    optimizer=dict(type=SGD, lr=0.02, momentum=0.9, weight_decay=0.0001))
+# Default setting for scaling LR automatically
+#   - `enable` means enable scaling LR automatically
+#       or not by default.
+#   - `base_batch_size` = (8 GPUs) x (2 samples per GPU).
+auto_scale_lr = dict(enable=False, base_batch_size=16)

mmdet/configs/common/ms_3x_coco_instance.py

+# Copyright (c) OpenMMLab. All rights reserved.
+
+# Please refer to https://mmengine.readthedocs.io/en/latest/advanced_tutorials/config.html#a-pure-python-style-configuration-file-beta for more details. # noqa
+# mmcv >= 2.0.1
+# mmengine >= 0.8.0
+
+from mmengine.config import read_base
+
+with read_base():
+    from .._base_.default_runtime import *
+
+from mmcv.transforms import RandomChoiceResize
+from mmengine.dataset import RepeatDataset
+from mmengine.dataset.sampler import DefaultSampler, InfiniteSampler
+from mmengine.optim import OptimWrapper
+from mmengine.optim.scheduler.lr_scheduler import LinearLR, MultiStepLR
+from mmengine.runner.loops import IterBasedTrainLoop, TestLoop, ValLoop
+from torch.optim import SGD
+
+from mmdet.datasets import AspectRatioBatchSampler, CocoDataset
+from mmdet.datasets.transforms.formatting import PackDetInputs
+from mmdet.datasets.transforms.loading import (FilterAnnotations,
+                                               LoadAnnotations,
+                                               LoadImageFromFile)
+from mmdet.datasets.transforms.transforms import (CachedMixUp, CachedMosaic,
+                                                  Pad, RandomCrop, RandomFlip,
+                                                  RandomResize, Resize)
+from mmdet.evaluation import CocoMetric
+
+# dataset settings
+dataset_type = CocoDataset
+data_root = 'data/coco/'
+
+# Example to use different file client
+# Method 1: simply set the data root and let the file I/O module
+# automatically infer from prefix (not support LMDB and Memcache yet)
+
+# data_root = 's3://openmmlab/datasets/detection/coco/'
+
+# Method 2: Use `backend_args`, `file_client_args` in versions before 3.0.0rc6
+# backend_args = dict(
+#     backend='petrel',
+#     path_mapping=dict({
+#         './data/': 's3://openmmlab/datasets/detection/',
+#         'data/': 's3://openmmlab/datasets/detection/'
+#     }))
+backend_args = None
+
+train_pipeline = [
+    dict(type=LoadImageFromFile, backend_args=backend_args),
+    dict(type=LoadAnnotations, with_bbox=True, with_mask=True),
+    dict(
+        type='RandomResize', scale=[(1333, 640), (1333, 800)],
+        keep_ratio=True),
+    dict(type=RandomFlip, prob=0.5),
+    dict(type=PackDetInputs)
+]
+test_pipeline = [
+    dict(type=LoadImageFromFile, backend_args=backend_args),
+    dict(type=Resize, scale=(1333, 800), keep_ratio=True),
+    dict(type=LoadAnnotations, with_bbox=True, with_mask=True),
+    dict(
+        type=PackDetInputs,
+        meta_keys=('img_id', 'img_path', 'ori_shape', 'img_shape',
+                   'scale_factor'))
+]
+train_dataloader.update(
+    dict(
+        batch_size=2,
+        num_workers=2,
+        persistent_workers=True,
+        sampler=dict(type=DefaultSampler, shuffle=True),
+        batch_sampler=dict(type=AspectRatioBatchSampler),
+        dataset=dict(
+            type=RepeatDataset,
+            times=3,
+            dataset=dict(
+                type=dataset_type,
+                data_root=data_root,
+                ann_file='annotations/instances_train2017.json',
+                data_prefix=dict(img='train2017/'),
+                filter_cfg=dict(filter_empty_gt=True, min_size=32),
+                pipeline=train_pipeline,
+                backend_args=backend_args))))
+val_dataloader.update(
+    dict(
+        batch_size=1,
+        num_workers=2,
+        persistent_workers=True,
+        drop_last=False,
+        sampler=dict(type=DefaultSampler, shuffle=False),
+        dataset=dict(
+            type=dataset_type,
+            data_root=data_root,
+            ann_file='annotations/instances_val2017.json',
+            data_prefix=dict(img='val2017/'),
+            test_mode=True,
+            pipeline=test_pipeline,
+            backend_args=backend_args)))
+test_dataloader = val_dataloader
+
+val_evaluator.update(
+    dict(
+        type=CocoMetric,
+        ann_file=data_root + 'annotations/instances_val2017.json',
+        metric='bbox',
+        backend_args=backend_args))
+test_evaluator = val_evaluator
+
+# training schedule for 3x with `RepeatDataset`
+train_cfg.update(dict(type=EpochBasedTrainLoop, max_epochs=12, val_interval=1))
+val_cfg.update(dict(type=ValLoop))
+test_cfg.update(dict(type=TestLoop))
+
+# learning rate
+param_scheduler = [
+    dict(type=LinearLR, start_factor=0.001, by_epoch=False, begin=0, end=500),
+    dict(
+        type=MultiStepLR,
+        begin=0,
+        end=12,
+        by_epoch=False,
+        milestones=[9, 11],
+        gamma=0.1)
+]
+
+# optimizer
+optim_wrapper.update(
+    dict(
+        type=OptimWrapper,
+        optimizer=dict(type=SGD, lr=0.02, momentum=0.9, weight_decay=0.0001)))
+# Default setting for scaling LR automatically
+#   - `enable` means enable scaling LR automatically
+#       or not by default.
+#   - `base_batch_size` = (8 GPUs) x (2 samples per GPU).
+auto_scale_lr.update(dict(enable=False, base_batch_size=16))

