auto_scale_lr = dict(base_batch_size=256) backend_args = dict(backend='local') codec = dict( heatmap_size=( 64, 64, ), input_size=( 256, 256, ), sigma=2, type='MSRAHeatmap', unbiased=True) custom_hooks = [ dict(type='SyncBuffersHook'), ] data_mode = 'topdown' data_root = 'A:\\NonEnclosureProjects\\inprep\\PrimateFace\\data\\annos_from_rex' dataset_info = dict( dataset_name='coco_wholebody_face', joint_weights=[ 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, ], keypoint_info=dict({ 0: dict( color=[ 255, 0, 0, ], id=0, name='face-0', swap='face-16', type=''), 1: dict( color=[ 255, 0, 0, ], id=1, name='face-1', swap='face-15', type=''), 10: dict( color=[ 255, 0, 0, ], id=10, name='face-10', swap='face-6', type=''), 11: dict( color=[ 255, 0, 0, ], id=11, name='face-11', swap='face-5', type=''), 12: dict( color=[ 255, 0, 0, ], id=12, name='face-12', swap='face-4', type=''), 13: dict( color=[ 255, 0, 0, ], id=13, name='face-13', swap='face-3', type=''), 14: dict( color=[ 255, 0, 0, ], id=14, name='face-14', swap='face-2', type=''), 15: dict( color=[ 255, 0, 0, ], id=15, name='face-15', swap='face-1', type=''), 16: dict( color=[ 255, 0, 0, ], id=16, name='face-16', swap='face-0', type=''), 17: dict( color=[ 255, 0, 0, ], id=17, name='face-17', swap='face-26', type=''), 18: dict( color=[ 255, 0, 0, ], id=18, name='face-18', swap='face-25', type=''), 19: dict( color=[ 255, 0, 0, ], id=19, name='face-19', swap='face-24', type=''), 2: dict( color=[ 255, 0, 0, ], id=2, name='face-2', swap='face-14', type=''), 20: dict( color=[ 255, 0, 0, ], id=20, name='face-20', swap='face-23', type=''), 21: dict( color=[ 255, 0, 0, ], id=21, name='face-21', swap='face-22', type=''), 22: dict( color=[ 255, 0, 0, ], id=22, name='face-22', swap='face-21', type=''), 23: dict( color=[ 255, 0, 0, ], id=23, name='face-23', swap='face-20', type=''), 24: dict( color=[ 255, 0, 0, ], id=24, name='face-24', swap='face-19', type=''), 25: dict( color=[ 255, 0, 0, ], id=25, name='face-25', swap='face-18', type=''), 26: dict( color=[ 255, 0, 0, ], id=26, name='face-26', swap='face-17', type=''), 27: dict(color=[ 255, 0, 0, ], id=27, name='face-27', swap='', type=''), 28: dict(color=[ 255, 0, 0, ], id=28, name='face-28', swap='', type=''), 29: dict(color=[ 255, 0, 0, ], id=29, name='face-29', swap='', type=''), 3: dict( color=[ 255, 0, 0, ], id=3, name='face-3', swap='face-13', type=''), 30: dict(color=[ 255, 0, 0, ], id=30, name='face-30', swap='', type=''), 31: dict( color=[ 255, 0, 0, ], id=31, name='face-31', swap='face-35', type=''), 32: dict( color=[ 255, 0, 0, ], id=32, name='face-32', swap='face-34', type=''), 33: dict(color=[ 255, 0, 0, ], id=33, name='face-33', swap='', type=''), 34: dict( color=[ 255, 0, 0, ], id=34, name='face-34', swap='face-32', type=''), 35: dict( color=[ 255, 0, 0, ], id=35, name='face-35', swap='face-31', type=''), 36: dict( color=[ 255, 0, 0, ], id=36, name='face-36', swap='face-45', type=''), 37: dict( color=[ 255, 0, 0, ], id=37, name='face-37', swap='face-44', type=''), 