atss_swin-l-p4-w12_fpn_dyhead_ms-2x_coco.py

_base_ = [
    '../_base_/datasets/coco_detection.py',
    '../_base_/schedules/schedule_1x.py', '../_base_/default_runtime.py'
]

pretrained = 'https://github.com/SwinTransformer/storage/releases/download/v1.0.0/swin_large_patch4_window12_384_22k.pth'  # noqa
model = dict(
    type='ATSS',
    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=128),
    backbone=dict(
        type='SwinTransformer',
        pretrain_img_size=384,
        embed_dims=192,
        depths=[2, 2, 18, 2],
        num_heads=[6, 12, 24, 48],
        window_size=12,
        mlp_ratio=4,
        qkv_bias=True,
        qk_scale=None,
        drop_rate=0.,
        attn_drop_rate=0.,
        drop_path_rate=0.2,
        patch_norm=True,
        out_indices=(1, 2, 3),
        # Please only add indices that would be used
        # in FPN, otherwise some parameter will not be used
        with_cp=False,
        convert_weights=True,
        init_cfg=dict(type='Pretrained', checkpoint=pretrained)),
    neck=[
        dict(
            type='FPN',
            in_channels=[384, 768, 1536],
            out_channels=256,
            start_level=0,
            add_extra_convs='on_output',
            num_outs=5),
        dict(
            type='DyHead',
            in_channels=256,
            out_channels=256,
            num_blocks=6,
            # disable zero_init_offset to follow official implementation
            zero_init_offset=False)
    ],
    bbox_head=dict(
        type='ATSSHead',
        num_classes=80,
        in_channels=256,
        pred_kernel_size=1,  # follow DyHead official implementation
        stacked_convs=0,
        feat_channels=256,
        anchor_generator=dict(
            type='AnchorGenerator',
            ratios=[1.0],
            octave_base_scale=8,
            scales_per_octave=1,
            strides=[8, 16, 32, 64, 128],
            center_offset=0.5),  # follow DyHead official implementation
        bbox_coder=dict(
            type='DeltaXYWHBBoxCoder',
            target_means=[.0, .0, .0, .0],
            target_stds=[0.1, 0.1, 0.2, 0.2]),
        loss_cls=dict(
            type='FocalLoss',
            use_sigmoid=True,
            gamma=2.0,
            alpha=0.25,
            loss_weight=1.0),
        loss_bbox=dict(type='GIoULoss', loss_weight=2.0),
        loss_centerness=dict(
            type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0)),
    # training and testing settings
    train_cfg=dict(
        assigner=dict(type='ATSSAssigner', topk=9),
        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=0.6),
        max_per_img=100))

# dataset settings
train_pipeline = [
    dict(type='LoadImageFromFile', backend_args={{_base_.backend_args}}),
    dict(type='LoadAnnotations', with_bbox=True),
    dict(
        type='RandomResize',
        scale=[(2000, 480), (2000, 1200)],
        keep_ratio=True,
        backend='pillow'),
    dict(type='RandomFlip', prob=0.5),
    dict(type='PackDetInputs')
]
test_pipeline = [
    dict(type='LoadImageFromFile', backend_args={{_base_.backend_args}}),
    dict(type='Resize', scale=(2000, 1200), keep_ratio=True, backend='pillow'),
    dict(type='LoadAnnotations', with_bbox=True),
    dict(
        type='PackDetInputs',
        meta_keys=('img_id', 'img_path', 'ori_shape', 'img_shape',
                   'scale_factor'))
]
train_dataloader = dict(
    dataset=dict(
        _delete_=True,
        type='RepeatDataset',
        times=2,
        dataset=dict(
            type={{_base_.dataset_type}},
            data_root={{_base_.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={{_base_.backend_args}})))
val_dataloader = dict(dataset=dict(pipeline=test_pipeline))
test_dataloader = val_dataloader

# optimizer
optim_wrapper = dict(
    _delete_=True,
    type='OptimWrapper',
    optimizer=dict(
        type='AdamW', lr=0.00005, betas=(0.9, 0.999), weight_decay=0.05),
    paramwise_cfg=dict(
        custom_keys={
            'absolute_pos_embed': dict(decay_mult=0.),
            'relative_position_bias_table': dict(decay_mult=0.),
            'norm': dict(decay_mult=0.)
        }),
    clip_grad=None)