mmdet/configs/common/ms_90k_coco.py

+# Copyright (c) OpenMMLab. All rights reserved.
+
+# Please refer to https://mmengine.readthedocs.io/en/latest/advanced_tutorials/config.html#a-pure-python-style-configuration-file-beta for more details. # noqa
+# mmcv >= 2.0.1
+# mmengine >= 0.8.0
+
+from mmengine.config import read_base
+
+with read_base():
+    from .._base_.default_runtime import *
+
+from mmcv.transforms import RandomChoiceResize
+from mmengine.dataset import RepeatDataset
+from mmengine.dataset.sampler import DefaultSampler, InfiniteSampler
+from mmengine.optim import OptimWrapper
+from mmengine.optim.scheduler.lr_scheduler import LinearLR, MultiStepLR
+from mmengine.runner.loops import IterBasedTrainLoop, TestLoop, ValLoop
+from torch.optim import SGD
+
+from mmdet.datasets import AspectRatioBatchSampler, CocoDataset
+from mmdet.datasets.transforms.formatting import PackDetInputs
+from mmdet.datasets.transforms.loading import (FilterAnnotations,
+                                               LoadAnnotations,
+                                               LoadImageFromFile)
+from mmdet.datasets.transforms.transforms import (CachedMixUp, CachedMosaic,
+                                                  Pad, RandomCrop, RandomFlip,
+                                                  RandomResize, Resize)
+from mmdet.evaluation import CocoMetric
+
+# dataset settings
+dataset_type = CocoDataset
+data_root = 'data/coco/'
+# Example to use different file client
+# Method 1: simply set the data root and let the file I/O module
+# automatically infer from prefix (not support LMDB and Memcache yet)
+
+# data_root = 's3://openmmlab/datasets/detection/coco/'
+
+# Method 2: Use `backend_args`, `file_client_args` in versions before 3.0.0rc6
+# backend_args = dict(
+#     backend='petrel',
+#     path_mapping=dict({
+#         './data/': 's3://openmmlab/datasets/detection/',
+#         'data/': 's3://openmmlab/datasets/detection/'
+#     }))
+backend_args = None
+
+# Align with Detectron2
+backend = 'pillow'
+train_pipeline = [
+    dict(
+        type=LoadImageFromFile,
+        backend_args=backend_args,
+        imdecode_backend=backend),
+    dict(type=LoadAnnotations, with_bbox=True),
+    dict(
+        type=RandomChoiceResize,
+        scales=[(1333, 640), (1333, 672), (1333, 704), (1333, 736),
+                (1333, 768), (1333, 800)],
+        keep_ratio=True,
+        backend=backend),
+    dict(type=RandomFlip, prob=0.5),
+    dict(type=PackDetInputs)
+]
+test_pipeline = [
+    dict(
+        type=LoadImageFromFile,
+        backend_args=backend_args,
+        imdecode_backend=backend),
+    dict(type=Resize, scale=(1333, 800), keep_ratio=True, backend=backend),
+    dict(type=LoadAnnotations, with_bbox=True),
+    dict(
+        type=PackDetInputs,
+        meta_keys=('img_id', 'img_path', 'ori_shape', 'img_shape',
+                   'scale_factor'))
+]
+train_dataloader.update(
+    dict(
+        batch_size=2,
+        num_workers=2,
+        persistent_workers=True,
+        pin_memory=True,
+        sampler=dict(type=InfiniteSampler, shuffle=True),
+        batch_sampler=dict(type=AspectRatioBatchSampler),
+        dataset=dict(
+            type=dataset_type,
+            data_root=data_root,
+            ann_file='annotations/instances_train2017.json',
+            data_prefix=dict(img='train2017/'),
+            filter_cfg=dict(filter_empty_gt=True, min_size=32),
+            pipeline=train_pipeline,
+            backend_args=backend_args)))
+val_dataloader.update(
+    dict(
+        batch_size=1,
+        num_workers=2,
+        persistent_workers=True,
+        drop_last=False,
+        pin_memory=True,
+        sampler=dict(type=DefaultSampler, shuffle=False),
+        dataset=dict(
+            type=dataset_type,
+            data_root=data_root,
+            ann_file='annotations/instances_val2017.json',
+            data_prefix=dict(img='val2017/'),
+            test_mode=True,
+            pipeline=test_pipeline,
+            backend_args=backend_args)))
+test_dataloader = val_dataloader
+
+val_evaluator.update(
+    dict(
+        type=CocoMetric,
+        ann_file=data_root + 'annotations/instances_val2017.json',
+        metric='bbox',
+        format_only=False,
+        backend_args=backend_args))
+test_evaluator = val_evaluator
+
+# training schedule for 90k
+max_iter = 90000
+train_cfg.update(
+    dict(type=IterBasedTrainLoop, max_iters=max_iter, val_interval=10000))
+val_cfg.update(dict(type=ValLoop))
+test_cfg.update(dict(type=TestLoop))
+
+# learning rate
+param_scheduler = [
+    dict(type=LinearLR, start_factor=0.001, by_epoch=False, begin=0, end=1000),
+    dict(
+        type=MultiStepLR,
+        begin=0,
+        end=max_iter,
+        by_epoch=False,
+        milestones=[60000, 80000],
+        gamma=0.1)
+]
+
+# optimizer
+optim_wrapper.update(
+    dict(
+        type=OptimWrapper,
+        optimizer=dict(type=SGD, lr=0.02, momentum=0.9, weight_decay=0.0001)))
+# Default setting for scaling LR automatically
+#   - `enable` means enable scaling LR automatically
+#       or not by default.
+#   - `base_batch_size` = (8 GPUs) x (2 samples per GPU).
+auto_scale_lr.update(dict(enable=False, base_batch_size=16))
+
+default_hooks.update(dict(checkpoint=dict(by_epoch=False, interval=10000)))
+log_processor.update(dict(by_epoch=False))

