add new model

configs/body/2d_kpt_sview_rgb_img/topdown_heatmap/coco/scnet_coco.yml

+Collections:
+- Name: SCNet
+  Paper:
+    Title: Improving Convolutional Networks with Self-Calibrated Convolutions
+    URL: http://openaccess.thecvf.com/content_CVPR_2020/html/Liu_Improving_Convolutional_Networks_With_Self-Calibrated_Convolutions_CVPR_2020_paper.html
+  README: https://github.com/open-mmlab/mmpose/blob/master/docs/en/papers/backbones/scnet.md
+Models:
+- Config: configs/body/2d_kpt_sview_rgb_img/topdown_heatmap/coco/scnet50_coco_256x192.py
+  In Collection: SCNet
+  Metadata:
+    Architecture: &id001
+    - SCNet
+    Training Data: COCO
+  Name: topdown_heatmap_scnet50_coco_256x192
+  Results:
+  - Dataset: COCO
+    Metrics:
+      AP: 0.728
+      AP@0.5: 0.899
+      AP@0.75: 0.807
+      AR: 0.784
+      AR@0.5: 0.938
+    Task: Body 2D Keypoint
+  Weights: https://download.openmmlab.com/mmpose/top_down/scnet/scnet50_coco_256x192-6920f829_20200709.pth
+- Config: configs/body/2d_kpt_sview_rgb_img/topdown_heatmap/coco/scnet50_coco_384x288.py
+  In Collection: SCNet
+  Metadata:
+    Architecture: *id001
+    Training Data: COCO
+  Name: topdown_heatmap_scnet50_coco_384x288
+  Results:
+  - Dataset: COCO
+    Metrics:
+      AP: 0.751
+      AP@0.5: 0.906
+      AP@0.75: 0.818
+      AR: 0.802
+      AR@0.5: 0.943
+    Task: Body 2D Keypoint
+  Weights: https://download.openmmlab.com/mmpose/top_down/scnet/scnet50_coco_384x288-9cacd0ea_20200709.pth
+- Config: configs/body/2d_kpt_sview_rgb_img/topdown_heatmap/coco/scnet101_coco_256x192.py
+  In Collection: SCNet
+  Metadata:
+    Architecture: *id001
+    Training Data: COCO
+  Name: topdown_heatmap_scnet101_coco_256x192
+  Results:
+  - Dataset: COCO
+    Metrics:
+      AP: 0.733
+      AP@0.5: 0.903
+      AP@0.75: 0.813
+      AR: 0.79
+      AR@0.5: 0.941
+    Task: Body 2D Keypoint
+  Weights: https://download.openmmlab.com/mmpose/top_down/scnet/scnet101_coco_256x192-6d348ef9_20200709.pth
+- Config: configs/body/2d_kpt_sview_rgb_img/topdown_heatmap/coco/scnet101_coco_384x288.py
+  In Collection: SCNet
+  Metadata:
+    Architecture: *id001
+    Training Data: COCO
+  Name: topdown_heatmap_scnet101_coco_384x288
+  Results:
+  - Dataset: COCO
+    Metrics:
+      AP: 0.752
+      AP@0.5: 0.906
+      AP@0.75: 0.823
+      AR: 0.804
+      AR@0.5: 0.943
+    Task: Body 2D Keypoint
+  Weights: https://download.openmmlab.com/mmpose/top_down/scnet/scnet101_coco_384x288-0b6e631b_20200709.pth

configs/body/2d_kpt_sview_rgb_img/topdown_heatmap/coco/seresnet101_coco_256x192.py

+_base_ = [
+    '../../../../_base_/default_runtime.py',
+    '../../../../_base_/datasets/coco.py'
+]
+evaluation = dict(interval=10, metric='mAP', save_best='AP')
+
+optimizer = dict(
+    type='Adam',
+    lr=5e-4,
+)
+optimizer_config = dict(grad_clip=None)
+# learning policy
+lr_config = dict(
+    policy='step',
+    warmup='linear',
+    warmup_iters=500,
+    warmup_ratio=0.001,
+    step=[170, 200])
+total_epochs = 210
+channel_cfg = dict(
+    num_output_channels=17,
+    dataset_joints=17,
+    dataset_channel=[
+        [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16],
+    ],
+    inference_channel=[
+        0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16
+    ])
+
+# model settings
+model = dict(
+    type='TopDown',
+    pretrained='mmcls://se-resnet101',
+    backbone=dict(type='SEResNet', depth=101),
+    keypoint_head=dict(
+        type='TopdownHeatmapSimpleHead',
+        in_channels=2048,
+        out_channels=channel_cfg['num_output_channels'],
+        loss_keypoint=dict(type='JointsMSELoss', use_target_weight=True)),
+    train_cfg=dict(),
+    test_cfg=dict(
+        flip_test=True,
+        post_process='default',
+        shift_heatmap=True,
+        modulate_kernel=11))
+
+data_cfg = dict(
+    image_size=[192, 256],
+    heatmap_size=[48, 64],
+    num_output_channels=channel_cfg['num_output_channels'],
+    num_joints=channel_cfg['dataset_joints'],
+    dataset_channel=channel_cfg['dataset_channel'],
+    inference_channel=channel_cfg['inference_channel'],
+    soft_nms=False,
+    nms_thr=1.0,
+    oks_thr=0.9,
+    vis_thr=0.2,
+    use_gt_bbox=False,
+    det_bbox_thr=0.0,
+    bbox_file='data/coco/person_detection_results/'
+    'COCO_val2017_detections_AP_H_56_person.json',
+)
+
+train_pipeline = [
+    dict(type='LoadImageFromFile'),
+    dict(type='TopDownGetBboxCenterScale', padding=1.25),
+    dict(type='TopDownRandomShiftBboxCenter', shift_factor=0.16, prob=0.3),
+    dict(type='TopDownRandomFlip', flip_prob=0.5),
+    dict(
+        type='TopDownHalfBodyTransform',
+        num_joints_half_body=8,
+        prob_half_body=0.3),
+    dict(
+        type='TopDownGetRandomScaleRotation', rot_factor=40, scale_factor=0.5),
+    dict(type='TopDownAffine'),
+    dict(type='ToTensor'),
+    dict(
+        type='NormalizeTensor',
+        mean=[0.485, 0.456, 0.406],
+        std=[0.229, 0.224, 0.225]),
+    dict(type='TopDownGenerateTarget', sigma=2),
+    dict(
+        type='Collect',
+        keys=['img', 'target', 'target_weight'],
+        meta_keys=[
+            'image_file', 'joints_3d', 'joints_3d_visible', 'center', 'scale',
+            'rotation', 'bbox_score', 'flip_pairs'
+        ]),
+]
+
+val_pipeline = [
+    dict(type='LoadImageFromFile'),
+    dict(type='TopDownGetBboxCenterScale', padding=1.25),
+    dict(type='TopDownAffine'),
+    dict(type='ToTensor'),
+    dict(
+        type='NormalizeTensor',
+        mean=[0.485, 0.456, 0.406],
+        std=[0.229, 0.224, 0.225]),
+    dict(
+        type='Collect',
+        keys=['img'],
+        meta_keys=[
+            'image_file', 'center', 'scale', 'rotation', 'bbox_score',
+            'flip_pairs'
+        ]),
+]
+
+test_pipeline = val_pipeline
+
+data_root = 'data/coco'
+data = dict(
+    samples_per_gpu=64,
+    workers_per_gpu=2,
+    val_dataloader=dict(samples_per_gpu=32),
+    test_dataloader=dict(samples_per_gpu=32),
+    train=dict(
+        type='TopDownCocoDataset',
+        ann_file=f'{data_root}/annotations/person_keypoints_train2017.json',
+        img_prefix=f'{data_root}/train2017/',
+        data_cfg=data_cfg,
+        pipeline=train_pipeline,
+        dataset_info={{_base_.dataset_info}}),
+    val=dict(
+        type='TopDownCocoDataset',
+        ann_file=f'{data_root}/annotations/person_keypoints_val2017.json',
+        img_prefix=f'{data_root}/val2017/',
+        data_cfg=data_cfg,
+        pipeline=val_pipeline,
+        dataset_info={{_base_.dataset_info}}),
+    test=dict(
+        type='TopDownCocoDataset',
+        ann_file=f'{data_root}/annotations/person_keypoints_val2017.json',
+        img_prefix=f'{data_root}/val2017/',
+        data_cfg=data_cfg,
+        pipeline=test_pipeline,
+        dataset_info={{_base_.dataset_info}}),
+)

configs/body/2d_kpt_sview_rgb_img/topdown_heatmap/coco/seresnet101_coco_384x288.py

