add swin backbones and support solider pretrain

configs/body/2d_kpt_sview_rgb_img/topdown_heatmap/coco/swin_base_coco_384x288_release.py

@@ -0,0 +1,166 @@
+_base_ = ['../../../../_base_/datasets/coco.py']
+log_level = 'INFO'
+load_from = None
+resume_from = None
+dist_params = dict(backend='nccl')
+workflow = [('train', 1)]
+checkpoint_config = dict(interval=5, create_symlink=False)
+evaluation = dict(interval=5, metric='mAP', save_best='AP')
+
+optimizer = dict(
+    type='AdamW',
+    lr=1e-3,
+    betas=(0.9, 0.999),
+    weight_decay=0.01,
+    paramwise_cfg=dict(
+        custom_keys={'relative_position_bias_table': dict(decay_mult=0.)}))
+
+optimizer_config = dict(grad_clip=None)
+# learning policy
+lr_config = dict(
+    policy='step',
+    warmup='linear',
+    warmup_iters=11710,
+    warmup_ratio=0.001,
+    step=[120, 150])
+total_epochs = 160
+log_config = dict(
+    interval=50, hooks=[
+        dict(type='TextLoggerHook'),
+    ])
+
+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
+norm_cfg = dict(type='SyncBN', requires_grad=True)
+model = dict(
+    type='TopDown',
+    pretrained='./pretrain_models/swin_base.pth',
+    backbone=dict(
+        type='SwinTransformer',
+        in_channels=3,
+        pretrain_img_size=224,
+        patch_size=4,
+        window_size=7,
+        embed_dims=128,
+        strides=(4, 2, 1, 1),
+        depths=(2, 2, 18, 2),
+        num_heads=(4, 8, 16, 32),
+        drop_path_rate=0.0,
+        drop_rate=0.0,
+        attn_drop_rate=0.0,
+        semantic_weight=0.8,),
+    keypoint_head=dict(
+        type='TopdownHeatmapSimpleHead',
+        in_channels=1024,
+        in_index=3,
+        out_channels=channel_cfg['num_output_channels'],
+        num_deconv_layers=1,
+        num_deconv_kernels=(4, ),
+        num_deconv_filters=(256, ),
+        #in_index=-1,
+        extra=dict(final_conv_kernel=1, ),
+        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_root = 'data/coco'
+data_cfg = dict(
+    image_size=[288, 384],
+    heatmap_size=[72, 96], #[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=f'{data_root}/person_detection_results/'
+    'COCO_val2017_detections_AP_H_56_person.json',
+)
+
+train_pipeline = [
+    dict(type='LoadImageFromFile'),
+    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='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 = dict(
+    samples_per_gpu=12,
+    workers_per_gpu=2,
+    val_dataloader=dict(samples_per_gpu=12),
+    test_dataloader=dict(samples_per_gpu=12),
+    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),
+    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),
+    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=val_pipeline),
+)
+
+# fp16 settings
+fp16 = dict(loss_scale='dynamic')

configs/body/2d_kpt_sview_rgb_img/topdown_heatmap/coco/swin_small_coco_384x288_release.py

@@ -0,0 +1,166 @@
+_base_ = ['../../../../_base_/datasets/coco.py']
+log_level = 'INFO'
+load_from = None
+resume_from = None
+dist_params = dict(backend='nccl')
+workflow = [('train', 1)]
+checkpoint_config = dict(interval=5, create_symlink=False)
+evaluation = dict(interval=5, metric='mAP', save_best='AP')
+
+optimizer = dict(
+    type='AdamW',
+    lr=1e-3,
+    betas=(0.9, 0.999),
+    weight_decay=0.01,
+    paramwise_cfg=dict(
+        custom_keys={'relative_position_bias_table': dict(decay_mult=0.)}))
+
+optimizer_config = dict(grad_clip=None)
+# learning policy
+lr_config = dict(
+    policy='step',
+    warmup='linear',
+    warmup_iters=11710,
+    warmup_ratio=0.001,
+    step=[170, 200])
+total_epochs = 210
+log_config = dict(
+    interval=50, hooks=[
+        dict(type='TextLoggerHook'),
+    ])
+
+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
+norm_cfg = dict(type='SyncBN', requires_grad=True)
+model = dict(
+    type='TopDown',
+    pretrained='./pretrain_models/swin_small.pth',
+    backbone=dict(
+        type='SwinTransformer',
+        in_channels=3,
+        pretrain_img_size=224,
+        patch_size=4,
+        window_size=7,
+        embed_dims=96,
+        strides=(4, 2, 1, 1),
+        depths=(2, 2, 18, 2),
+        num_heads=(3, 6, 12, 24),
+        drop_path_rate=0.0,
+        drop_rate=0.0,
+        attn_drop_rate=0.0,
+        semantic_weight=0.8),
+    keypoint_head=dict(
+        type='TopdownHeatmapSimpleHead',
+        in_channels=768,
+        in_index=3,
+        out_channels=channel_cfg['num_output_channels'],
+        num_deconv_layers=1,
+        num_deconv_kernels=(4, ),
+        num_deconv_filters=(256, ),
+        #in_index=-1,
+        extra=dict(final_conv_kernel=1, ),
+        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_root = 'data/coco'
+data_cfg = dict(
+    image_size=[288, 384],
+    heatmap_size=[72, 96], #[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=f'{data_root}/person_detection_results/'
+    'COCO_val2017_detections_AP_H_56_person.json',
+)
+
+train_pipeline = [
+    dict(type='LoadImageFromFile'),
+    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='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 = dict(
+    samples_per_gpu=8,
+    workers_per_gpu=2,
+    val_dataloader=dict(samples_per_gpu=8),
+    test_dataloader=dict(samples_per_gpu=8),
+    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),
+    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),
+    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=val_pipeline),
+)
+
+# fp16 settings
+fp16 = dict(loss_scale='dynamic')