mmdet/configs/common/ms_poly_3x_coco_instance.py

+# Copyright (c) OpenMMLab. All rights reserved.
+
+# Please refer to https://mmengine.readthedocs.io/en/latest/advanced_tutorials/config.html#a-pure-python-style-configuration-file-beta for more details. # noqa
+# mmcv >= 2.0.1
+# mmengine >= 0.8.0
+
+from mmengine.config import read_base
+
+with read_base():
+    from .._base_.default_runtime import *
+
+from mmcv.transforms import RandomChoiceResize
+from mmengine.dataset import RepeatDataset
+from mmengine.dataset.sampler import DefaultSampler, InfiniteSampler
+from mmengine.optim import OptimWrapper
+from mmengine.optim.scheduler.lr_scheduler import LinearLR, MultiStepLR
+from mmengine.runner.loops import IterBasedTrainLoop, TestLoop, ValLoop
+from torch.optim import SGD
+
+from mmdet.datasets import AspectRatioBatchSampler, CocoDataset
+from mmdet.datasets.transforms.formatting import PackDetInputs
+from mmdet.datasets.transforms.loading import (FilterAnnotations,
+                                               LoadAnnotations,
+                                               LoadImageFromFile)
+from mmdet.datasets.transforms.transforms import (CachedMixUp, CachedMosaic,
+                                                  Pad, RandomCrop, RandomFlip,
+                                                  RandomResize, Resize)
+from mmdet.evaluation import CocoMetric
+
+# dataset settings
+dataset_type = CocoDataset
+data_root = 'data/coco/'
+# Example to use different file client
+# Method 1: simply set the data root and let the file I/O module
+# automatically infer from prefix (not support LMDB and Memcache yet)
+
+# data_root = 's3://openmmlab/datasets/detection/coco/'
+
+# Method 2: Use `backend_args`, `file_client_args` in versions before 3.0.0rc6
+# backend_args = dict(
+#     backend='petrel',
+#     path_mapping=dict({
+#         './data/': 's3://openmmlab/datasets/detection/',
+#         'data/': 's3://openmmlab/datasets/detection/'
+#     }))
+backend_args = None
+
+# In mstrain 3x config, img_scale=[(1333, 640), (1333, 800)],
+# multiscale_mode='range'
+train_pipeline = [
+    dict(type=LoadImageFromFile, backend_args=backend_args),
+    dict(
+        type=LoadAnnotations, with_bbox=True, with_mask=True, poly2mask=False),
+    dict(
+        type='RandomResize', scale=[(1333, 640), (1333, 800)],
+        keep_ratio=True),
+    dict(type=RandomFlip, prob=0.5),
+    dict(type=PackDetInputs)
+]
+test_pipeline = [
+    dict(type=LoadImageFromFile, backend_args=backend_args),
+    dict(type=Resize, scale=(1333, 800), keep_ratio=True),
+    dict(
+        type=LoadAnnotations, with_bbox=True, with_mask=True, poly2mask=False),
+    dict(
+        type=PackDetInputs,
+        meta_keys=('img_id', 'img_path', 'ori_shape', 'img_shape',
+                   'scale_factor'))
+]
+train_dataloader.update(
+    dict(
+        batch_size=2,
+        num_workers=2,
+        persistent_workers=True,
+        pin_memory=True,
+        sampler=dict(type=DefaultSampler, shuffle=True),
+        batch_sampler=dict(type=AspectRatioBatchSampler),
+        dataset=dict(
+            type=RepeatDataset,
+            data_root=data_root,
+            ann_file='annotations/instances_train2017.json',
+            data_prefix=dict(img='train2017/'),
+            filter_cfg=dict(filter_empty_gt=True, min_size=32),
+            pipeline=train_pipeline,
+            backend_args=backend_args)))
+val_dataloader.update(
+    dict(
+        batch_size=2,
+        num_workers=2,
+        persistent_workers=True,
+        drop_last=False,
+        pin_memory=True,
+        sampler=dict(type=DefaultSampler, shuffle=False),
+        dataset=dict(
+            type=dataset_type,
+            data_root=data_root,
+            ann_file='annotations/instances_val2017.json',
+            data_prefix=dict(img='val2017/'),
+            test_mode=True,
+            pipeline=test_pipeline,
+            backend_args=backend_args)))
+test_dataloader = val_dataloader
+
+val_evaluator.update(
+    dict(
+        type=CocoMetric,
+        ann_file=data_root + 'annotations/instances_val2017.json',
+        metric=['bbox', 'segm'],
+        backend_args=backend_args))
+test_evaluator = val_evaluator
+
+# training schedule for 3x with `RepeatDataset`
+train_cfg.update(dict(type=EpochBasedTrainLoop, max_iters=12, val_interval=1))
+val_cfg.update(dict(type=ValLoop))
+test_cfg.update(dict(type=TestLoop))
+
+# learning rate
+param_scheduler = [
+    dict(type=LinearLR, start_factor=0.001, by_epoch=False, begin=0, end=500),
+    dict(
+        type=MultiStepLR,
+        begin=0,
+        end=12,
+        by_epoch=False,
+        milestones=[9, 11],
+        gamma=0.1)
+]
+
+# optimizer
+optim_wrapper.update(
+    dict(
+        type=OptimWrapper,
+        optimizer=dict(type=SGD, lr=0.02, momentum=0.9, weight_decay=0.0001)))
+# Default setting for scaling LR automatically
+#   - `enable` means enable scaling LR automatically
+#       or not by default.
+#   - `base_batch_size` = (8 GPUs) x (2 samples per GPU).
+auto_scale_lr.update(dict(enable=False, base_batch_size=16))