+_base_ = [
+    '../../../../_base_/default_runtime.py',
+    '../../../../_base_/datasets/coco.py'
+]
+evaluation = dict(interval=10, metric='mAP', save_best='AP')
+
+optimizer = dict(
+    type='Adam',
+    lr=5e-4,
+)
+optimizer_config = dict(grad_clip=None)
+# learning policy
+lr_config = dict(
+    policy='step',
+    warmup='linear',
+    warmup_iters=500,
+    warmup_ratio=0.001,
+    step=[170, 200])
+total_epochs = 210
+channel_cfg = dict(
+    num_output_channels=17,
+    dataset_joints=17,
+    dataset_channel=[
+        [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16],
+    ],
+    inference_channel=[
+        0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16
+    ])
+
+# model settings
+model = dict(
+    type='TopDown',
+    pretrained='mmcls://se-resnet101',
+    backbone=dict(type='SEResNet', depth=101),
+    keypoint_head=dict(
+        type='TopdownHeatmapSimpleHead',
+        in_channels=2048,
+        out_channels=channel_cfg['num_output_channels'],
+        loss_keypoint=dict(type='JointsMSELoss', use_target_weight=True)),
+    train_cfg=dict(),
+    test_cfg=dict(
+        flip_test=True,
+        post_process='default',
+        shift_heatmap=True,
+        modulate_kernel=11))
+
+data_cfg = dict(
+    image_size=[288, 384],
+    heatmap_size=[72, 96],
+    num_output_channels=channel_cfg['num_output_channels'],
+    num_joints=channel_cfg['dataset_joints'],
+    dataset_channel=channel_cfg['dataset_channel'],
+    inference_channel=channel_cfg['inference_channel'],
+    soft_nms=False,
+    nms_thr=1.0,
+    oks_thr=0.9,
+    vis_thr=0.2,
+    use_gt_bbox=False,
+    det_bbox_thr=0.0,
+    bbox_file='data/coco/person_detection_results/'
+    'COCO_val2017_detections_AP_H_56_person.json',
+)
+
+train_pipeline = [
+    dict(type='LoadImageFromFile'),
+    dict(type='TopDownGetBboxCenterScale', padding=1.25),
+    dict(type='TopDownRandomShiftBboxCenter', shift_factor=0.16, prob=0.3),
+    dict(type='TopDownRandomFlip', flip_prob=0.5),
+    dict(
+        type='TopDownHalfBodyTransform',
+        num_joints_half_body=8,
+        prob_half_body=0.3),
+    dict(
+        type='TopDownGetRandomScaleRotation', rot_factor=40, scale_factor=0.5),
+    dict(type='TopDownAffine'),
+    dict(type='ToTensor'),
+    dict(
+        type='NormalizeTensor',
+        mean=[0.485, 0.456, 0.406],
+        std=[0.229, 0.224, 0.225]),
+    dict(type='TopDownGenerateTarget', sigma=3),
+    dict(
+        type='Collect',
+        keys=['img', 'target', 'target_weight'],
+        meta_keys=[
+            'image_file', 'joints_3d', 'joints_3d_visible', 'center', 'scale',
+            'rotation', 'bbox_score', 'flip_pairs'
+        ]),
+]
+
+val_pipeline = [
+    dict(type='LoadImageFromFile'),
+    dict(type='TopDownGetBboxCenterScale', padding=1.25),
+    dict(type='TopDownAffine'),
+    dict(type='ToTensor'),
+    dict(
+        type='NormalizeTensor',
+        mean=[0.485, 0.456, 0.406],
+        std=[0.229, 0.224, 0.225]),
+    dict(
+        type='Collect',
+        keys=['img'],
+        meta_keys=[
+            'image_file', 'center', 'scale', 'rotation', 'bbox_score',
+            'flip_pairs'
+        ]),
+]
+
+test_pipeline = val_pipeline
+
+data_root = 'data/coco'
+data = dict(
+    samples_per_gpu=32,
+    workers_per_gpu=2,
+    val_dataloader=dict(samples_per_gpu=32),
+    test_dataloader=dict(samples_per_gpu=32),
+    train=dict(
+        type='TopDownCocoDataset',
+        ann_file=f'{data_root}/annotations/person_keypoints_train2017.json',
+        img_prefix=f'{data_root}/train2017/',
+        data_cfg=data_cfg,
+        pipeline=train_pipeline,
+        dataset_info={{_base_.dataset_info}}),
+    val=dict(
+        type='TopDownCocoDataset',
+        ann_file=f'{data_root}/annotations/person_keypoints_val2017.json',
+        img_prefix=f'{data_root}/val2017/',
+        data_cfg=data_cfg,
+        pipeline=val_pipeline,
+        dataset_info={{_base_.dataset_info}}),
+    test=dict(
+        type='TopDownCocoDataset',
+        ann_file=f'{data_root}/annotations/person_keypoints_val2017.json',
+        img_prefix=f'{data_root}/val2017/',
+        data_cfg=data_cfg,
+        pipeline=test_pipeline,
+        dataset_info={{_base_.dataset_info}}),
+)

configs/body/2d_kpt_sview_rgb_img/topdown_heatmap/coco/seresnet152_coco_256x192.py