38: dict( color=[ 255, 0, 0, ], id=38, name='face-38', swap='face-43', type=''), 39: dict( color=[ 255, 0, 0, ], id=39, name='face-39', swap='face-42', type=''), 4: dict( color=[ 255, 0, 0, ], id=4, name='face-4', swap='face-12', type=''), 40: dict( color=[ 255, 0, 0, ], id=40, name='face-40', swap='face-47', type=''), 41: dict( color=[ 255, 0, 0, ], id=41, name='face-41', swap='face-46', type=''), 42: dict( color=[ 255, 0, 0, ], id=42, name='face-42', swap='face-39', type=''), 43: dict( color=[ 255, 0, 0, ], id=43, name='face-43', swap='face-38', type=''), 44: dict( color=[ 255, 0, 0, ], id=44, name='face-44', swap='face-37', type=''), 45: dict( color=[ 255, 0, 0, ], id=45, name='face-45', swap='face-36', type=''), 46: dict( color=[ 255, 0, 0, ], id=46, name='face-46', swap='face-41', type=''), 47: dict( color=[ 255, 0, 0, ], id=47, name='face-47', swap='face-40', type=''), 48: dict( color=[ 255, 0, 0, ], id=48, name='face-48', swap='face-54', type=''), 49: dict( color=[ 255, 0, 0, ], id=49, name='face-49', swap='face-53', type=''), 5: dict( color=[ 255, 0, 0, ], id=5, name='face-5', swap='face-11', type=''), 50: dict( color=[ 255, 0, 0, ], id=50, name='face-50', swap='face-52', type=''), 51: dict(color=[ 255, 0, 0, ], id=52, name='face-51', swap='', type=''), 52: dict( color=[ 255, 0, 0, ], id=52, name='face-52', swap='face-50', type=''), 53: dict( color=[ 255, 0, 0, ], id=53, name='face-53', swap='face-49', type=''), 54: dict( color=[ 255, 0, 0, ], id=54, name='face-54', swap='face-48', type=''), 55: dict( color=[ 255, 0, 0, ], id=55, name='face-55', swap='face-59', type=''), 56: dict( color=[ 255, 0, 0, ], id=56, name='face-56', swap='face-58', type=''), 57: dict(color=[ 255, 0, 0, ], id=57, name='face-57', swap='', type=''), 58: dict( color=[ 255, 0, 0, ], id=58, name='face-58', swap='face-56', type=''), 59: dict( color=[ 255, 0, 0, ], id=59, name='face-59', swap='face-55', type=''), 6: dict( color=[ 255, 0, 0, ], id=6, name='face-6', swap='face-10', type=''), 60: dict( color=[ 255, 0, 0, ], id=60, name='face-60', swap='face-64', type=''), 61: dict( color=[ 255, 0, 0, ], id=61, name='face-61', swap='face-63', type=''), 62: dict(color=[ 255, 0, 0, ], id=62, name='face-62', swap='', type=''), 63: dict( color=[ 255, 0, 0, ], id=63, name='face-63', swap='face-61', type=''), 64: dict( color=[ 255, 0, 0, ], id=64, name='face-64', swap='face-60', type=''), 65: dict( color=[ 255, 0, 0, ], id=65, name='face-65', swap='face-67', type=''), 66: dict(color=[ 255, 0, 0, ], id=66, name='face-66', swap='', type=''), 67: dict( color=[ 255, 0, 0, ], id=67, name='face-67', swap='face-65', type=''), 7: dict(color=[ 255, 0, 0, ], id=7, name='face-7', swap='face-9', type=''), 8: dict(color=[ 255, 0, 0, ], id=8, name='face-8', swap='', type=''), 9: dict(color=[ 255, 0, 0, ], id=9, name='face-9', swap='face-7', type='') }), paper_info=dict( author= 'Jin, Sheng and Xu, Lumin and Xu, Jin and Wang, Can and Liu, Wentao and Qian, Chen and Ouyang, Wanli