mmdet/configs/common/ms_poly_90k_coco_instance.py

+# Copyright (c) OpenMMLab. All rights reserved.
+
+# Please refer to https://mmengine.readthedocs.io/en/latest/advanced_tutorials/config.html#a-pure-python-style-configuration-file-beta for more details. # noqa
+# mmcv >= 2.0.1
+# mmengine >= 0.8.0
+
+from mmengine.config import read_base
+
+with read_base():
+    from .._base_.default_runtime import *
+
+from mmcv.transforms import RandomChoiceResize
+from mmengine.dataset import RepeatDataset
+from mmengine.dataset.sampler import DefaultSampler, InfiniteSampler
+from mmengine.optim import OptimWrapper
+from mmengine.optim.scheduler.lr_scheduler import LinearLR, MultiStepLR
+from mmengine.runner.loops import IterBasedTrainLoop, TestLoop, ValLoop
+from torch.optim import SGD
+
+from mmdet.datasets import AspectRatioBatchSampler, CocoDataset
+from mmdet.datasets.transforms.formatting import PackDetInputs
+from mmdet.datasets.transforms.loading import (FilterAnnotations,
+                                               LoadAnnotations,
+                                               LoadImageFromFile)
+from mmdet.datasets.transforms.transforms import (CachedMixUp, CachedMosaic,
+                                                  Pad, RandomCrop, RandomFlip,
+                                                  RandomResize, Resize)
+from mmdet.evaluation import CocoMetric
+
+# dataset settings
+dataset_type = CocoDataset
+data_root = 'data/coco/'
+# Example to use different file client
+# Method 1: simply set the data root and let the file I/O module
+# automatically infer from prefix (not support LMDB and Memcache yet)
+
+# data_root = 's3://openmmlab/datasets/detection/coco/'
+
+# Method 2: Use `backend_args`, `file_client_args` in versions before 3.0.0rc6
+# backend_args = dict(
+#     backend='petrel',
+#     path_mapping=dict({
+#         './data/': 's3://openmmlab/datasets/detection/',
+#         'data/': 's3://openmmlab/datasets/detection/'
+#     }))
+backend_args = None
+
+# Align with Detectron2
+backend = 'pillow'
+train_pipeline = [
+    dict(
+        type=LoadImageFromFile,
+        backend_args=backend_args,
+        imdecode_backend=backend),
+    dict(
+        type=LoadAnnotations, with_bbox=True, with_mask=True, poly2mask=False),
+    dict(
+        type=RandomChoiceResize,
+        scales=[(1333, 640), (1333, 672), (1333, 704), (1333, 736),
+                (1333, 768), (1333, 800)],
+        keep_ratio=True,
+        backend=backend),
+    dict(type=RandomFlip, prob=0.5),
+    dict(type=PackDetInputs)
+]
+test_pipeline = [
+    dict(
+        type=LoadImageFromFile,
+        backend_args=backend_args,
+        imdecode_backend=backend),
+    dict(type=Resize, scale=(1333, 800), keep_ratio=True, backend=backend),
+    dict(
+        type=LoadAnnotations, with_bbox=True, with_mask=True, poly2mask=False),
+    dict(
+        type=PackDetInputs,
+        meta_keys=('img_id', 'img_path', 'ori_shape', 'img_shape',
+                   'scale_factor'))
+]
+train_dataloader.update(
+    dict(
+        batch_size=2,
+        num_workers=2,
+        persistent_workers=True,
+        pin_memory=True,
+        sampler=dict(type=InfiniteSampler, shuffle=True),
+        batch_sampler=dict(type=AspectRatioBatchSampler),
+        dataset=dict(
+            type=dataset_type,
+            data_root=data_root,
+            ann_file='annotations/instances_train2017.json',
+            data_prefix=dict(img='train2017/'),
+            filter_cfg=dict(filter_empty_gt=True, min_size=32),
+            pipeline=train_pipeline,
+            backend_args=backend_args)))
+val_dataloader.update(
+    dict(
+        batch_size=1,
+        num_workers=2,
+        persistent_workers=True,
+        drop_last=False,
+        pin_memory=True,
+        sampler=dict(type=DefaultSampler, shuffle=False),
+        dataset=dict(
+            type=dataset_type,
+            data_root=data_root,
+            ann_file='annotations/instances_val2017.json',
+            data_prefix=dict(img='val2017/'),
+            test_mode=True,
+            pipeline=test_pipeline,
+            backend_args=backend_args)))
+test_dataloader = val_dataloader
+
+val_evaluator.update(
+    dict(
+        type=CocoMetric,
+        ann_file=data_root + 'annotations/instances_val2017.json',
+        metric=['bbox', 'segm'],
+        format_only=False,
+        backend_args=backend_args))
+test_evaluator = val_evaluator
+
+# training schedule for 90k
+max_iter = 90000
+train_cfg.update(
+    dict(type=IterBasedTrainLoop, max_iters=max_iter, val_interval=10000))
+val_cfg.update(dict(type=ValLoop))
+test_cfg.update(dict(type=TestLoop))
+
+# learning rate
+param_scheduler = [
+    dict(type=LinearLR, start_factor=0.001, by_epoch=False, begin=0, end=1000),
+    dict(
+        type=MultiStepLR,
+        begin=0,
+        end=max_iter,
+        by_epoch=False,
+        milestones=[60000, 80000],
+        gamma=0.1)
+]
+
+# optimizer
+optim_wrapper.update(
+    dict(
+        type=OptimWrapper,
+        optimizer=dict(type=SGD, lr=0.02, momentum=0.9, weight_decay=0.0001)))
+# Default setting for scaling LR automatically
+#   - `enable` means enable scaling LR automatically
+#       or not by default.
+#   - `base_batch_size` = (8 GPUs) x (2 samples per GPU).
+auto_scale_lr.update(dict(enable=False, base_batch_size=16))
+
+default_hooks.update(dict(checkpoint=dict(by_epoch=False, interval=10000)))
+log_processor.update(dict(by_epoch=False))