+_base_ = [
+    '../../../../_base_/default_runtime.py',
+    '../../../../_base_/datasets/coco.py'
+]
+evaluation = dict(interval=10, metric='mAP', save_best='AP')
+
+optimizer = dict(
+    type='Adam',
+    lr=5e-4,
+)
+optimizer_config = dict(grad_clip=None)
+# learning policy
+lr_config = dict(
+    policy='step',
+    warmup='linear',
+    warmup_iters=500,
+    warmup_ratio=0.001,
+    step=[170, 200])
+total_epochs = 210
+channel_cfg = dict(
+    num_output_channels=17,
+    dataset_joints=17,
+    dataset_channel=[
+        [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16],
+    ],
+    inference_channel=[
+        0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16
+    ])
+
+# model settings
+model = dict(
+    type='TopDown',
+    pretrained=None,
+    backbone=dict(type='SEResNet', depth=152),
+    keypoint_head=dict(
+        type='TopdownHeatmapSimpleHead',
+        in_channels=2048,
+        out_channels=channel_cfg['num_output_channels'],
+        loss_keypoint=dict(type='JointsMSELoss', use_target_weight=True)),
+    train_cfg=dict(),
+    test_cfg=dict(
+        flip_test=True,
+        post_process='default',
+        shift_heatmap=True,
+        modulate_kernel=11))
+
+data_cfg = dict(
+    image_size=[192, 256],
+    heatmap_size=[48, 64],
+    num_output_channels=channel_cfg['num_output_channels'],
+    num_joints=channel_cfg['dataset_joints'],
+    dataset_channel=channel_cfg['dataset_channel'],
+    inference_channel=channel_cfg['inference_channel'],
+    soft_nms=False,
+    nms_thr=1.0,
+    oks_thr=0.9,
+    vis_thr=0.2,
+    use_gt_bbox=False,
+    det_bbox_thr=0.0,
+    bbox_file='data/coco/person_detection_results/'
+    'COCO_val2017_detections_AP_H_56_person.json',
+)
+
+train_pipeline = [
+    dict(type='LoadImageFromFile'),
+    dict(type='TopDownGetBboxCenterScale', padding=1.25),
+    dict(type='TopDownRandomShiftBboxCenter', shift_factor=0.16, prob=0.3),
+    dict(type='TopDownRandomFlip', flip_prob=0.5),
+    dict(
+        type='TopDownHalfBodyTransform',
+        num_joints_half_body=8,
+        prob_half_body=0.3),
+    dict(
+        type='TopDownGetRandomScaleRotation', rot_factor=40, scale_factor=0.5),
+    dict(type='TopDownAffine'),
+    dict(type='ToTensor'),
+    dict(
+        type='NormalizeTensor',
+        mean=[0.485, 0.456, 0.406],
+        std=[0.229, 0.224, 0.225]),
+    dict(type='TopDownGenerateTarget', sigma=2),
+    dict(
+        type='Collect',
+        keys=['img', 'target', 'target_weight'],
+        meta_keys=[
+            'image_file', 'joints_3d', 'joints_3d_visible', 'center', 'scale',
+            'rotation', 'bbox_score', 'flip_pairs'
+        ]),
+]
+
+val_pipeline = [
+    dict(type='LoadImageFromFile'),
+    dict(type='TopDownGetBboxCenterScale', padding=1.25),
+    dict(type='TopDownAffine'),
+    dict(type='ToTensor'),
+    dict(
+        type='NormalizeTensor',
+        mean=[0.485, 0.456, 0.406],
+        std=[0.229, 0.224, 0.225]),
+    dict(
+        type='Collect',
+        keys=['img'],
+        meta_keys=[
+            'image_file', 'center', 'scale', 'rotation', 'bbox_score',
+            'flip_pairs'
+        ]),
+]
+
+test_pipeline = val_pipeline
+
+data_root = 'data/coco'
+data = dict(
+    samples_per_gpu=32,
+    workers_per_gpu=2,
+    val_dataloader=dict(samples_per_gpu=32),
+    test_dataloader=dict(samples_per_gpu=32),
+    train=dict(
+        type='TopDownCocoDataset',
+        ann_file=f'{data_root}/annotations/person_keypoints_train2017.json',
+        img_prefix=f'{data_root}/train2017/',
+        data_cfg=data_cfg,
+        pipeline=train_pipeline,
+        dataset_info={{_base_.dataset_info}}),
+    val=dict(
+        type='TopDownCocoDataset',
+        ann_file=f'{data_root}/annotations/person_keypoints_val2017.json',
+        img_prefix=f'{data_root}/val2017/',
+        data_cfg=data_cfg,
+        pipeline=val_pipeline,
+        dataset_info={{_base_.dataset_info}}),
+    test=dict(
+        type='TopDownCocoDataset',
+        ann_file=f'{data_root}/annotations/person_keypoints_val2017.json',
+        img_prefix=f'{data_root}/val2017/',
+        data_cfg=data_cfg,
+        pipeline=test_pipeline,
+        dataset_info={{_base_.dataset_info}}),
+)

configs/body/2d_kpt_sview_rgb_img/topdown_heatmap/coco/seresnet152_coco_384x288.py

+_base_ = [
+    '../../../../_base_/default_runtime.py',
+    '../../../../_base_/datasets/coco.py'
+]
+evaluation = dict(interval=10, metric='mAP', save_best='AP')
+
+optimizer = dict(
+    type='Adam',
+    lr=5e-4,
+)
+optimizer_config = dict(grad_clip=None)
+# learning policy
+lr_config = dict(
+    policy='step',
+    warmup='linear',
+    warmup_iters=500,
+    warmup_ratio=0.001,
+    step=[170, 200])
+total_epochs = 210
+channel_cfg = dict(
+    num_output_channels=17,
+    dataset_joints=17,
+    dataset_channel=[
+        [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16],
+    ],
+    inference_channel=[
+        0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16
+    ])
+
+# model settings
+model = dict(
+    type='TopDown',
+    pretrained=None,
+    backbone=dict(type='SEResNet', depth=152),
+    keypoint_head=dict(
+        type='TopdownHeatmapSimpleHead',
+        in_channels=2048,
+        out_channels=channel_cfg['num_output_channels'],
+        loss_keypoint=dict(type='JointsMSELoss', use_target_weight=True)),
+    train_cfg=dict(),
+    test_cfg=dict(
+        flip_test=True,
+        post_process='default',
+        shift_heatmap=True,
+        modulate_kernel=11))
+
+data_cfg = dict(
+    image_size=[288, 384],
+    heatmap_size=[72, 96],
+    num_output_channels=channel_cfg['num_output_channels'],
+    num_joints=channel_cfg['dataset_joints'],
+    dataset_channel=channel_cfg['dataset_channel'],
+    inference_channel=channel_cfg['inference_channel'],
+    soft_nms=False,
+    nms_thr=1.0,
+    oks_thr=0.9,
+    vis_thr=0.2,
+    use_gt_bbox=False,
+    det_bbox_thr=0.0,
+    bbox_file='data/coco/person_detection_results/'
+    'COCO_val2017_detections_AP_H_56_person.json',
+)
+
+train_pipeline = [
+    dict(type='LoadImageFromFile'),
+    dict(type='TopDownGetBboxCenterScale', padding=1.25),
+    dict(type='TopDownRandomShiftBboxCenter', shift_factor=0.16, prob=0.3),
+    dict(type='TopDownRandomFlip', flip_prob=0.5),
+    dict(
+        type='TopDownHalfBodyTransform',
+        num_joints_half_body=8,
+        prob_half_body=0.3),
+    dict(
+        type='TopDownGetRandomScaleRotation', rot_factor=40, scale_factor=0.5),
+    dict(type='TopDownAffine'),
+    dict(type='ToTensor'),
+    dict(
+        type='NormalizeTensor',
+        mean=[0.485, 0.456, 0.406],
+        std=[0.229, 0.224, 0.225]),
+    dict(type='TopDownGenerateTarget', sigma=3),
+    dict(
+        type='Collect',
+        keys=['img', 'target', 'target_weight'],
+        meta_keys=[
+            'image_file', 'joints_3d', 'joints_3d_visible', 'center', 'scale',
+            'rotation', 'bbox_score', 'flip_pairs'
+        ]),
+]
+
+val_pipeline = [
+    dict(type='LoadImageFromFile'),
+    dict(type='TopDownGetBboxCenterScale', padding=1.25),
+    dict(type='TopDownAffine'),
+    dict(type='ToTensor'),
+    dict(
+        type='NormalizeTensor',
+        mean=[0.485, 0.456, 0.406],
+        std=[0.229, 0.224, 0.225]),
+    dict(
+        type='Collect',
+        keys=['img'],
+        meta_keys=[
+            'image_file', 'center', 'scale', 'rotation', 'bbox_score',
+            'flip_pairs'
+        ]),
+]
+
+test_pipeline = val_pipeline
+
+data_root = 'data/coco'
+data = dict(
+    samples_per_gpu=48,
+    workers_per_gpu=2,
+    val_dataloader=dict(samples_per_gpu=32),
+    test_dataloader=dict(samples_per_gpu=32),
+    train=dict(
+        type='TopDownCocoDataset',
+        ann_file=f'{data_root}/annotations/person_keypoints_train2017.json',
+        img_prefix=f'{data_root}/train2017/',
+        data_cfg=data_cfg,
+        pipeline=train_pipeline,
+        dataset_info={{_base_.dataset_info}}),
+    val=dict(
+        type='TopDownCocoDataset',
+        ann_file=f'{data_root}/annotations/person_keypoints_val2017.json',
+        img_prefix=f'{data_root}/val2017/',
+        data_cfg=data_cfg,
+        pipeline=val_pipeline,
+        dataset_info={{_base_.dataset_info}}),
+    test=dict(
+        type='TopDownCocoDataset',
+        ann_file=f'{data_root}/annotations/person_keypoints_val2017.json',
+        img_prefix=f'{data_root}/val2017/',
+        data_cfg=data_cfg,
+        pipeline=test_pipeline,
+        dataset_info={{_base_.dataset_info}}),
+)

configs/body/2d_kpt_sview_rgb_img/topdown_heatmap/coco/seresnet50_coco_256x192.py