and Luo, Ping', container= 'Proceedings of the European Conference on Computer Vision (ECCV)', homepage='https://github.com/jin-s13/COCO-WholeBody/', title='Whole-Body Human Pose Estimation in the Wild', year='2020'), sigmas=[ 0.042, 0.043, 0.044, 0.043, 0.04, 0.035, 0.031, 0.025, 0.02, 0.023, 0.029, 0.032, 0.037, 0.038, 0.043, 0.041, 0.045, 0.013, 0.012, 0.011, 0.011, 0.012, 0.012, 0.011, 0.011, 0.013, 0.015, 0.009, 0.007, 0.007, 0.007, 0.012, 0.009, 0.008, 0.016, 0.01, 0.017, 0.011, 0.009, 0.011, 0.009, 0.007, 0.013, 0.008, 0.011, 0.012, 0.01, 0.034, 0.008, 0.008, 0.009, 0.008, 0.008, 0.007, 0.01, 0.008, 0.009, 0.009, 0.009, 0.007, 0.007, 0.008, 0.011, 0.008, 0.008, 0.008, 0.01, 0.008, ], skeleton_info=dict()) dataset_type = 'CocoWholeBodyFaceDataset' default_hooks = dict( badcase=dict( badcase_thr=5, enable=False, metric_type='loss', out_dir='badcase', type='BadCaseAnalysisHook'), checkpoint=dict( interval=1, rule='less', save_best='NME', type='CheckpointHook'), logger=dict(interval=50, type='LoggerHook'), param_scheduler=dict(type='ParamSchedulerHook'), sampler_seed=dict(type='DistSamplerSeedHook'), timer=dict(type='IterTimerHook'), visualization=dict(enable=False, type='PoseVisualizationHook')) default_scope = 'mmpose' env_cfg = dict( cudnn_benchmark=False, dist_cfg=dict(backend='gloo'), mp_cfg=dict(mp_start_method='fork', opencv_num_threads=0)) launcher = 'pytorch' load_from = None log_level = 'INFO' log_processor = dict( by_epoch=True, num_digits=6, type='LogProcessor', window_size=50) model = dict( backbone=dict( extra=dict( stage1=dict( block='BOTTLENECK', num_blocks=(4, ), num_branches=1, num_channels=(64, ), num_modules=1), stage2=dict( block='BASIC', num_blocks=( 4, 4, ), num_branches=2, num_channels=( 18, 36, ), num_modules=1), stage3=dict( block='BASIC', num_blocks=( 4, 4, 4, ), num_branches=3, num_channels=( 18, 36, 72, ), num_modules=4), stage4=dict( block='BASIC', multiscale_output=True, num_blocks=( 4, 4, 4, 4, ), num_branches=4, num_channels=( 18, 36, 72, 144, ), num_modules=3), upsample=dict(align_corners=False, mode='bilinear')), in_channels=3, init_cfg=dict( checkpoint='open-mmlab://msra/hrnetv2_w18', type='Pretrained'), type='HRNet'), data_preprocessor=dict( bgr_to_rgb=True, mean=[ 123.675, 116.28, 103.53, ], std=[ 58.395, 57.12, 57.375, ], type='PoseDataPreprocessor'), head=dict( conv_kernel_sizes=(1, ), conv_out_channels=(270, ), decoder=dict( heatmap_size=( 64, 64, ), input_size=( 256, 256, ), sigma=2, type='MSRAHeatmap', unbiased=True), deconv_out_channels=None, in_channels=270, loss=dict(type='KeypointMSELoss', use_target_weight=True), out_channels=68, type='HeatmapHead'), neck=dict(concat=True, type='FeatureMapProcessor'), test_cfg=dict(flip_mode='heatmap', flip_test=True, shift_heatmap=True), type='TopdownPoseEstimator') optim_wrapper = dict( loss_scale='dynamic', optimizer=dict(lr=0.002, type='Adam'), type='AmpOptimWrapper') param_scheduler = [ dict( begin=0, by_epoch=False, end=500, start_factor=0.001, type='LinearLR'), dict( begin=0, by_epoch=True, end=210, gamma=0.1, milestones=[ 