mmdet/configs/common/ssj_270_coco_instance.py

+# Copyright (c) OpenMMLab. All rights reserved.
+
+# Please refer to https://mmengine.readthedocs.io/en/latest/advanced_tutorials/config.html#a-pure-python-style-configuration-file-beta for more details. # noqa
+# mmcv >= 2.0.1
+# mmengine >= 0.8.0
+
+from mmengine.config import read_base
+
+with read_base():
+    from .._base_.default_runtime import *
+
+from mmcv.transforms import RandomChoiceResize
+from mmengine.dataset import RepeatDataset
+from mmengine.dataset.sampler import DefaultSampler, InfiniteSampler
+from mmengine.optim import OptimWrapper
+from mmengine.optim.scheduler.lr_scheduler import LinearLR, MultiStepLR
+from mmengine.runner.loops import IterBasedTrainLoop, TestLoop, ValLoop
+from torch.optim import SGD
+
+from mmdet.datasets import AspectRatioBatchSampler, CocoDataset
+from mmdet.datasets.transforms.formatting import PackDetInputs
+from mmdet.datasets.transforms.loading import (FilterAnnotations,
+                                               LoadAnnotations,
+                                               LoadImageFromFile)
+from mmdet.datasets.transforms.transforms import (CachedMixUp, CachedMosaic,
+                                                  Pad, RandomCrop, RandomFlip,
+                                                  RandomResize, Resize)
+from mmdet.evaluation import CocoMetric
+
+# dataset settings
+dataset_type = CocoDataset
+data_root = 'data/coco/'
+# Example to use different file client
+# Method 1: simply set the data root and let the file I/O module
+# automatically infer from prefix (not support LMDB and Memcache yet)
+
+# data_root = 's3://openmmlab/datasets/detection/coco/'
+
+# Method 2: Use `backend_args`, `file_client_args` in versions before 3.0.0rc6
+# backend_args = dict(
+#     backend='petrel',
+#     path_mapping=dict({
+#         './data/': 's3://openmmlab/datasets/detection/',
+#         'data/': 's3://openmmlab/datasets/detection/'
+#     }))
+backend_args = None
+
+# Standard Scale Jittering (SSJ) resizes and crops an image
+# with a resize range of 0.8 to 1.25 of the original image size.
+train_pipeline = [
+    dict(type=LoadImageFromFile, backend_args=backend_args),
+    dict(type=LoadAnnotations, with_bbox=True, with_mask=True),
+    dict(
+        type=RandomResize,
+        scale=image_size,
+        ratio_range=(0.8, 1.25),
+        keep_ratio=True),
+    dict(
+        type='RandomCrop',
+        crop_type='absolute_range',
+        crop_size=image_size,
+        recompute_bbox=True,
+        allow_negative_crop=True),
+    dict(type='FilterAnnotations', min_gt_bbox_wh=(1e-2, 1e-2)),
+    dict(type=RandomFlip, prob=0.5),
+    dict(type=PackDetInputs)
+]
+test_pipeline = [
+    dict(type=LoadImageFromFile, backend_args=backend_args),
+    dict(type=Resize, scale=(1333, 800), keep_ratio=True),
+    dict(type=LoadAnnotations, with_bbox=True, with_mask=True),
+    dict(
+        type=PackDetInputs,
+        meta_keys=('img_id', 'img_path', 'ori_shape', 'img_shape',
+                   'scale_factor'))
+]
+train_dataloader.update(
+    dict(
+        batch_size=2,
+        num_workers=2,
+        persistent_workers=True,
+        sampler=dict(type=InfiniteSampler),
+        dataset=dict(
+            type=dataset_type,
+            data_root=data_root,
+            ann_file='annotations/instances_train2017.json',
+            data_prefix=dict(img='train2017/'),
+            filter_cfg=dict(filter_empty_gt=True, min_size=32),
+            pipeline=train_pipeline,
+            backend_args=backend_args)))
+val_dataloader.update(
+    dict(
+        batch_size=1,
+        num_workers=2,
+        persistent_workers=True,
+        drop_last=False,
+        sampler=dict(type=DefaultSampler, shuffle=False),
+        dataset=dict(
+            type=dataset_type,
+            data_root=data_root,
+            ann_file='annotations/instances_val2017.json',
+            data_prefix=dict(img='val2017/'),
+            test_mode=True,
+            pipeline=test_pipeline,
+            backend_args=backend_args)))
+test_dataloader = val_dataloader
+
+val_evaluator.update(
+    dict(
+        type=CocoMetric,
+        ann_file=data_root + 'annotations/instances_val2017.json',
+        metric=['bbox', 'segm'],
+        format_only=False,
+        backend_args=backend_args))
+test_evaluator = val_evaluator
+
+val_evaluator = dict(
+    type=CocoMetric,
+    ann_file=data_root + 'annotations/instances_val2017.json',
+    metric=['bbox', 'segm'],
+    format_only=False,
+    backend_args=backend_args)
+test_evaluator = val_evaluator
+
+# The model is trained by 270k iterations with batch_size 64,
+# which is roughly equivalent to 144 epochs.
+
+max_iter = 270000
+train_cfg.update(
+    dict(type=IterBasedTrainLoop, max_iters=max_iter, val_interval=10000))
+val_cfg.update(dict(type=ValLoop))
+test_cfg.update(dict(type=TestLoop))
+
+# learning rate
+param_scheduler = [
+    dict(type=LinearLR, start_factor=0.001, by_epoch=False, begin=0, end=1000),
+    dict(
+        type=MultiStepLR,
+        begin=0,
+        end=max_iter,
+        by_epoch=False,
+        milestones=[243000, 256500, 263250],
+        gamma=0.1)
+]
+
+# optimizer
+optim_wrapper.update(
+    dict(
+        type=OptimWrapper,
+        optimizer=dict(type=SGD, lr=0.1, momentum=0.9, weight_decay=0.00004)))
+# Default setting for scaling LR automatically
+#   - `enable` means enable scaling LR automatically
+#       or not by default.
+#   - `base_batch_size` = (8 GPUs) x (2 samples per GPU).
+auto_scale_lr.update(dict(base_batch_size=64))
+
+default_hooks.update(dict(checkpoint=dict(by_epoch=False, interval=10000)))
+log_processor.update(dict(by_epoch=False))