+_base_ = [
+    '../../../../_base_/default_runtime.py',
+    '../../../../_base_/datasets/coco.py'
+]
+evaluation = dict(interval=10, metric='mAP', save_best='AP')
+
+optimizer = dict(
+    type='Adam',
+    lr=5e-4,
+)
+optimizer_config = dict(grad_clip=None)
+# learning policy
+lr_config = dict(
+    policy='step',
+    warmup='linear',
+    warmup_iters=500,
+    warmup_ratio=0.001,
+    step=[170, 200])
+total_epochs = 210
+channel_cfg = dict(
+    num_output_channels=17,
+    dataset_joints=17,
+    dataset_channel=[
+        [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16],
+    ],
+    inference_channel=[
+        0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16
+    ])
+
+# model settings
+model = dict(
+    type='TopDown',
+    pretrained='mmcls://se-resnet50',
+    backbone=dict(type='SEResNet', depth=50),
+    keypoint_head=dict(
+        type='TopdownHeatmapSimpleHead',
+        in_channels=2048,
+        out_channels=channel_cfg['num_output_channels'],
+        loss_keypoint=dict(type='JointsMSELoss', use_target_weight=True)),
+    train_cfg=dict(),
+    test_cfg=dict(
+        flip_test=True,
+        post_process='default',
+        shift_heatmap=True,
+        modulate_kernel=11))
+
+data_cfg = dict(
+    image_size=[192, 256],
+    heatmap_size=[48, 64],
+    num_output_channels=channel_cfg['num_output_channels'],
+    num_joints=channel_cfg['dataset_joints'],
+    dataset_channel=channel_cfg['dataset_channel'],
+    inference_channel=channel_cfg['inference_channel'],
+    soft_nms=False,
+    nms_thr=1.0,
+    oks_thr=0.9,
+    vis_thr=0.2,
+    use_gt_bbox=False,
+    det_bbox_thr=0.0,
+    bbox_file='data/coco/person_detection_results/'
+    'COCO_val2017_detections_AP_H_56_person.json',
+)
+
+train_pipeline = [
+    dict(type='LoadImageFromFile'),
+    dict(type='TopDownGetBboxCenterScale', padding=1.25),
+    dict(type='TopDownRandomShiftBboxCenter', shift_factor=0.16, prob=0.3),
+    dict(type='TopDownRandomFlip', flip_prob=0.5),
+    dict(
+        type='TopDownHalfBodyTransform',
+        num_joints_half_body=8,
+        prob_half_body=0.3),
+    dict(
+        type='TopDownGetRandomScaleRotation', rot_factor=40, scale_factor=0.5),
+    dict(type='TopDownAffine'),
+    dict(type='ToTensor'),
+    dict(
+        type='NormalizeTensor',
+        mean=[0.485, 0.456, 0.406],
+        std=[0.229, 0.224, 0.225]),
+    dict(type='TopDownGenerateTarget', sigma=2),
+    dict(
+        type='Collect',
+        keys=['img', 'target', 'target_weight'],
+        meta_keys=[
+            'image_file', 'joints_3d', 'joints_3d_visible', 'center', 'scale',
+            'rotation', 'bbox_score', 'flip_pairs'
+        ]),
+]
+
+val_pipeline = [
+    dict(type='LoadImageFromFile'),
+    dict(type='TopDownGetBboxCenterScale', padding=1.25),
+    dict(type='TopDownAffine'),
+    dict(type='ToTensor'),
+    dict(
+        type='NormalizeTensor',
+        mean=[0.485, 0.456, 0.406],
+        std=[0.229, 0.224, 0.225]),
+    dict(
+        type='Collect',
+        keys=['img'],
+        meta_keys=[
+            'image_file', 'center', 'scale', 'rotation', 'bbox_score',
+            'flip_pairs'
+        ]),
+]
+
+test_pipeline = val_pipeline
+
+data_root = 'data/coco'
+data = dict(
+    samples_per_gpu=64,
+    workers_per_gpu=2,
+    val_dataloader=dict(samples_per_gpu=32),
+    test_dataloader=dict(samples_per_gpu=32),
+    train=dict(
+        type='TopDownCocoDataset',
+        ann_file=f'{data_root}/annotations/person_keypoints_train2017.json',
+        img_prefix=f'{data_root}/train2017/',
+        data_cfg=data_cfg,
+        pipeline=train_pipeline,
+        dataset_info={{_base_.dataset_info}}),
+    val=dict(
+        type='TopDownCocoDataset',
+        ann_file=f'{data_root}/annotations/person_keypoints_val2017.json',
+        img_prefix=f'{data_root}/val2017/',
+        data_cfg=data_cfg,
+        pipeline=val_pipeline,
+        dataset_info={{_base_.dataset_info}}),
+    test=dict(
+        type='TopDownCocoDataset',
+        ann_file=f'{data_root}/annotations/person_keypoints_val2017.json',
+        img_prefix=f'{data_root}/val2017/',
+        data_cfg=data_cfg,
+        pipeline=test_pipeline,
+        dataset_info={{_base_.dataset_info}}),
+)

configs/body/2d_kpt_sview_rgb_img/topdown_heatmap/coco/seresnet50_coco_384x288.py

+_base_ = [
+    '../../../../_base_/default_runtime.py',
+    '../../../../_base_/datasets/coco.py'
+]
+evaluation = dict(interval=10, metric='mAP', save_best='AP')
+
+optimizer = dict(
+    type='Adam',
+    lr=5e-4,
+)
+optimizer_config = dict(grad_clip=None)
+# learning policy
+lr_config = dict(
+    policy='step',
+    warmup='linear',
+    warmup_iters=500,
+    warmup_ratio=0.001,
+    step=[170, 200])
+total_epochs = 210
+channel_cfg = dict(
+    num_output_channels=17,
+    dataset_joints=17,
+    dataset_channel=[
+        [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16],
+    ],
+    inference_channel=[
+        0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16
+    ])
+
+# model settings
+model = dict(
+    type='TopDown',
+    pretrained='mmcls://se-resnet50',
+    backbone=dict(type='SEResNet', depth=50),
+    keypoint_head=dict(
+        type='TopdownHeatmapSimpleHead',
+        in_channels=2048,
+        out_channels=channel_cfg['num_output_channels'],
+        loss_keypoint=dict(type='JointsMSELoss', use_target_weight=True)),
+    train_cfg=dict(),
+    test_cfg=dict(
+        flip_test=True,
+        post_process='default',
+        shift_heatmap=True,
+        modulate_kernel=11))
+
+data_cfg = dict(
+    image_size=[288, 384],
+    heatmap_size=[72, 96],
+    num_output_channels=channel_cfg['num_output_channels'],
+    num_joints=channel_cfg['dataset_joints'],
+    dataset_channel=channel_cfg['dataset_channel'],
+    inference_channel=channel_cfg['inference_channel'],
+    soft_nms=False,
+    nms_thr=1.0,
+    oks_thr=0.9,
+    vis_thr=0.2,
+    use_gt_bbox=False,
+    det_bbox_thr=0.0,
+    bbox_file='data/coco/person_detection_results/'
+    'COCO_val2017_detections_AP_H_56_person.json',
+)
+
+train_pipeline = [
+    dict(type='LoadImageFromFile'),
+    dict(type='TopDownGetBboxCenterScale', padding=1.25),
+    dict(type='TopDownRandomShiftBboxCenter', shift_factor=0.16, prob=0.3),
+    dict(type='TopDownRandomFlip', flip_prob=0.5),
+    dict(
+        type='TopDownHalfBodyTransform',
+        num_joints_half_body=8,
+        prob_half_body=0.3),
+    dict(
+        type='TopDownGetRandomScaleRotation', rot_factor=40, scale_factor=0.5),
+    dict(type='TopDownAffine'),
+    dict(type='ToTensor'),
+    dict(
+        type='NormalizeTensor',
+        mean=[0.485, 0.456, 0.406],
+        std=[0.229, 0.224, 0.225]),
+    dict(type='TopDownGenerateTarget', sigma=3),
+    dict(
+        type='Collect',
+        keys=['img', 'target', 'target_weight'],
+        meta_keys=[
+            'image_file', 'joints_3d', 'joints_3d_visible', 'center', 'scale',
+            'rotation', 'bbox_score', 'flip_pairs'
+        ]),
+]
+
+val_pipeline = [
+    dict(type='LoadImageFromFile'),
+    dict(type='TopDownGetBboxCenterScale', padding=1.25),
+    dict(type='TopDownAffine'),
+    dict(type='ToTensor'),
+    dict(
+        type='NormalizeTensor',
+        mean=[0.485, 0.456, 0.406],
+        std=[0.229, 0.224, 0.225]),
+    dict(
+        type='Collect',
+        keys=['img'],
+        meta_keys=[
+            'image_file', 'center', 'scale', 'rotation', 'bbox_score',
+            'flip_pairs'
+        ]),
+]
+
+test_pipeline = val_pipeline
+
+data_root = 'data/coco'
+data = dict(
+    samples_per_gpu=64,
+    workers_per_gpu=2,
+    val_dataloader=dict(samples_per_gpu=32),
+    test_dataloader=dict(samples_per_gpu=32),
+    train=dict(
+        type='TopDownCocoDataset',
+        ann_file=f'{data_root}/annotations/person_keypoints_train2017.json',
+        img_prefix=f'{data_root}/train2017/',
+        data_cfg=data_cfg,
+        pipeline=train_pipeline,
+        dataset_info={{_base_.dataset_info}}),
+    val=dict(
+        type='TopDownCocoDataset',
+        ann_file=f'{data_root}/annotations/person_keypoints_val2017.json',
+        img_prefix=f'{data_root}/val2017/',
+        data_cfg=data_cfg,
+        pipeline=val_pipeline,
+        dataset_info={{_base_.dataset_info}}),
+    test=dict(
+        type='TopDownCocoDataset',
+        ann_file=f'{data_root}/annotations/person_keypoints_val2017.json',
+        img_prefix=f'{data_root}/val2017/',
+        data_cfg=data_cfg,
+        pipeline=test_pipeline,
+        dataset_info={{_base_.dataset_info}}),
+)

