_base_ = [
    '../_base_/datasets/coco_detection.py', '../_base_/default_runtime.py'
]
model = dict(
    type='DeformableDETR',
    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),
    bbox_head=dict(
        type='DeformableDETRHead',
        with_box_refine=True,
        as_two_stage=True,
        num_query=300,
        num_classes=11,
        in_channels=2048,
        sync_cls_avg_factor=True,
        transformer=dict(
            type='DeformableDetrTransformer',
            encoder=dict(
                type='DetrTransformerEncoder',
                num_layers=6,
                transformerlayers=dict(
                    type='BaseTransformerLayer',
                    attn_cfgs=dict(
                        type='MultiScaleDeformableAttention', embed_dims=256),
                    feedforward_channels=1024,
                    ffn_dropout=0.1,
                    operation_order=('self_attn', 'norm', 'ffn', 'norm'))),
            decoder=dict(
                type='DeformableDetrTransformerDecoder',
                num_layers=6,
                return_intermediate=True,
                transformerlayers=dict(
                    type='DetrTransformerDecoderLayer',
                    attn_cfgs=[
                        dict(
                            type='MultiheadAttention',
                            embed_dims=256,
                            num_heads=8,
                            dropout=0.1),
                        dict(
                            type='MultiScaleDeformableAttention',
                            embed_dims=256)
                    ],
                    feedforward_channels=1024,
                    ffn_dropout=0.1,
                    operation_order=('self_attn', 'norm', 'cross_attn', 'norm',
                                     'ffn', 'norm')))),
        positional_encoding=dict(
            type='SinePositionalEncoding',
            num_feats=128,
            normalize=True,
            offset=-0.5),
        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',
            cls_cost=dict(type='FocalLossCost', weight=2.0),
            reg_cost=dict(type='BBoxL1Cost', weight=5.0, box_format='xywh'),
            iou_cost=dict(type='IoUCost', iou_mode='giou', weight=2.0))),
    test_cfg=dict(max_per_img=100))
img_norm_cfg = dict(
    mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
# 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'),
    dict(type='LoadAnnotations', with_bbox=True),
    dict(
        type='Resize',
        img_scale=[(400, 300), (500, 400)],
        multiscale_mode='value',
        keep_ratio=True),
    dict(type='RandomFlip', flip_ratio=[0.2,0.2,0.2], direction=['horizontal', 'vertical', 'diagonal']),
    dict(type='BrightnessTransform', level=5, prob=0.5),
    dict(type='ContrastTransform', level=5, prob=0.5),
    dict(type='RandomShift', shift_ratio=0.5),
    dict(type='MinIoURandomCrop', min_ious=(0.5, 0.7, 0.9), min_crop_size=0.8),
    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=(400, 300),
        flip=False,
        transforms=[
            dict(type='Resize', keep_ratio=True),
            dict(type='RandomFlip'),
            dict(type='Normalize', **img_norm_cfg),
            dict(type='Pad', size_divisor=32),
            dict(type='ImageToTensor', keys=['img']),
            dict(type='Collect', keys=['img']),
        ])
]

dataset_type = 'CocoDataset'
classes = ('yiwei','loujian','celi','libei','fantie','lianxi','duojian','shunjian','shaoxi','jiahan','yiwu')

data = dict(
    samples_per_gpu=16,
    workers_per_gpu=8,
    train=dict(
        type=dataset_type,
        img_prefix='/home/shanwei-luo/userdata/datasets/dsxw_dataset_v4/dsxw_train/images/',
        classes=classes,
        ann_file='/home/shanwei-luo/userdata/datasets/dsxw_dataset_v4/dsxw_train/annotations/train.json',
        pipeline=train_pipeline),
    val=dict(
        type=dataset_type,
        img_prefix='/home/shanwei-luo/userdata/datasets/dsxw_dataset_v4/dsxw_test/images/',
        classes=classes,
        ann_file='/home/shanwei-luo/userdata/datasets/dsxw_dataset_v4/dsxw_test/annotations/test.json',
        pipeline=test_pipeline),
    test=dict(
        type=dataset_type,
        img_prefix='/home/shanwei-luo/userdata/datasets/dsxw_dataset_v4/dsxw_test/images/',
        classes=classes,
        ann_file='/home/shanwei-luo/userdata/datasets/dsxw_dataset_v4/dsxw_test/annotations/test.json',
        pipeline=test_pipeline))
# optimizer
optimizer = dict(
    type='AdamW',
    lr=2e-4,
    weight_decay=0.0001,
    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)
        }))
optimizer_config = dict(grad_clip=dict(max_norm=0.1, norm_type=2))
# learning policy
lr_config = dict(policy='step', step=[40])
runner = dict(type='EpochBasedRunner', max_epochs=60)

evaluation = dict(interval=5, metric='bbox')