mmdet/configs/common/ssj_scp_270k_coco_instance.py

+# Copyright (c) OpenMMLab. All rights reserved.
+
+# Please refer to https://mmengine.readthedocs.io/en/latest/advanced_tutorials/config.html#a-pure-python-style-configuration-file-beta for more details. # noqa
+# mmcv >= 2.0.1
+# mmengine >= 0.8.0
+
+from mmengine.config import read_base
+
+with read_base():
+    from .ssj_270_coco_instance import *
+
+from mmdet.datasets import MultiImageMixDataset
+from mmdet.datasets.transforms import CopyPaste
+
+# dataset settings
+dataset_type = CocoDataset
+data_root = 'data/coco/'
+image_size = (1024, 1024)
+# Example to use different file client
+# Method 1: simply set the data root and let the file I/O module
+# automatically infer from prefix (not support LMDB and Memcache yet)
+
+# data_root = 's3://openmmlab/datasets/detection/coco/'
+
+# Method 2: Use `backend_args`, `file_client_args` in versions before 3.0.0rc6
+# backend_args = dict(
+#     backend='petrel',
+#     path_mapping=dict({
+#         './data/': 's3://openmmlab/datasets/detection/',
+#         'data/': 's3://openmmlab/datasets/detection/'
+#     }))
+backend_args = None
+
+# Standard Scale Jittering (SSJ) resizes and crops an image
+# with a resize range of 0.8 to 1.25 of the original image size.
+load_pipeline = [
+    dict(type=LoadImageFromFile, backend_args=backend_args),
+    dict(type=LoadAnnotations, with_bbox=True, with_mask=True),
+    dict(
+        type=RandomResize,
+        scale=image_size,
+        ratio_range=(0.8, 1.25),
+        keep_ratio=True),
+    dict(
+        type='RandomCrop',
+        crop_type='absolute_range',
+        crop_size=image_size,
+        recompute_bbox=True,
+        allow_negative_crop=True),
+    dict(type='FilterAnnotations', min_gt_bbox_wh=(1e-2, 1e-2)),
+    dict(type=RandomFlip, prob=0.5),
+    dict(type=Pad, size=image_size),
+]
+train_pipeline = [
+    dict(type=CopyPaste, max_num_pasted=100),
+    dict(type=PackDetInputs)
+]
+
+train_dataloader.update(
+    dict(
+        type=MultiImageMixDataset,
+        dataset=dict(
+            type=dataset_type,
+            data_root=data_root,
+            ann_file='annotations/instances_train2017.json',
+            data_prefix=dict(img='train2017/'),
+            filter_cfg=dict(filter_empty_gt=True, min_size=32),
+            pipeline=load_pipeline,
+            backend_args=backend_args),
+        pipeline=train_pipeline))