configs/body/2d_kpt_sview_rgb_img/topdown_heatmap/coco/seresnet_coco.md

+<!-- [BACKBONE] -->
+
+<details>
+<summary align="right"><a href="http://openaccess.thecvf.com/content_cvpr_2018/html/Hu_Squeeze-and-Excitation_Networks_CVPR_2018_paper">SEResNet (CVPR'2018)</a></summary>
+
+```bibtex
+@inproceedings{hu2018squeeze,
+  title={Squeeze-and-excitation networks},
+  author={Hu, Jie and Shen, Li and Sun, Gang},
+  booktitle={Proceedings of the IEEE conference on computer vision and pattern recognition},
+  pages={7132--7141},
+  year={2018}
+}
+```
+
+</details>
+
+<!-- [DATASET] -->
+
+<details>
+<summary align="right"><a href="https://link.springer.com/chapter/10.1007/978-3-319-10602-1_48">COCO (ECCV'2014)</a></summary>
+
+```bibtex
+@inproceedings{lin2014microsoft,
+  title={Microsoft coco: Common objects in context},
+  author={Lin, Tsung-Yi and Maire, Michael and Belongie, Serge and Hays, James and Perona, Pietro and Ramanan, Deva and Doll{\'a}r, Piotr and Zitnick, C Lawrence},
+  booktitle={European conference on computer vision},
+  pages={740--755},
+  year={2014},
+  organization={Springer}
+}
+```
+
+</details>
+
+Results on COCO val2017 with detector having human AP of 56.4 on COCO val2017 dataset
+
+| Arch                                          | Input Size |  AP   | AP<sup>50</sup> | AP<sup>75</sup> |  AR   | AR<sup>50</sup> |                     ckpt                      |                      log                      |
+| :-------------------------------------------- | :--------: | :---: | :-------------: | :-------------: | :---: | :-------------: | :-------------------------------------------: | :-------------------------------------------: |
+| [pose_seresnet_50](/configs/body/2d_kpt_sview_rgb_img/topdown_heatmap/coco/seresnet50_coco_256x192.py) |  256x192   | 0.728 |      0.900      |      0.809      | 0.784 |      0.940      | [ckpt](https://download.openmmlab.com/mmpose/top_down/seresnet/seresnet50_coco_256x192-25058b66_20200727.pth) | [log](https://download.openmmlab.com/mmpose/top_down/seresnet/seresnet50_coco_256x192_20200727.log.json) |
+| [pose_seresnet_50](/configs/body/2d_kpt_sview_rgb_img/topdown_heatmap/coco/seresnet50_coco_384x288.py) |  384x288   | 0.748 |      0.905      |      0.819      | 0.799 |      0.941      | [ckpt](https://download.openmmlab.com/mmpose/top_down/seresnet/seresnet50_coco_384x288-bc0b7680_20200727.pth) | [log](https://download.openmmlab.com/mmpose/top_down/seresnet/seresnet50_coco_384x288_20200727.log.json) |
+| [pose_seresnet_101](/configs/body/2d_kpt_sview_rgb_img/topdown_heatmap/coco/seresnet101_coco_256x192.py) |  256x192   | 0.734 |      0.904      |      0.815      | 0.790 |      0.942      | [ckpt](https://download.openmmlab.com/mmpose/top_down/seresnet/seresnet101_coco_256x192-83f29c4d_20200727.pth) | [log](https://download.openmmlab.com/mmpose/top_down/seresnet/seresnet101_coco_256x192_20200727.log.json) |
+| [pose_seresnet_101](/configs/body/2d_kpt_sview_rgb_img/topdown_heatmap/coco/seresnet101_coco_384x288.py) |  384x288   | 0.753 |      0.907      |      0.823      | 0.805 |      0.943      | [ckpt](https://download.openmmlab.com/mmpose/top_down/seresnet/seresnet101_coco_384x288-48de1709_20200727.pth) | [log](https://download.openmmlab.com/mmpose/top_down/seresnet/seresnet101_coco_384x288_20200727.log.json) |
+| [pose_seresnet_152\*](/configs/body/2d_kpt_sview_rgb_img/topdown_heatmap/coco/seresnet152_coco_256x192.py) |  256x192   | 0.730 |      0.899      |      0.810      | 0.786 |      0.940      | [ckpt](https://download.openmmlab.com/mmpose/top_down/seresnet/seresnet152_coco_256x192-1c628d79_20200727.pth) | [log](https://download.openmmlab.com/mmpose/top_down/seresnet/seresnet152_coco_256x192_20200727.log.json) |
+| [pose_seresnet_152\*](/configs/body/2d_kpt_sview_rgb_img/topdown_heatmap/coco/seresnet152_coco_384x288.py) |  384x288   | 0.753 |      0.906      |      0.823      | 0.806 |      0.945      | [ckpt](https://download.openmmlab.com/mmpose/top_down/seresnet/seresnet152_coco_384x288-58b23ee8_20200727.pth) | [log](https://download.openmmlab.com/mmpose/top_down/seresnet/seresnet152_coco_384x288_20200727.log.json) |
+
+Note that * means without imagenet pre-training.