40, 55, ], type='MultiStepLR'), ] resume = False test_cfg = dict() test_dataloader = dict( batch_size=32, dataset=dict( ann_file= 'A:\\NonEnclosureProjects\\inprep\\PrimateFace\\data\\annos_from_rex\\test.json', data_mode='topdown', pipeline=[ dict(type='LoadImage'), dict(type='GetBBoxCenterScale'), dict(input_size=( 256, 256, ), type='TopdownAffine'), dict(type='PackPoseInputs'), ], test_mode=True, type='CocoWholeBodyFaceDataset'), drop_last=False, num_workers=2, persistent_workers=True, sampler=dict(round_up=False, shuffle=False, type='DefaultSampler')) test_evaluator = dict(norm_mode='keypoint_distance', type='NME') train_cfg = dict(by_epoch=True, max_epochs=60, val_interval=1) train_dataloader = dict( batch_size=64, dataset=dict( ann_file= 'A:\\NonEnclosureProjects\\inprep\\PrimateFace\\data\\annos_from_rex\\train.json', data_mode='topdown', pipeline=[ dict(type='LoadImage'), dict(type='GetBBoxCenterScale'), dict(direction='horizontal', type='RandomFlip'), dict(direction='vertical', type='RandomFlip'), dict( rotate_factor=60, scale_factor=( 0.75, 1.25, ), type='RandomBBoxTransform'), dict(input_size=( 256, 256, ), type='TopdownAffine'), dict( encoder=dict( heatmap_size=( 64, 64, ), input_size=( 256, 256, ), sigma=2, type='MSRAHeatmap', unbiased=True), type='GenerateTarget'), dict(type='PackPoseInputs'), ], type='CocoWholeBodyFaceDataset'), num_workers=4, persistent_workers=True, sampler=dict(shuffle=True, type='DefaultSampler')) train_pipeline = [ dict(type='LoadImage'), dict(type='GetBBoxCenterScale'), dict(direction='horizontal', type='RandomFlip'), dict(direction='vertical', type='RandomFlip'), dict( rotate_factor=60, scale_factor=( 0.75, 1.25, ), type='RandomBBoxTransform'), dict(input_size=( 256, 256, ), type='TopdownAffine'), dict( encoder=dict( heatmap_size=( 64, 64, ), input_size=( 256, 256, ), sigma=2, type='MSRAHeatmap', unbiased=True), type='GenerateTarget'), dict(type='PackPoseInputs'), ] val_cfg = dict() val_dataloader = dict( batch_size=32, dataset=dict( ann_file= 'A:\\NonEnclosureProjects\\inprep\\PrimateFace\\data\\annos_from_rex\\val.json', data_mode='topdown', pipeline=[ dict(type='LoadImage'), dict(type='GetBBoxCenterScale'), dict(input_size=( 256, 256, ), type='TopdownAffine'), dict(type='PackPoseInputs'), ], test_mode=True, type='CocoWholeBodyFaceDataset'), drop_last=False, num_workers=2, persistent_workers=True, sampler=dict(round_up=False, shuffle=False, type='DefaultSampler')) val_evaluator = dict(norm_mode='keypoint_distance', type='NME') val_pipeline = [ dict(type='LoadImage'), dict(type='GetBBoxCenterScale'), dict(input_size=( 256, 256, ), type='TopdownAffine'), dict(type='PackPoseInputs'), ] vis_backends = [ dict(type='LocalVisBackend'), ] visualizer = dict( name='visualizer', type='PoseLocalVisualizer', vis_backends=[ dict(type='LocalVisBackend'), ]) work_dir = 'A:\\NonEnclosureProjects\\inprep\\PrimateFace\\results\\hrnet-dark_rex_4.5k_68kpt_250528'