mmdet/configs/deformable_detr/deformable_detr_r50_16xb2_50e_coco.py

+# Copyright (c) OpenMMLab. All rights reserved.
+
+# Please refer to https://mmengine.readthedocs.io/en/latest/advanced_tutorials/config.html#a-pure-python-style-configuration-file-beta for more details. # noqa
+# mmcv >= 2.0.1
+# mmengine >= 0.8.0
+
+from mmengine.config import read_base
+
+with read_base():
+    from .._base_.datasets.coco_detection import *
+    from .._base_.default_runtime import *
+
+from mmcv.transforms import LoadImageFromFile, RandomChoice, RandomChoiceResize
+from mmengine.optim.optimizer import OptimWrapper
+from mmengine.optim.scheduler import MultiStepLR
+from mmengine.runner.loops import EpochBasedTrainLoop, TestLoop, ValLoop
+from torch.optim.adamw import AdamW
+
+from mmdet.datasets.transforms import (LoadAnnotations, PackDetInputs,
+                                       RandomCrop, RandomFlip, Resize)
+from mmdet.models.backbones import ResNet
+from mmdet.models.data_preprocessors import DetDataPreprocessor
+from mmdet.models.dense_heads import DeformableDETRHead
+from mmdet.models.detectors import DeformableDETR
+from mmdet.models.losses import FocalLoss, GIoULoss, L1Loss
+from mmdet.models.necks import ChannelMapper
+from mmdet.models.task_modules import (BBoxL1Cost, FocalLossCost,
+                                       HungarianAssigner, IoUCost)
+
+model = dict(
+    type=DeformableDETR,
+    num_queries=300,
+    num_feature_levels=4,
+    with_box_refine=False,
+    as_two_stage=False,
+    data_preprocessor=dict(
+        type=DetDataPreprocessor,
+        mean=[123.675, 116.28, 103.53],
+        std=[58.395, 57.12, 57.375],
+        bgr_to_rgb=True,
+        pad_size_divisor=1),
+    backbone=dict(
+        type=ResNet,
+        depth=50,
+        num_stages=4,
+        out_indices=(1, 2, 3),
+        frozen_stages=1,
+        norm_cfg=dict(type='BN', requires_grad=False),
+        norm_eval=True,
+        style='pytorch',
+        init_cfg=dict(type='Pretrained', checkpoint='torchvision://resnet50')),
+    neck=dict(
+        type=ChannelMapper,
+        in_channels=[512, 1024, 2048],
+        kernel_size=1,
+        out_channels=256,
+        act_cfg=None,
+        norm_cfg=dict(type='GN', num_groups=32),
+        num_outs=4),
+    encoder=dict(  # DeformableDetrTransformerEncoder
+        num_layers=6,
+        layer_cfg=dict(  # DeformableDetrTransformerEncoderLayer
+            self_attn_cfg=dict(  # MultiScaleDeformableAttention
+                embed_dims=256,
+                batch_first=True),
+            ffn_cfg=dict(
+                embed_dims=256, feedforward_channels=1024, ffn_drop=0.1))),
+    decoder=dict(  # DeformableDetrTransformerDecoder
+        num_layers=6,
+        return_intermediate=True,
+        layer_cfg=dict(  # DeformableDetrTransformerDecoderLayer
+            self_attn_cfg=dict(  # MultiheadAttention
+                embed_dims=256,
+                num_heads=8,
+                dropout=0.1,
+                batch_first=True),
+            cross_attn_cfg=dict(  # MultiScaleDeformableAttention
+                embed_dims=256,
+                batch_first=True),
+            ffn_cfg=dict(
+                embed_dims=256, feedforward_channels=1024, ffn_drop=0.1)),
+        post_norm_cfg=None),
+    positional_encoding=dict(num_feats=128, normalize=True, offset=-0.5),
+    bbox_head=dict(
+        type=DeformableDETRHead,
+        num_classes=80,
+        sync_cls_avg_factor=True,
+        loss_cls=dict(
+            type=FocalLoss,
+            use_sigmoid=True,
+            gamma=2.0,
+            alpha=0.25,
+            loss_weight=2.0),
+        loss_bbox=dict(type=L1Loss, loss_weight=5.0),
+        loss_iou=dict(type=GIoULoss, loss_weight=2.0)),
+    # training and testing settings
+    train_cfg=dict(
+        assigner=dict(
+            type=HungarianAssigner,
+            match_costs=[
+                dict(type=FocalLossCost, weight=2.0),