configs/body/2d_kpt_sview_rgb_img/topdown_heatmap/coco/seresnet_coco.yml

+Collections:
+- Name: SEResNet
+  Paper:
+    Title: Squeeze-and-excitation networks
+    URL: http://openaccess.thecvf.com/content_cvpr_2018/html/Hu_Squeeze-and-Excitation_Networks_CVPR_2018_paper
+  README: https://github.com/open-mmlab/mmpose/blob/master/docs/en/papers/backbones/seresnet.md
+Models:
+- Config: configs/body/2d_kpt_sview_rgb_img/topdown_heatmap/coco/seresnet50_coco_256x192.py
+  In Collection: SEResNet
+  Metadata:
+    Architecture: &id001
+    - SEResNet
+    Training Data: COCO
+  Name: topdown_heatmap_seresnet50_coco_256x192
+  Results:
+  - Dataset: COCO
+    Metrics:
+      AP: 0.728
+      AP@0.5: 0.9
+      AP@0.75: 0.809
+      AR: 0.784
+      AR@0.5: 0.94
+    Task: Body 2D Keypoint
+  Weights: https://download.openmmlab.com/mmpose/top_down/seresnet/seresnet50_coco_256x192-25058b66_20200727.pth
+- Config: configs/body/2d_kpt_sview_rgb_img/topdown_heatmap/coco/seresnet50_coco_384x288.py
+  In Collection: SEResNet
+  Metadata:
+    Architecture: *id001
+    Training Data: COCO
+  Name: topdown_heatmap_seresnet50_coco_384x288
+  Results:
+  - Dataset: COCO
+    Metrics:
+      AP: 0.748
+      AP@0.5: 0.905
+      AP@0.75: 0.819
+      AR: 0.799
+      AR@0.5: 0.941
+    Task: Body 2D Keypoint
+  Weights: https://download.openmmlab.com/mmpose/top_down/seresnet/seresnet50_coco_384x288-bc0b7680_20200727.pth
+- Config: configs/body/2d_kpt_sview_rgb_img/topdown_heatmap/coco/seresnet101_coco_256x192.py
+  In Collection: SEResNet
+  Metadata:
+    Architecture: *id001
+    Training Data: COCO
+  Name: topdown_heatmap_seresnet101_coco_256x192
+  Results:
+  - Dataset: COCO
+    Metrics:
+      AP: 0.734
+      AP@0.5: 0.904
+      AP@0.75: 0.815
+      AR: 0.79
+      AR@0.5: 0.942
+    Task: Body 2D Keypoint
+  Weights: https://download.openmmlab.com/mmpose/top_down/seresnet/seresnet101_coco_256x192-83f29c4d_20200727.pth
+- Config: configs/body/2d_kpt_sview_rgb_img/topdown_heatmap/coco/seresnet101_coco_384x288.py
+  In Collection: SEResNet
+  Metadata:
+    Architecture: *id001
+    Training Data: COCO
+  Name: topdown_heatmap_seresnet101_coco_384x288
+  Results:
+  - Dataset: COCO
+    Metrics:
+      AP: 0.753
+      AP@0.5: 0.907
+      AP@0.75: 0.823
+      AR: 0.805
+      AR@0.5: 0.943
+    Task: Body 2D Keypoint
+  Weights: https://download.openmmlab.com/mmpose/top_down/seresnet/seresnet101_coco_384x288-48de1709_20200727.pth
+- Config: configs/body/2d_kpt_sview_rgb_img/topdown_heatmap/coco/seresnet152_coco_256x192.py
+  In Collection: SEResNet
+  Metadata:
+    Architecture: *id001
+    Training Data: COCO
+  Name: topdown_heatmap_seresnet152_coco_256x192
+  Results:
+  - Dataset: COCO
+    Metrics:
+      AP: 0.73
+      AP@0.5: 0.899
+      AP@0.75: 0.81
+      AR: 0.786
+      AR@0.5: 0.94
+    Task: Body 2D Keypoint
+  Weights: https://download.openmmlab.com/mmpose/top_down/seresnet/seresnet152_coco_256x192-1c628d79_20200727.pth
+- Config: configs/body/2d_kpt_sview_rgb_img/topdown_heatmap/coco/seresnet152_coco_384x288.py
+  In Collection: SEResNet
+  Metadata:
+    Architecture: *id001
+    Training Data: COCO
+  Name: topdown_heatmap_seresnet152_coco_384x288
+  Results:
+  - Dataset: COCO
+    Metrics:
+      AP: 0.753
+      AP@0.5: 0.906
+      AP@0.75: 0.823
+      AR: 0.806
+      AR@0.5: 0.945
+    Task: Body 2D Keypoint
+  Weights: https://download.openmmlab.com/mmpose/top_down/seresnet/seresnet152_coco_384x288-58b23ee8_20200727.pth

configs/body/2d_kpt_sview_rgb_img/topdown_heatmap/coco/shufflenetv1_coco.md

+<!-- [BACKBONE] -->
+
+<details>
+<summary align="right"><a href="http://openaccess.thecvf.com/content_cvpr_2018/html/Zhang_ShuffleNet_An_Extremely_CVPR_2018_paper.html">ShufflenetV1 (CVPR'2018)</a></summary>
+
+```bibtex
+@inproceedings{zhang2018shufflenet,
+  title={Shufflenet: An extremely efficient convolutional neural network for mobile devices},
+  author={Zhang, Xiangyu and Zhou, Xinyu and Lin, Mengxiao and Sun, Jian},
+  booktitle={Proceedings of the IEEE conference on computer vision and pattern recognition},
+  pages={6848--6856},
+  year={2018}
+}
+```
+
+</details>
+
+<!-- [DATASET] -->
+
+<details>
+<summary align="right"><a href="https://link.springer.com/chapter/10.1007/978-3-319-10602-1_48">COCO (ECCV'2014)</a></summary>
+
+```bibtex
+@inproceedings{lin2014microsoft,
+  title={Microsoft coco: Common objects in context},
+  author={Lin, Tsung-Yi and Maire, Michael and Belongie, Serge and Hays, James and Perona, Pietro and Ramanan, Deva and Doll{\'a}r, Piotr and Zitnick, C Lawrence},
+  booktitle={European conference on computer vision},
+  pages={740--755},
+  year={2014},
+  organization={Springer}
+}
+```
+
+</details>
+
+Results on COCO val2017 with detector having human AP of 56.4 on COCO val2017 dataset
+
+| Arch                                          | Input Size |  AP   | AP<sup>50</sup> | AP<sup>75</sup> |  AR   | AR<sup>50</sup> |                     ckpt                      |                      log                      |
+| :-------------------------------------------- | :--------: | :---: | :-------------: | :-------------: | :---: | :-------------: | :-------------------------------------------: | :-------------------------------------------: |
+| [pose_shufflenetv1](/configs/body/2d_kpt_sview_rgb_img/topdown_heatmap/coco/shufflenetv1_coco_256x192.py) |  256x192   | 0.585 |      0.845      |      0.650      | 0.651 |      0.894      | [ckpt](https://download.openmmlab.com/mmpose/top_down/shufflenetv1/shufflenetv1_coco_256x192-353bc02c_20200727.pth) | [log](https://download.openmmlab.com/mmpose/top_down/shufflenetv1/shufflenetv1_coco_256x192_20200727.log.json) |
+| [pose_shufflenetv1](/configs/body/2d_kpt_sview_rgb_img/topdown_heatmap/coco/shufflenetv1_coco_384x288.py) |  384x288   | 0.622 |      0.859      |      0.685      | 0.684 |      0.901      | [ckpt](https://download.openmmlab.com/mmpose/top_down/shufflenetv1/shufflenetv1_coco_384x288-b2930b24_20200804.pth) | [log](https://download.openmmlab.com/mmpose/top_down/shufflenetv1/shufflenetv1_coco_384x288_20200804.log.json) |

configs/body/2d_kpt_sview_rgb_img/topdown_heatmap/coco/shufflenetv1_coco.yml

+Collections:
+- Name: ShufflenetV1
+  Paper:
+    Title: 'Shufflenet: An extremely efficient convolutional neural network for mobile
+      devices'
+    URL: http://openaccess.thecvf.com/content_cvpr_2018/html/Zhang_ShuffleNet_An_Extremely_CVPR_2018_paper.html
+  README: https://github.com/open-mmlab/mmpose/blob/master/docs/en/papers/backbones/shufflenetv1.md
+Models:
+- Config: configs/body/2d_kpt_sview_rgb_img/topdown_heatmap/coco/shufflenetv1_coco_256x192.py
+  In Collection: ShufflenetV1
+  Metadata:
+    Architecture: &id001
+    - ShufflenetV1
+    Training Data: COCO
+  Name: topdown_heatmap_shufflenetv1_coco_256x192
+  Results:
+  - Dataset: COCO
+    Metrics:
+      AP: 0.585
+      AP@0.5: 0.845
+      AP@0.75: 0.65
+      AR: 0.651
+      AR@0.5: 0.894
+    Task: Body 2D Keypoint
+  Weights: https://download.openmmlab.com/mmpose/top_down/shufflenetv1/shufflenetv1_coco_256x192-353bc02c_20200727.pth
+- Config: configs/body/2d_kpt_sview_rgb_img/topdown_heatmap/coco/shufflenetv1_coco_384x288.py
+  In Collection: ShufflenetV1
+  Metadata:
+    Architecture: *id001
+    Training Data: COCO
+  Name: topdown_heatmap_shufflenetv1_coco_384x288
+  Results:
+  - Dataset: COCO
+    Metrics:
+      AP: 0.622
+      AP@0.5: 0.859
+      AP@0.75: 0.685
+      AR: 0.684
+      AR@0.5: 0.901
+    Task: Body 2D Keypoint
+  Weights: https://download.openmmlab.com/mmpose/top_down/shufflenetv1/shufflenetv1_coco_384x288-b2930b24_20200804.pth

configs/body/2d_kpt_sview_rgb_img/topdown_heatmap/coco/shufflenetv1_coco_256x192.py

+_base_ = [
+    '../../../../_base_/default_runtime.py',
+    '../../../../_base_/datasets/coco.py'
+]
+evaluation = dict(interval=10, metric='mAP', save_best='AP')
+
+optimizer = dict(
+    type='Adam',
+    lr=5e-4,
+)
+optimizer_config = dict(grad_clip=None)
+# learning policy
+lr_config = dict(
+    policy='step',
+    warmup='linear',
+    warmup_iters=500,
+    warmup_ratio=0.001,
+    step=[170, 200])
+total_epochs = 210
+channel_cfg = dict(
+    num_output_channels=17,
+    dataset_joints=17,
+    dataset_channel=[
+        [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16],
+    ],
+    inference_channel=[
+        0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16
+    ])
+
+# model settings
+model = dict(
+    type='TopDown',
+    pretrained='mmcls://shufflenet_v1',
+    backbone=dict(type='ShuffleNetV1', groups=3),
+    keypoint_head=dict(
+        type='TopdownHeatmapSimpleHead',
+        in_channels=960,
+        out_channels=channel_cfg['num_output_channels'],
+        loss_keypoint=dict(type='JointsMSELoss', use_target_weight=True)),
+    train_cfg=dict(),
+    test_cfg=dict(
+        flip_test=True,
+        post_process='default',
+        shift_heatmap=True,
+        modulate_kernel=11))
+
+data_cfg = dict(
+    image_size=[192, 256],
+    heatmap_size=[48, 64],
+    num_output_channels=channel_cfg['num_output_channels'],
+    num_joints=channel_cfg['dataset_joints'],
+    dataset_channel=channel_cfg['dataset_channel'],
+    inference_channel=channel_cfg['inference_channel'],
+    soft_nms=False,
+    nms_thr=1.0,
+    oks_thr=0.9,
+    vis_thr=0.2,
+    use_gt_bbox=False,
+    det_bbox_thr=0.0,
+    bbox_file='data/coco/person_detection_results/'
+    'COCO_val2017_detections_AP_H_56_person.json',
+)
+
+train_pipeline = [
+    dict(type='LoadImageFromFile'),
+    dict(type='TopDownGetBboxCenterScale', padding=1.25),
+    dict(type='TopDownRandomShiftBboxCenter', shift_factor=0.16, prob=0.3),
+    dict(type='TopDownRandomFlip', flip_prob=0.5),
+    dict(
+        type='TopDownHalfBodyTransform',
+        num_joints_half_body=8,
+        prob_half_body=0.3),
+    dict(
+        type='TopDownGetRandomScaleRotation', rot_factor=40, scale_factor=0.5),
+    dict(type='TopDownAffine'),
+    dict(type='ToTensor'),
+    dict(
+        type='NormalizeTensor',
+        mean=[0.485, 0.456, 0.406],
+        std=[0.229, 0.224, 0.225]),
+    dict(type='TopDownGenerateTarget', sigma=2),
+    dict(
+        type='Collect',
+        keys=['img', 'target', 'target_weight'],
+        meta_keys=[
+            'image_file', 'joints_3d', 'joints_3d_visible', 'center', 'scale',
+            'rotation', 'bbox_score', 'flip_pairs'
+        ]),
+]
+
+val_pipeline = [
+    dict(type='LoadImageFromFile'),
+    dict(type='TopDownGetBboxCenterScale', padding=1.25),
+    dict(type='TopDownAffine'),
+    dict(type='ToTensor'),
+    dict(
+        type='NormalizeTensor',
+        mean=[0.485, 0.456, 0.406],
+        std=[0.229, 0.224, 0.225]),
+    dict(
+        type='Collect',
+        keys=['img'],
+        meta_keys=[
+            'image_file', 'center', 'scale', 'rotation', 'bbox_score',
+            'flip_pairs'
+        ]),
+]
+
+test_pipeline = val_pipeline
+
+data_root = 'data/coco'
+data = dict(
+    samples_per_gpu=64,
+    workers_per_gpu=2,
+    val_dataloader=dict(samples_per_gpu=32),
+    test_dataloader=dict(samples_per_gpu=32),
+    train=dict(
+        type='TopDownCocoDataset',
+        ann_file=f'{data_root}/annotations/person_keypoints_train2017.json',
+        img_prefix=f'{data_root}/train2017/',
+        data_cfg=data_cfg,
+        pipeline=train_pipeline,
+        dataset_info={{_base_.dataset_info}}),
+    val=dict(
+        type='TopDownCocoDataset',
+        ann_file=f'{data_root}/annotations/person_keypoints_val2017.json',
+        img_prefix=f'{data_root}/val2017/',
+        data_cfg=data_cfg,
+        pipeline=val_pipeline,
+        dataset_info={{_base_.dataset_info}}),
+    test=dict(
+        type='TopDownCocoDataset',
+        ann_file=f'{data_root}/annotations/person_keypoints_val2017.json',
+        img_prefix=f'{data_root}/val2017/',
+        data_cfg=data_cfg,
+        pipeline=test_pipeline,
+        dataset_info={{_base_.dataset_info}}),
+)

configs/body/2d_kpt_sview_rgb_img/topdown_heatmap/coco/shufflenetv1_coco_384x288.py