+                dict(type=BBoxL1Cost, weight=5.0, box_format='xywh'),
+                dict(type=IoUCost, iou_mode='giou', weight=2.0)
+            ])),
+    test_cfg=dict(max_per_img=100))
+
+# train_pipeline, NOTE the img_scale and the Pad's size_divisor is different
+# from the default setting in mmdet.
+train_pipeline = [
+    dict(type=LoadImageFromFile, backend_args=backend_args),
+    dict(type=LoadAnnotations, with_bbox=True),
+    dict(type=RandomFlip, prob=0.5),
+    dict(
+        type=RandomChoice,
+        transforms=[
+            [
+                dict(
+                    type=RandomChoiceResize,
+                    scales=[(480, 1333), (512, 1333), (544, 1333), (576, 1333),
+                            (608, 1333), (640, 1333), (672, 1333), (704, 1333),
+                            (736, 1333), (768, 1333), (800, 1333)],
+                    resize_type=Resize,
+                    keep_ratio=True)
+            ],
+            [
+                dict(
+                    type=RandomChoiceResize,
+                    # The radio of all image in train dataset < 7
+                    # follow the original implement
+                    scales=[(400, 4200), (500, 4200), (600, 4200)],
+                    resize_type=Resize,
+                    keep_ratio=True),
+                dict(
+                    type=RandomCrop,
+                    crop_type='absolute_range',
+                    crop_size=(384, 600),
+                    allow_negative_crop=True),
+                dict(
+                    type=RandomChoiceResize,
+                    scales=[(480, 1333), (512, 1333), (544, 1333), (576, 1333),
+                            (608, 1333), (640, 1333), (672, 1333), (704, 1333),
+                            (736, 1333), (768, 1333), (800, 1333)],
+                    resize_type=Resize,
+                    keep_ratio=True)
+            ]
+        ]),
+    dict(type=PackDetInputs)
+]
+train_dataloader.update(
+    dict(
+        dataset=dict(
+            filter_cfg=dict(filter_empty_gt=False), pipeline=train_pipeline)))
+
+# optimizer
+optim_wrapper = dict(
+    type=OptimWrapper,
+    optimizer=dict(type=AdamW, lr=0.0002, weight_decay=0.0001),
+    clip_grad=dict(max_norm=0.1, norm_type=2),
+    paramwise_cfg=dict(
+        custom_keys={
+            'backbone': dict(lr_mult=0.1),
+            'sampling_offsets': dict(lr_mult=0.1),
+            'reference_points': dict(lr_mult=0.1)
+        }))
+
+# learning policy
+max_epochs = 50
+train_cfg = dict(
+    type=EpochBasedTrainLoop, max_epochs=max_epochs, val_interval=1)
+val_cfg = dict(type=ValLoop)
+test_cfg = dict(type=TestLoop)
+
+param_scheduler = [
+    dict(
+        type=MultiStepLR,
+        begin=0,
+        end=max_epochs,
+        by_epoch=True,
+        milestones=[40],
+        gamma=0.1)
+]
+
+# NOTE: `auto_scale_lr` is for automatically scaling LR,
+# USER SHOULD NOT CHANGE ITS VALUES.
+# base_batch_size = (16 GPUs) x (2 samples per GPU)
+auto_scale_lr = dict(base_batch_size=32)

mmdet/configs/deformable_detr/deformable_detr_refine_r50_16xb2_50e_coco.py

+# Copyright (c) OpenMMLab. All rights reserved.
+
+# Please refer to https://mmengine.readthedocs.io/en/latest/advanced_tutorials/config.html#a-pure-python-style-configuration-file-beta for more details. # noqa
+# mmcv >= 2.0.1
+# mmengine >= 0.8.0
+
+from mmengine.config import read_base
+
+with read_base():
+    from .deformable_detr_r50_16xb2_50e_coco import *
+
+model.update(dict(with_box_refine=True))

mmdet/configs/deformable_detr/deformable_detr_refine_twostage_r50_16xb2_50e_coco.py

+# Copyright (c) OpenMMLab. All rights reserved.
+
+# Please refer to https://mmengine.readthedocs.io/en/latest/advanced_tutorials/config.html#a-pure-python-style-configuration-file-beta for more details. # noqa
+# mmcv >= 2.0.1
+# mmengine >= 0.8.0
+
+from mmengine.config import read_base
+
+with read_base():
+    from .deformable_detr_refine_r50_16xb2_50e_coco import *
+
+model.update(dict(as_two_stage=True))