+_base_ = [
+    '../../../../_base_/default_runtime.py',
+    '../../../../_base_/datasets/coco.py'
+]
+evaluation = dict(interval=10, metric='mAP', save_best='AP')
+
+optimizer = dict(
+    type='Adam',
+    lr=5e-4,
+)
+optimizer_config = dict(grad_clip=None)
+# learning policy
+lr_config = dict(
+    policy='step',
+    warmup='linear',
+    warmup_iters=500,
+    warmup_ratio=0.001,
+    step=[170, 200])
+total_epochs = 210
+channel_cfg = dict(
+    num_output_channels=17,
+    dataset_joints=17,
+    dataset_channel=[
+        [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16],
+    ],
+    inference_channel=[
+        0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16
+    ])
+
+# model settings
+model = dict(
+    type='TopDown',
+    pretrained='mmcls://shufflenet_v1',
+    backbone=dict(type='ShuffleNetV1', groups=3),
+    keypoint_head=dict(
+        type='TopdownHeatmapSimpleHead',
+        in_channels=960,
+        out_channels=channel_cfg['num_output_channels'],
+        loss_keypoint=dict(type='JointsMSELoss', use_target_weight=True)),
+    train_cfg=dict(),
+    test_cfg=dict(
+        flip_test=True,
+        post_process='default',
+        shift_heatmap=True,
+        modulate_kernel=11))
+
+data_cfg = dict(
+    image_size=[288, 384],
+    heatmap_size=[72, 96],
+    num_output_channels=channel_cfg['num_output_channels'],
+    num_joints=channel_cfg['dataset_joints'],
+    dataset_channel=channel_cfg['dataset_channel'],
+    inference_channel=channel_cfg['inference_channel'],
+    soft_nms=False,
+    nms_thr=1.0,
+    oks_thr=0.9,
+    vis_thr=0.2,
+    use_gt_bbox=False,
+    det_bbox_thr=0.0,
+    bbox_file='data/coco/person_detection_results/'
+    'COCO_val2017_detections_AP_H_56_person.json',
+)
+
+train_pipeline = [
+    dict(type='LoadImageFromFile'),
+    dict(type='TopDownGetBboxCenterScale', padding=1.25),
+    dict(type='TopDownRandomShiftBboxCenter', shift_factor=0.16, prob=0.3),
+    dict(type='TopDownRandomFlip', flip_prob=0.5),
+    dict(
+        type='TopDownHalfBodyTransform',
+        num_joints_half_body=8,
+        prob_half_body=0.3),
+    dict(
+        type='TopDownGetRandomScaleRotation', rot_factor=40, scale_factor=0.5),
+    dict(type='TopDownAffine'),
+    dict(type='ToTensor'),
+    dict(
+        type='NormalizeTensor',
+        mean=[0.485, 0.456, 0.406],
+        std=[0.229, 0.224, 0.225]),
+    dict(type='TopDownGenerateTarget', sigma=3),
+    dict(
+        type='Collect',
+        keys=['img', 'target', 'target_weight'],
+        meta_keys=[
+            'image_file', 'joints_3d', 'joints_3d_visible', 'center', 'scale',
+            'rotation', 'bbox_score', 'flip_pairs'
+        ]),
+]
+
+val_pipeline = [
+    dict(type='LoadImageFromFile'),
+    dict(type='TopDownGetBboxCenterScale', padding=1.25),
+    dict(type='TopDownAffine'),
+    dict(type='ToTensor'),
+    dict(
+        type='NormalizeTensor',
+        mean=[0.485, 0.456, 0.406],
+        std=[0.229, 0.224, 0.225]),
+    dict(
+        type='Collect',
+        keys=['img'],
+        meta_keys=[
+            'image_file', 'center', 'scale', 'rotation', 'bbox_score',
+            'flip_pairs'
+        ]),
+]
+
+test_pipeline = val_pipeline
+
+data_root = 'data/coco'
+data = dict(
+    samples_per_gpu=64,
+    workers_per_gpu=2,
+    val_dataloader=dict(samples_per_gpu=32),
+    test_dataloader=dict(samples_per_gpu=32),
+    train=dict(
+        type='TopDownCocoDataset',
+        ann_file=f'{data_root}/annotations/person_keypoints_train2017.json',
+        img_prefix=f'{data_root}/train2017/',
+        data_cfg=data_cfg,
+        pipeline=train_pipeline,
+        dataset_info={{_base_.dataset_info}}),
+    val=dict(
+        type='TopDownCocoDataset',
+        ann_file=f'{data_root}/annotations/person_keypoints_val2017.json',
+        img_prefix=f'{data_root}/val2017/',
+        data_cfg=data_cfg,
+        pipeline=val_pipeline,
+        dataset_info={{_base_.dataset_info}}),
+    test=dict(
+        type='TopDownCocoDataset',
+        ann_file=f'{data_root}/annotations/person_keypoints_val2017.json',
+        img_prefix=f'{data_root}/val2017/',
+        data_cfg=data_cfg,
+        pipeline=test_pipeline,
+        dataset_info={{_base_.dataset_info}}),
+)

configs/body/2d_kpt_sview_rgb_img/topdown_heatmap/coco/shufflenetv2_coco.md

+<!-- [BACKBONE] -->
+
+<details>
+<summary align="right"><a href="http://openaccess.thecvf.com/content_ECCV_2018/html/Ningning_Light-weight_CNN_Architecture_ECCV_2018_paper.html">ShufflenetV2 (ECCV'2018)</a></summary>
+
+```bibtex
+@inproceedings{ma2018shufflenet,
+  title={Shufflenet v2: Practical guidelines for efficient cnn architecture design},
+  author={Ma, Ningning and Zhang, Xiangyu and Zheng, Hai-Tao and Sun, Jian},
+  booktitle={Proceedings of the European conference on computer vision (ECCV)},
+  pages={116--131},
+  year={2018}
+}
+```
+
+</details>
+
+<!-- [DATASET] -->
+
+<details>
+<summary align="right"><a href="https://link.springer.com/chapter/10.1007/978-3-319-10602-1_48">COCO (ECCV'2014)</a></summary>
+
+```bibtex
+@inproceedings{lin2014microsoft,
+  title={Microsoft coco: Common objects in context},
+  author={Lin, Tsung-Yi and Maire, Michael and Belongie, Serge and Hays, James and Perona, Pietro and Ramanan, Deva and Doll{\'a}r, Piotr and Zitnick, C Lawrence},
+  booktitle={European conference on computer vision},
+  pages={740--755},
+  year={2014},
+  organization={Springer}
+}
+```
+
+</details>
+
+Results on COCO val2017 with detector having human AP of 56.4 on COCO val2017 dataset
+
+| Arch                                          | Input Size |  AP   | AP<sup>50</sup> | AP<sup>75</sup> |  AR   | AR<sup>50</sup> |                     ckpt                      |                      log                      |
+| :-------------------------------------------- | :--------: | :---: | :-------------: | :-------------: | :---: | :-------------: | :-------------------------------------------: | :-------------------------------------------: |
+| [pose_shufflenetv2](/configs/body/2d_kpt_sview_rgb_img/topdown_heatmap/coco/shufflenetv2_coco_256x192.py) |  256x192   | 0.599 |      0.854      |      0.663      | 0.664 |      0.899      | [ckpt](https://download.openmmlab.com/mmpose/top_down/shufflenetv2/shufflenetv2_coco_256x192-0aba71c7_20200921.pth) | [log](https://download.openmmlab.com/mmpose/top_down/shufflenetv2/shufflenetv2_coco_256x192_20200921.log.json) |
+| [pose_shufflenetv2](/configs/body/2d_kpt_sview_rgb_img/topdown_heatmap/coco/shufflenetv2_coco_384x288.py) |  384x288   | 0.636 |      0.865      |      0.705      | 0.697 |      0.909      | [ckpt](https://download.openmmlab.com/mmpose/top_down/shufflenetv2/shufflenetv2_coco_384x288-fb38ac3a_20200921.pth) | [log](https://download.openmmlab.com/mmpose/top_down/shufflenetv2/shufflenetv2_coco_384x288_20200921.log.json) |

configs/body/2d_kpt_sview_rgb_img/topdown_heatmap/coco/shufflenetv2_coco.yml

+Collections:
+- Name: ShufflenetV2
+  Paper:
+    Title: 'Shufflenet v2: Practical guidelines for efficient cnn architecture design'
+    URL: http://openaccess.thecvf.com/content_ECCV_2018/html/Ningning_Light-weight_CNN_Architecture_ECCV_2018_paper.html
+  README: https://github.com/open-mmlab/mmpose/blob/master/docs/en/papers/backbones/shufflenetv2.md
+Models:
+- Config: configs/body/2d_kpt_sview_rgb_img/topdown_heatmap/coco/shufflenetv2_coco_256x192.py
+  In Collection: ShufflenetV2
+  Metadata:
+    Architecture: &id001
+    - ShufflenetV2
+    Training Data: COCO
+  Name: topdown_heatmap_shufflenetv2_coco_256x192
+  Results:
+  - Dataset: COCO
+    Metrics:
+      AP: 0.599
+      AP@0.5: 0.854
+      AP@0.75: 0.663
+      AR: 0.664
+      AR@0.5: 0.899
+    Task: Body 2D Keypoint
+  Weights: https://download.openmmlab.com/mmpose/top_down/shufflenetv2/shufflenetv2_coco_256x192-0aba71c7_20200921.pth
+- Config: configs/body/2d_kpt_sview_rgb_img/topdown_heatmap/coco/shufflenetv2_coco_384x288.py
+  In Collection: ShufflenetV2
+  Metadata:
+    Architecture: *id001
+    Training Data: COCO
+  Name: topdown_heatmap_shufflenetv2_coco_384x288
+  Results:
+  - Dataset: COCO
+    Metrics:
+      AP: 0.636
+      AP@0.5: 0.865
+      AP@0.75: 0.705
+      AR: 0.697
+      AR@0.5: 0.909
+    Task: Body 2D Keypoint
+  Weights: https://download.openmmlab.com/mmpose/top_down/shufflenetv2/shufflenetv2_coco_384x288-fb38ac3a_20200921.pth